University of Pretoria

Department of Economics Working Paper Series Halloween Effect in Developed Stock Markets A US Perspective Alex Plastun Sumy State University Xolani Sibande University of Pretoria Rangan Gupta University of Pretoria Mark E Wohar University of Nebraska at Omaha and Loughborough University Working Paper 2019-14 February 2019 __________________________________________________________ Department of Economics University of Pretoria 0002 Pretoria South Africa Tel +27 12 420 2413

Halloween Effect in Developed Stock Markets AUS Perspective

Alex Plastunlowast Xolani Sibandedagger Rangan GuptaDagger Mark E Woharsect

February 19 2019

Abstract

In this paper we conduct a comprehensive investigation of the Halloweeneffect evolution in the US stock market over its entire history We employvarious statistical techniques (average analysis Studentrsquos t-test ANOVAand the Mann-Whitney test) and the trading simulation approach to analysethe evolution of the Halloween effect The results suggest that in the USstock market the Halloween effect became more persistent since the middleof the 20th century Despite the decline in its prevalence since that timenowadays it is still present in the US stock market and provides opportunitiesto build a trading strategy which can beat the market These results arewell in line with other developed stock markets Therefore in the main ourresults are inconsistent with the Efficient Market Hypothesis

Keywords Calendar Anomalies Halloween Effect Stock Market Efficient Mar-ket HypothesisJEL Codes G12 C63

lowastFaculty of Economics and Management Sumy State University Sumy Ukraine Emailoplastungmailcom This author gratefully acknowledges financial support from the Ministryof Education and Science of Ukraine (0117U003936)

daggerDepartment of Economics University of Pretoria Pretoria South Africa Emailxolanissgmailcom

DaggerDepartment of Economics University of Pretoria Pretoria South Africa Email ran-ganguptaupacza

sectCorresponding Author College of Business Administration University of Nebraska USSchool of Business and Economics Loughborough University Leicestershire LE11 3TU UKEmail mwoharunomahaedu

1

1 Introduction

Bouman amp Jacobsen (2002) observed that every year around May the Europeanpress would remark that buyers in stock markets seem to have sold in May andgone away This was referring to a phenomenon where the month of May wouldsignal the beginning of a bear market and investors would be better off sellingand holding cash That is until November when the bull market would returnThis phenomenon is known as the Halloween effect

Essentially the Halloween effect indicates that returns between November andApril are higher than in the other months of the year Bouman amp Jacobsen (2002)notes that this difference in returns cannot be explained by factors such as riskvolatility or the January effect amongst others Furthermore these profits wereeconomically exploitable However some in the literature suggest that the Hal-loween effect is a statistical aberration due to for example data outliers (Maberlyamp Pierce (2003) Maberly amp Pierce (2004) Dichtl amp Drobetz (2015) and Dichtlamp Drobetz (2014)) whilst others confirm its continued existence (Jacobsen ampVisaltanachoti (2009) Guo et al (2014) and Lean (2011)) and others suggestthat it has since disappeared (Siriopoulos amp Giannopoulo (2006))

If stock markets follow the Efficient Market Hypothesis (EMH) then marketanomalies (price firm-size and calendar) should not exist at all (Fama (1965)Fama (1970) and Jensen (1978)) In instances where abnormal profits are ex-ploitable their discovery of should result in them disappearing (Schwert (2003))However as noted by Bouman amp Jacobsen (2002) the Halloween effect does notseem to suffer from Murphyrsquos law and has persisted Several reasons have beenadvanced for this in the literature with the most prominent focusing on psycholog-ical and environmental factors underpinning investor behaviour (Kelly amp Meschke(2010) Kamstra et al (2003) Cao amp Wei (2005) and Jacobsen amp Marquering(2008))

Several studies focused on the United States (US) stock market Bouman ampJacobsen (2002) were the first to confirm the existence of the Halloween effect inthe US stock market amongst others Furthermore Bouman amp Jacobsen (2002)showed that the Halloween effect was economically significant In another studyJacobsen amp Visaltanachoti (2009) studied the Halloween effect in US sectors andfound that in more than two-thirds of these sectors the Halloween effect was sta-tistically significant

2

However Jacobsen amp Visaltanachoti (2009) also found differences across sec-tors with the Halloween effect absent in consumer sectors but present in pro-duction sectors Finally Jacobsen amp Visaltanachoti (2009) showed that thesedifferences across sectors could be used to profit from the Halloween effect usingsector rotation

In another US study Dichtl amp Drobetz (2015) using a bootstrap simulationto avoid possible data snooping biases Dichtl amp Drobetz (2015) found that theHalloween effect has weakened in the US and that inline with Maberly amp Pierce(2004) the Halloween effect may have been driven by extreme monthly returnobservations However Haggard amp Witte (2010) that the Halloween effect in theUS was robust to outliers Furthermore Lucey amp Zhao (2008) found no evidenceof the Halloween effect in the US Lucey amp Zhao (2008) instead found that whatwas seen as the Halloween effect in the US was a reflection of the January effectTherefore Lucey amp Zhao (2008) concluded that the existence of the Halloweeneffect in the US was uncertain

More recently Jacobsen amp Zhang (2018) conducted a worldwide study includ-ing the US of the Halloween effect and found that the Halloween effect was robustOn average across the world returns between November to April were 4 higherthan for the months of May to October Lloyd et al (2017) also found that theHalloween effect was robust in 34 out of 35 countries including the US Arendaset al (2018) focused directly on the US and studied the Halloween effect usingdata on blue-chip stocks from the Dow Jones Industrial Average between 1980and 2007 Arendas et al (2018) found that in 18 out of 35 stocks the Halloweeneffect was statistically significant however the strength of the Halloween effectvaried significantly from stock to stock

Other profitable strategies can coincide or at times be indistinguishable fromthe Halloween effect Afik amp Lahav (2015) showed that although the is significantin the US it can easily be outperformed by just holding a market index from themonths of March to November each year Afik amp Lahav (2015) show that thisstrategy persistent for 43 years on the SampP 500 Index however they concede thatit is difficult to distinguish this effect from the Halloween effect Schabek amp Castro(2017) also showed that strategies starting in October November and Decemberalso generated abnormal returns Schabek amp Castro (2017) also disputed the find-ing of Bouman amp Jacobsen (2002) and showed that the Halloween effect detectedin this study could be subject to sample bias

3

In this paper we therefore revisit the Halloween effect in the US stock marketover its entire history This is in order to avoid concerns about the validity ofsome results in the literature caused by for example noise and selection biasesTo achieve this we conduct average analysis Studentrsquos t-tests Analysis of Vari-ance (ANOVA) tests Mann-Whitney tests and trading simulations on the dataSection 2 describes the data and the empirical methodology and Section 3 presentsthe empirical results Lastly Section 4 offers some concluding remarks

2 Data and Methodology

We use monthly data on various stock markets from the Global Financial Database1That is the US stock market over the period 1791 to 2015 the UK stock marketover the period 1693 to 2017 the Canadian stock market over the period 1915 to2014 the French stock market over the period 1898 to 2018 the German stockmarket over the period 1870 to 2018 the Italian stock market over the period 1905to 2018 the Japanese stock market over the period 1914 to 2013 and the Swissstock market over the period 1916 to 2015 Table 3 in Appendix A summarises thedata To test the robustness of our results the analysis is repeated using this datafor all these countries In order to explore the evolution of the Halloween effect wesplit the overall samples into 25-year sub-periods This explains the different endyears amongst the eight stock markets However the 25-year sub-periods allow forsufficient data for analysis of the dynamics of the evolution of the Halloween effect

To this end the following hypotheses are tested

bull H1 The Halloween effect is not a market myth or legend

bull H2 The Halloween effect evolves over time

bull H3 The Halloween effect can be exploited to get abnormal profits fromtrading in the stock market

We employ average analysis the Studentrsquos t-test ANOVA analysis and theMann-Whitney test The average analysis provides preliminary evidence of thedifference between the months of November to April and the months of May toOctober The parametric and non-parametric tests are carried out on returnswhich were calculated as follows

Rt =

983061ClosetClosetminus1

minus 1

983062times 100 (1)

1The data is available for download at httpswwwglobalfinancialdatacom

4

Halloween Effect in Developed Stock Markets AUS Perspective

Alex Plastunlowast Xolani Sibandedagger Rangan GuptaDagger Mark E Woharsect

February 19 2019

Abstract

In this paper we conduct a comprehensive investigation of the Halloweeneffect evolution in the US stock market over its entire history We employvarious statistical techniques (average analysis Studentrsquos t-test ANOVAand the Mann-Whitney test) and the trading simulation approach to analysethe evolution of the Halloween effect The results suggest that in the USstock market the Halloween effect became more persistent since the middleof the 20th century Despite the decline in its prevalence since that timenowadays it is still present in the US stock market and provides opportunitiesto build a trading strategy which can beat the market These results arewell in line with other developed stock markets Therefore in the main ourresults are inconsistent with the Efficient Market Hypothesis

Keywords Calendar Anomalies Halloween Effect Stock Market Efficient Mar-ket HypothesisJEL Codes G12 C63

lowastFaculty of Economics and Management Sumy State University Sumy Ukraine Emailoplastungmailcom This author gratefully acknowledges financial support from the Ministryof Education and Science of Ukraine (0117U003936)

daggerDepartment of Economics University of Pretoria Pretoria South Africa Emailxolanissgmailcom

DaggerDepartment of Economics University of Pretoria Pretoria South Africa Email ran-ganguptaupacza

sectCorresponding Author College of Business Administration University of Nebraska USSchool of Business and Economics Loughborough University Leicestershire LE11 3TU UKEmail mwoharunomahaedu

1

1 Introduction

Bouman amp Jacobsen (2002) observed that every year around May the Europeanpress would remark that buyers in stock markets seem to have sold in May andgone away This was referring to a phenomenon where the month of May wouldsignal the beginning of a bear market and investors would be better off sellingand holding cash That is until November when the bull market would returnThis phenomenon is known as the Halloween effect

Essentially the Halloween effect indicates that returns between November andApril are higher than in the other months of the year Bouman amp Jacobsen (2002)notes that this difference in returns cannot be explained by factors such as riskvolatility or the January effect amongst others Furthermore these profits wereeconomically exploitable However some in the literature suggest that the Hal-loween effect is a statistical aberration due to for example data outliers (Maberlyamp Pierce (2003) Maberly amp Pierce (2004) Dichtl amp Drobetz (2015) and Dichtlamp Drobetz (2014)) whilst others confirm its continued existence (Jacobsen ampVisaltanachoti (2009) Guo et al (2014) and Lean (2011)) and others suggestthat it has since disappeared (Siriopoulos amp Giannopoulo (2006))

If stock markets follow the Efficient Market Hypothesis (EMH) then marketanomalies (price firm-size and calendar) should not exist at all (Fama (1965)Fama (1970) and Jensen (1978)) In instances where abnormal profits are ex-ploitable their discovery of should result in them disappearing (Schwert (2003))However as noted by Bouman amp Jacobsen (2002) the Halloween effect does notseem to suffer from Murphyrsquos law and has persisted Several reasons have beenadvanced for this in the literature with the most prominent focusing on psycholog-ical and environmental factors underpinning investor behaviour (Kelly amp Meschke(2010) Kamstra et al (2003) Cao amp Wei (2005) and Jacobsen amp Marquering(2008))

Several studies focused on the United States (US) stock market Bouman ampJacobsen (2002) were the first to confirm the existence of the Halloween effect inthe US stock market amongst others Furthermore Bouman amp Jacobsen (2002)showed that the Halloween effect was economically significant In another studyJacobsen amp Visaltanachoti (2009) studied the Halloween effect in US sectors andfound that in more than two-thirds of these sectors the Halloween effect was sta-tistically significant

2

However Jacobsen amp Visaltanachoti (2009) also found differences across sec-tors with the Halloween effect absent in consumer sectors but present in pro-duction sectors Finally Jacobsen amp Visaltanachoti (2009) showed that thesedifferences across sectors could be used to profit from the Halloween effect usingsector rotation

In another US study Dichtl amp Drobetz (2015) using a bootstrap simulationto avoid possible data snooping biases Dichtl amp Drobetz (2015) found that theHalloween effect has weakened in the US and that inline with Maberly amp Pierce(2004) the Halloween effect may have been driven by extreme monthly returnobservations However Haggard amp Witte (2010) that the Halloween effect in theUS was robust to outliers Furthermore Lucey amp Zhao (2008) found no evidenceof the Halloween effect in the US Lucey amp Zhao (2008) instead found that whatwas seen as the Halloween effect in the US was a reflection of the January effectTherefore Lucey amp Zhao (2008) concluded that the existence of the Halloweeneffect in the US was uncertain

More recently Jacobsen amp Zhang (2018) conducted a worldwide study includ-ing the US of the Halloween effect and found that the Halloween effect was robustOn average across the world returns between November to April were 4 higherthan for the months of May to October Lloyd et al (2017) also found that theHalloween effect was robust in 34 out of 35 countries including the US Arendaset al (2018) focused directly on the US and studied the Halloween effect usingdata on blue-chip stocks from the Dow Jones Industrial Average between 1980and 2007 Arendas et al (2018) found that in 18 out of 35 stocks the Halloweeneffect was statistically significant however the strength of the Halloween effectvaried significantly from stock to stock

Other profitable strategies can coincide or at times be indistinguishable fromthe Halloween effect Afik amp Lahav (2015) showed that although the is significantin the US it can easily be outperformed by just holding a market index from themonths of March to November each year Afik amp Lahav (2015) show that thisstrategy persistent for 43 years on the SampP 500 Index however they concede thatit is difficult to distinguish this effect from the Halloween effect Schabek amp Castro(2017) also showed that strategies starting in October November and Decemberalso generated abnormal returns Schabek amp Castro (2017) also disputed the find-ing of Bouman amp Jacobsen (2002) and showed that the Halloween effect detectedin this study could be subject to sample bias

3

In this paper we therefore revisit the Halloween effect in the US stock marketover its entire history This is in order to avoid concerns about the validity ofsome results in the literature caused by for example noise and selection biasesTo achieve this we conduct average analysis Studentrsquos t-tests Analysis of Vari-ance (ANOVA) tests Mann-Whitney tests and trading simulations on the dataSection 2 describes the data and the empirical methodology and Section 3 presentsthe empirical results Lastly Section 4 offers some concluding remarks

2 Data and Methodology

We use monthly data on various stock markets from the Global Financial Database1That is the US stock market over the period 1791 to 2015 the UK stock marketover the period 1693 to 2017 the Canadian stock market over the period 1915 to2014 the French stock market over the period 1898 to 2018 the German stockmarket over the period 1870 to 2018 the Italian stock market over the period 1905to 2018 the Japanese stock market over the period 1914 to 2013 and the Swissstock market over the period 1916 to 2015 Table 3 in Appendix A summarises thedata To test the robustness of our results the analysis is repeated using this datafor all these countries In order to explore the evolution of the Halloween effect wesplit the overall samples into 25-year sub-periods This explains the different endyears amongst the eight stock markets However the 25-year sub-periods allow forsufficient data for analysis of the dynamics of the evolution of the Halloween effect

To this end the following hypotheses are tested

bull H1 The Halloween effect is not a market myth or legend

bull H2 The Halloween effect evolves over time

bull H3 The Halloween effect can be exploited to get abnormal profits fromtrading in the stock market

We employ average analysis the Studentrsquos t-test ANOVA analysis and theMann-Whitney test The average analysis provides preliminary evidence of thedifference between the months of November to April and the months of May toOctober The parametric and non-parametric tests are carried out on returnswhich were calculated as follows

Rt =

983061ClosetClosetminus1

minus 1

983062times 100 (1)

1The data is available for download at httpswwwglobalfinancialdatacom

4

1 Introduction

Bouman amp Jacobsen (2002) observed that every year around May the Europeanpress would remark that buyers in stock markets seem to have sold in May andgone away This was referring to a phenomenon where the month of May wouldsignal the beginning of a bear market and investors would be better off sellingand holding cash That is until November when the bull market would returnThis phenomenon is known as the Halloween effect

Essentially the Halloween effect indicates that returns between November andApril are higher than in the other months of the year Bouman amp Jacobsen (2002)notes that this difference in returns cannot be explained by factors such as riskvolatility or the January effect amongst others Furthermore these profits wereeconomically exploitable However some in the literature suggest that the Hal-loween effect is a statistical aberration due to for example data outliers (Maberlyamp Pierce (2003) Maberly amp Pierce (2004) Dichtl amp Drobetz (2015) and Dichtlamp Drobetz (2014)) whilst others confirm its continued existence (Jacobsen ampVisaltanachoti (2009) Guo et al (2014) and Lean (2011)) and others suggestthat it has since disappeared (Siriopoulos amp Giannopoulo (2006))

If stock markets follow the Efficient Market Hypothesis (EMH) then marketanomalies (price firm-size and calendar) should not exist at all (Fama (1965)Fama (1970) and Jensen (1978)) In instances where abnormal profits are ex-ploitable their discovery of should result in them disappearing (Schwert (2003))However as noted by Bouman amp Jacobsen (2002) the Halloween effect does notseem to suffer from Murphyrsquos law and has persisted Several reasons have beenadvanced for this in the literature with the most prominent focusing on psycholog-ical and environmental factors underpinning investor behaviour (Kelly amp Meschke(2010) Kamstra et al (2003) Cao amp Wei (2005) and Jacobsen amp Marquering(2008))

Several studies focused on the United States (US) stock market Bouman ampJacobsen (2002) were the first to confirm the existence of the Halloween effect inthe US stock market amongst others Furthermore Bouman amp Jacobsen (2002)showed that the Halloween effect was economically significant In another studyJacobsen amp Visaltanachoti (2009) studied the Halloween effect in US sectors andfound that in more than two-thirds of these sectors the Halloween effect was sta-tistically significant

2

However Jacobsen amp Visaltanachoti (2009) also found differences across sec-tors with the Halloween effect absent in consumer sectors but present in pro-duction sectors Finally Jacobsen amp Visaltanachoti (2009) showed that thesedifferences across sectors could be used to profit from the Halloween effect usingsector rotation

In another US study Dichtl amp Drobetz (2015) using a bootstrap simulationto avoid possible data snooping biases Dichtl amp Drobetz (2015) found that theHalloween effect has weakened in the US and that inline with Maberly amp Pierce(2004) the Halloween effect may have been driven by extreme monthly returnobservations However Haggard amp Witte (2010) that the Halloween effect in theUS was robust to outliers Furthermore Lucey amp Zhao (2008) found no evidenceof the Halloween effect in the US Lucey amp Zhao (2008) instead found that whatwas seen as the Halloween effect in the US was a reflection of the January effectTherefore Lucey amp Zhao (2008) concluded that the existence of the Halloweeneffect in the US was uncertain

More recently Jacobsen amp Zhang (2018) conducted a worldwide study includ-ing the US of the Halloween effect and found that the Halloween effect was robustOn average across the world returns between November to April were 4 higherthan for the months of May to October Lloyd et al (2017) also found that theHalloween effect was robust in 34 out of 35 countries including the US Arendaset al (2018) focused directly on the US and studied the Halloween effect usingdata on blue-chip stocks from the Dow Jones Industrial Average between 1980and 2007 Arendas et al (2018) found that in 18 out of 35 stocks the Halloweeneffect was statistically significant however the strength of the Halloween effectvaried significantly from stock to stock

Other profitable strategies can coincide or at times be indistinguishable fromthe Halloween effect Afik amp Lahav (2015) showed that although the is significantin the US it can easily be outperformed by just holding a market index from themonths of March to November each year Afik amp Lahav (2015) show that thisstrategy persistent for 43 years on the SampP 500 Index however they concede thatit is difficult to distinguish this effect from the Halloween effect Schabek amp Castro(2017) also showed that strategies starting in October November and Decemberalso generated abnormal returns Schabek amp Castro (2017) also disputed the find-ing of Bouman amp Jacobsen (2002) and showed that the Halloween effect detectedin this study could be subject to sample bias

3

In this paper we therefore revisit the Halloween effect in the US stock marketover its entire history This is in order to avoid concerns about the validity ofsome results in the literature caused by for example noise and selection biasesTo achieve this we conduct average analysis Studentrsquos t-tests Analysis of Vari-ance (ANOVA) tests Mann-Whitney tests and trading simulations on the dataSection 2 describes the data and the empirical methodology and Section 3 presentsthe empirical results Lastly Section 4 offers some concluding remarks

2 Data and Methodology

We use monthly data on various stock markets from the Global Financial Database1That is the US stock market over the period 1791 to 2015 the UK stock marketover the period 1693 to 2017 the Canadian stock market over the period 1915 to2014 the French stock market over the period 1898 to 2018 the German stockmarket over the period 1870 to 2018 the Italian stock market over the period 1905to 2018 the Japanese stock market over the period 1914 to 2013 and the Swissstock market over the period 1916 to 2015 Table 3 in Appendix A summarises thedata To test the robustness of our results the analysis is repeated using this datafor all these countries In order to explore the evolution of the Halloween effect wesplit the overall samples into 25-year sub-periods This explains the different endyears amongst the eight stock markets However the 25-year sub-periods allow forsufficient data for analysis of the dynamics of the evolution of the Halloween effect

To this end the following hypotheses are tested

bull H1 The Halloween effect is not a market myth or legend

bull H2 The Halloween effect evolves over time

bull H3 The Halloween effect can be exploited to get abnormal profits fromtrading in the stock market

We employ average analysis the Studentrsquos t-test ANOVA analysis and theMann-Whitney test The average analysis provides preliminary evidence of thedifference between the months of November to April and the months of May toOctober The parametric and non-parametric tests are carried out on returnswhich were calculated as follows

Rt =

983061ClosetClosetminus1

minus 1

983062times 100 (1)

1The data is available for download at httpswwwglobalfinancialdatacom

4

However Jacobsen amp Visaltanachoti (2009) also found differences across sec-tors with the Halloween effect absent in consumer sectors but present in pro-duction sectors Finally Jacobsen amp Visaltanachoti (2009) showed that thesedifferences across sectors could be used to profit from the Halloween effect usingsector rotation

In another US study Dichtl amp Drobetz (2015) using a bootstrap simulationto avoid possible data snooping biases Dichtl amp Drobetz (2015) found that theHalloween effect has weakened in the US and that inline with Maberly amp Pierce(2004) the Halloween effect may have been driven by extreme monthly returnobservations However Haggard amp Witte (2010) that the Halloween effect in theUS was robust to outliers Furthermore Lucey amp Zhao (2008) found no evidenceof the Halloween effect in the US Lucey amp Zhao (2008) instead found that whatwas seen as the Halloween effect in the US was a reflection of the January effectTherefore Lucey amp Zhao (2008) concluded that the existence of the Halloweeneffect in the US was uncertain

More recently Jacobsen amp Zhang (2018) conducted a worldwide study includ-ing the US of the Halloween effect and found that the Halloween effect was robustOn average across the world returns between November to April were 4 higherthan for the months of May to October Lloyd et al (2017) also found that theHalloween effect was robust in 34 out of 35 countries including the US Arendaset al (2018) focused directly on the US and studied the Halloween effect usingdata on blue-chip stocks from the Dow Jones Industrial Average between 1980and 2007 Arendas et al (2018) found that in 18 out of 35 stocks the Halloweeneffect was statistically significant however the strength of the Halloween effectvaried significantly from stock to stock

Other profitable strategies can coincide or at times be indistinguishable fromthe Halloween effect Afik amp Lahav (2015) showed that although the is significantin the US it can easily be outperformed by just holding a market index from themonths of March to November each year Afik amp Lahav (2015) show that thisstrategy persistent for 43 years on the SampP 500 Index however they concede thatit is difficult to distinguish this effect from the Halloween effect Schabek amp Castro(2017) also showed that strategies starting in October November and Decemberalso generated abnormal returns Schabek amp Castro (2017) also disputed the find-ing of Bouman amp Jacobsen (2002) and showed that the Halloween effect detectedin this study could be subject to sample bias

3

In this paper we therefore revisit the Halloween effect in the US stock marketover its entire history This is in order to avoid concerns about the validity ofsome results in the literature caused by for example noise and selection biasesTo achieve this we conduct average analysis Studentrsquos t-tests Analysis of Vari-ance (ANOVA) tests Mann-Whitney tests and trading simulations on the dataSection 2 describes the data and the empirical methodology and Section 3 presentsthe empirical results Lastly Section 4 offers some concluding remarks

2 Data and Methodology

We use monthly data on various stock markets from the Global Financial Database1That is the US stock market over the period 1791 to 2015 the UK stock marketover the period 1693 to 2017 the Canadian stock market over the period 1915 to2014 the French stock market over the period 1898 to 2018 the German stockmarket over the period 1870 to 2018 the Italian stock market over the period 1905to 2018 the Japanese stock market over the period 1914 to 2013 and the Swissstock market over the period 1916 to 2015 Table 3 in Appendix A summarises thedata To test the robustness of our results the analysis is repeated using this datafor all these countries In order to explore the evolution of the Halloween effect wesplit the overall samples into 25-year sub-periods This explains the different endyears amongst the eight stock markets However the 25-year sub-periods allow forsufficient data for analysis of the dynamics of the evolution of the Halloween effect

To this end the following hypotheses are tested

bull H1 The Halloween effect is not a market myth or legend

bull H2 The Halloween effect evolves over time

bull H3 The Halloween effect can be exploited to get abnormal profits fromtrading in the stock market

We employ average analysis the Studentrsquos t-test ANOVA analysis and theMann-Whitney test The average analysis provides preliminary evidence of thedifference between the months of November to April and the months of May toOctober The parametric and non-parametric tests are carried out on returnswhich were calculated as follows

Rt =

983061ClosetClosetminus1

minus 1

983062times 100 (1)

1The data is available for download at httpswwwglobalfinancialdatacom

4

In this paper we therefore revisit the Halloween effect in the US stock marketover its entire history This is in order to avoid concerns about the validity ofsome results in the literature caused by for example noise and selection biasesTo achieve this we conduct average analysis Studentrsquos t-tests Analysis of Vari-ance (ANOVA) tests Mann-Whitney tests and trading simulations on the dataSection 2 describes the data and the empirical methodology and Section 3 presentsthe empirical results Lastly Section 4 offers some concluding remarks

2 Data and Methodology

We use monthly data on various stock markets from the Global Financial Database1That is the US stock market over the period 1791 to 2015 the UK stock marketover the period 1693 to 2017 the Canadian stock market over the period 1915 to2014 the French stock market over the period 1898 to 2018 the German stockmarket over the period 1870 to 2018 the Italian stock market over the period 1905to 2018 the Japanese stock market over the period 1914 to 2013 and the Swissstock market over the period 1916 to 2015 Table 3 in Appendix A summarises thedata To test the robustness of our results the analysis is repeated using this datafor all these countries In order to explore the evolution of the Halloween effect wesplit the overall samples into 25-year sub-periods This explains the different endyears amongst the eight stock markets However the 25-year sub-periods allow forsufficient data for analysis of the dynamics of the evolution of the Halloween effect

To this end the following hypotheses are tested

bull H1 The Halloween effect is not a market myth or legend

bull H2 The Halloween effect evolves over time

bull H3 The Halloween effect can be exploited to get abnormal profits fromtrading in the stock market

We employ average analysis the Studentrsquos t-test ANOVA analysis and theMann-Whitney test The average analysis provides preliminary evidence of thedifference between the months of November to April and the months of May toOctober The parametric and non-parametric tests are carried out on returnswhich were calculated as follows

Rt =

983061ClosetClosetminus1

minus 1

983062times 100 (1)

1The data is available for download at httpswwwglobalfinancialdatacom

4

where Rt is the return on the tth day in percentage Closet is the close price onthe tth day and Closetminus1 is the open price on the tminus 1th day

Essentially both the parametric and non-parametric tests test whether re-turns from the months of November to April and returns from the May to Octobermonths are from the same population A rejection of the null hypothesis thereforeindicates the presence of a statistical anomaly in a stock market The Studentrsquos t-tests were conducted using a 95 level of confidence and Nminus12 degrees of freedom

To test whether profits on a Halloween effect based trading strategy were ex-ploitable we use the trading simulation approach The initial step in the tradingsimulation approach is to compute the percentage result (Result) from each dealin the following manner

Result =100times Popen

Pclose

(2)

where Popen is the opening price and Pclose is the closing price

After adding the financial results from all deals a positive financial resultindicates an exploitable stock market anomaly and a negative financial result rep-resents the opposite T-tests were carried out to ensure that financial results arestatistically different from random trading Since the sub-sample sizes are less than100 t-tests where carried out instead of z-tests A rejection of the null hypothesisthat is the mean of the population were profit can be exploited is not differentfrom that of the population were exploitable profits are equal to zero indicatesthat there are advantages to the simulated trading strategy

The results for the US stock market are presented in Appendix B We present asummary of the results for the other developed stock markets in Table 2 Howevera supplementary Appendix C is available with full results for these markets

2N = N1 +N2 in this case

5

3 Results

31 Average analysis

The results of the simple average analysis are displayed in Table 4 and Figure2 Table 4 reveals that the Halloween effect first appeared in the 1841 to 1865sub-period and continued until the 1891 to 1915 sub-period However the differ-ences between the November to April months and the May to October monthsbecame more pronounced in the middle of the 20thth century This continued un-til the last sub-period (1991 to 2015) which had the largest difference of all thesub-periods This difference is clearly demonstrated in Figure 2 which shows thehighest ANOVA multiplier3 in the last sub-period

32 Statistical tests

However the ANOVA analysis (Table 5) the Mann-Whitney test (Table 6) andthe t-tests (Table 7) show that these differences were not statistically significantThis is despite the rejection of the null hypothesis in 1991 to 2015 sub-period usingthe Mann-Whitney test This suggests that the Halloween effect is a myth in theUS stock market

33 Trading strategy

Practically it is interesting to consider how a trading strategy based on the Hal-loween effect would perform compared to a simple buy and hold strategy In Table8 and Figure 3 we show the results of this trading simulation and these revealthat between 1941 and 2015 the trading strategy based on the Halloween effectgenerated stable profits and these profits were statistically different from randomtrading This is confirmed in Figure 3 which shows that both the percentage ofsuccessful trades and the profit percentage increased significantly after 1941 Allother periods failed to generate significant trades based on the Halloween effect

34 Persistence over time

Is the Halloween effect a recent phenomenon or has it persisted in the past Thecombined effect of the average analysis statistical tests and trading simulation

3The ANOVA-multiplier uses the FFCV

ratio to test for the statistical significance of anomalieswhere FCV is the critical value of the F -statistic

6

shows the persistence of the Halloween effect over time Table 1 shows this com-bined effect A higher overall score indicates that the Halloween effect was moreprevalent during that period Clearly Table 1 shows that the Halloween effect wasnot expressly persistence in the US stock market

Table 1 Halloween effect persistence in the US stock market

Period Averageanalysis

Studentst-test

ANOVA Mann-Whitney

test

Tradingsimula-tion

Overall

1791-1815 - - - - - 01816-1840 - - - - - 01841-1865 - - - - - 01866-1890 - - - - - 01891-1915 - - - - - 01916-1940 - - - - - 01941-1965 - - - - + 11966-1990 + - - - + 21991-2015 + - - + + 3

Note + means the anomaly is present and - means that it is not present The Overallcolumn simply counts the number of + with a higher number indicating stronger evidenceof the anomaly

An argument can be made that the Halloween effect is a recent phenomenonin the US stock market The Halloween effect in the US stock market was presentfrom the 1960s Interestingly the Halloween effect only became more and moreprevalent after the 1960s The Halloween effect also allowed for abnormal profitgeneration since the 1960s and as a result was detectable

4 Robustness

The persistence of the Halloween effect in the US stock market is well in line withother developed markets Table 2 shows that the persistence of the Halloweeneffect in the UK Canadian French German Italian Japanese and Swiss stockmarkets In the main the Halloween effect was absent in most of these marketsprior to the middle of the 20th century with the exception of the UK stock marketwhere the Halloween effect briefly appeared between 1843 and 1867 and in theJapanese stock market

7

Similar to the US stock market the persistence of the Halloween effect increasedwith time However in the German and Swiss stock markets the Halloween effectdisappeared in the last sub-period It is not clear if this suggests that the Halloweeneffect has disappeared indefinitely from these markets What is clear is that theHalloween effect is not omnipresent but has been more persistence since the middleof the middle of the 20th century

Figure 1 Halloween effect in developed stock markets

Note The scale is from 0 to 5 where 0 is total absence of anomaly and 5 is the mostconvincing presence of the anomaly

Figure 1 clearly confirms this point The average overall score for all stockmarkets including the US stock market increases substantially after 1917 andthen subsequently begins to slightly decline after 1992 Therefore our main resultis that the US stock market along with other developed markets is not in line withthe EMH due to the prevalence of the Halloween effect

As mentioned earlier clear explanations for the existence of the Halloween effectremain the subject of debate However the fact that the Halloween effect becamemore persistent during the same period in which significant improvements in thenews media available significant improvements in labour reforms- particularly withvacations and in data processing ability must be noted

8

Tab

le2

Persistence

oftheHalloweeneff

ectin

develop

edstock

market

Period

Avera

geanalysis

Stu

dents

t-te

stANOVA

Mann-W

hitney

test

Tra

din

gsimulation

Overa

ll

UK

stock

market

1693-1

717

+-

--

-1

1718-1

742

+-

--

-1

1743-1

767

+-

--

-1

1768-1

792

+-

--

-1

1793-1

817

+-

--

-1

1818-1

842

+-

--

-1

1843-1

867

++

--

+3

1868-1

892

+-

--

-1

1893-1

917

+-

--

-1

1918-1

942

--

--

-1

1943-1

967

++

--

+3

1968-1

992

++

++

+5

1993-2

017

+-

--

+2

Canadia

nstock

market

1915-1

939

+-

--

-1

1940-1

964

+-

--

+2

1965-1

989

++

++

+5

1990-2

014

+-

--

+2

French

stock

market

1870-1

894

+-

--

-1

1895-1

919

+-

--

-1

1920-1

944

+-

--

-1

1945-1

969

+-

-+

+3

1970-1

994

++

+-

+4

1995-2

018

++

+-

+4

Germ

an

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

Italian

stock

market

1870-1

894

+-

--

-1

1895-1

919

+-

--

-1

1920-1

944

+-

--

-1

1945-1

969

+-

-+

+3

1970-1

994

++

+-

+4

1995-2

018

++

+-

+4

Japanese

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

1916-1

940

+-

--

-1

1941-1

965

+-

--

+2

1966-1

990

+-

--

-1

1991-2

015

+-

-+

+3

Swiss

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

Note+

meansthean

omalyis

present

and-meansthatit

isnotpresent

TheOverallcolumnsimply

counts

thenumber

of+

withahigher

number

indicatingstrongereviden

ceof

theanom

aly

9

5 Conclusion

In this paper we examined one of the most recognised calendar anomalies theHalloween effect in the US stock market According to the Halloween effect re-turns between November and April must have significantly stronger average growththan in other months This is impossible according to the EMH To achieve thiswe utilised average analysis the Students t-test and ANOVA the Mann-Whitneytest and the trading simulation approach The results revealed that in the USstock market the Halloween effect first appeared in the middle of the 20th centuryHowever nowadays the Halloween effect is less prevalent but continues to pro-vide opportunities to build a trading strategy which can beat the market Theseresults are well in line with other developed markets Therefore our results areinconsistent with the EMH An interesting observation from this study is that thepersistence of the Halloween effect in the various stock markets is consistent withthe Adaptive Market Hypothesis What is somewhat controversial is that the pe-riod in the middle of the 20th in which the Halloween effect was most persistentis in the literature thought of as the period when stock markets became moreefficient However this can be explained by the fact that unlike other the othercalendar anomalies the Halloween effect does not suffer from Murphyrsquos law

10

References

Afik Z amp Lahav Y (2015) A better autopilot than Sell-in-May 40 years inthe US market Journal of Asset Management 16 (1) 41ndash51

Arendas P Malacka V amp Schwarzova M (2018) A closer look at the Halloweeneffect The case of the Dow Jones Industrial Average International Journal ofFinancial Studies 6 (2) 42

Bouman S amp Jacobsen B (2002) The Halloween indicator rdquoSell in May andgo awayrdquo Another puzzle American Economic Review 92 (5) 1618ndash1635

Cao M amp Wei J (2005) Stock market returns A note on temperature anomalyJournal of Banking amp Finance 29 (6) 1559ndash1573

Dichtl H amp Drobetz W (2014) Are stock markets really so inefficient Thecase of the Halloween Indicator Finance Research Letters 11 (2) 112ndash121

Dichtl H amp Drobetz W (2015) Sell in May and go away Still good advice forinvestors International Review of Financial Analysis 38 29ndash43

Fama E F (1965) The behavior of stock-market prices The Journal of Business 38 (1) 34ndash105

Fama E F (1970) Efficient capital markets A review of theory and empiricalwork The Journal of Finance 25 (2) 383ndash417

Guo B Luo X amp Zhang Z (2014) Sell in May and go away Evidence fromChina Finance Research Letters 11 (4) 362ndash368

Haggard K S amp Witte H D (2010) The Halloween effect Trick or treatInternational Review of Financial Analysis 19 (5) 379 - 387

Jacobsen B amp Marquering W (2008) Is it the weather Journal of Banking ampFinance 32 (4) 526ndash540

Jacobsen B amp Visaltanachoti N (2009) The Halloween effect in US sectorsFinancial Review 44 (3) 437ndash459

Jacobsen B amp Zhang C Y (2018) The halloween indicator everywhere and allthe time Retrieved from httpsssrncomabstract=2154873

Jensen M (1978) Some anomalous evidence regarding market efficiency Journalof Financial Economics 6 (23 (1978)) 95ndash101

11

3 Results

31 Average analysis

The results of the simple average analysis are displayed in Table 4 and Figure2 Table 4 reveals that the Halloween effect first appeared in the 1841 to 1865sub-period and continued until the 1891 to 1915 sub-period However the differ-ences between the November to April months and the May to October monthsbecame more pronounced in the middle of the 20thth century This continued un-til the last sub-period (1991 to 2015) which had the largest difference of all thesub-periods This difference is clearly demonstrated in Figure 2 which shows thehighest ANOVA multiplier3 in the last sub-period

32 Statistical tests

However the ANOVA analysis (Table 5) the Mann-Whitney test (Table 6) andthe t-tests (Table 7) show that these differences were not statistically significantThis is despite the rejection of the null hypothesis in 1991 to 2015 sub-period usingthe Mann-Whitney test This suggests that the Halloween effect is a myth in theUS stock market

33 Trading strategy

Practically it is interesting to consider how a trading strategy based on the Hal-loween effect would perform compared to a simple buy and hold strategy In Table8 and Figure 3 we show the results of this trading simulation and these revealthat between 1941 and 2015 the trading strategy based on the Halloween effectgenerated stable profits and these profits were statistically different from randomtrading This is confirmed in Figure 3 which shows that both the percentage ofsuccessful trades and the profit percentage increased significantly after 1941 Allother periods failed to generate significant trades based on the Halloween effect

34 Persistence over time

Is the Halloween effect a recent phenomenon or has it persisted in the past Thecombined effect of the average analysis statistical tests and trading simulation

3The ANOVA-multiplier uses the FFCV

ratio to test for the statistical significance of anomalieswhere FCV is the critical value of the F -statistic

6

shows the persistence of the Halloween effect over time Table 1 shows this com-bined effect A higher overall score indicates that the Halloween effect was moreprevalent during that period Clearly Table 1 shows that the Halloween effect wasnot expressly persistence in the US stock market

Table 1 Halloween effect persistence in the US stock market

Period Averageanalysis

Studentst-test

ANOVA Mann-Whitney

test

Tradingsimula-tion

Overall

1791-1815 - - - - - 01816-1840 - - - - - 01841-1865 - - - - - 01866-1890 - - - - - 01891-1915 - - - - - 01916-1940 - - - - - 01941-1965 - - - - + 11966-1990 + - - - + 21991-2015 + - - + + 3

Note + means the anomaly is present and - means that it is not present The Overallcolumn simply counts the number of + with a higher number indicating stronger evidenceof the anomaly

An argument can be made that the Halloween effect is a recent phenomenonin the US stock market The Halloween effect in the US stock market was presentfrom the 1960s Interestingly the Halloween effect only became more and moreprevalent after the 1960s The Halloween effect also allowed for abnormal profitgeneration since the 1960s and as a result was detectable

4 Robustness

The persistence of the Halloween effect in the US stock market is well in line withother developed markets Table 2 shows that the persistence of the Halloweeneffect in the UK Canadian French German Italian Japanese and Swiss stockmarkets In the main the Halloween effect was absent in most of these marketsprior to the middle of the 20th century with the exception of the UK stock marketwhere the Halloween effect briefly appeared between 1843 and 1867 and in theJapanese stock market

7

Similar to the US stock market the persistence of the Halloween effect increasedwith time However in the German and Swiss stock markets the Halloween effectdisappeared in the last sub-period It is not clear if this suggests that the Halloweeneffect has disappeared indefinitely from these markets What is clear is that theHalloween effect is not omnipresent but has been more persistence since the middleof the middle of the 20th century

Figure 1 Halloween effect in developed stock markets

Note The scale is from 0 to 5 where 0 is total absence of anomaly and 5 is the mostconvincing presence of the anomaly

Figure 1 clearly confirms this point The average overall score for all stockmarkets including the US stock market increases substantially after 1917 andthen subsequently begins to slightly decline after 1992 Therefore our main resultis that the US stock market along with other developed markets is not in line withthe EMH due to the prevalence of the Halloween effect

As mentioned earlier clear explanations for the existence of the Halloween effectremain the subject of debate However the fact that the Halloween effect becamemore persistent during the same period in which significant improvements in thenews media available significant improvements in labour reforms- particularly withvacations and in data processing ability must be noted

8

Tab

le2

Persistence

oftheHalloweeneff

ectin

develop

edstock

market

Period

Avera

geanalysis

Stu

dents

t-te

stANOVA

Mann-W

hitney

test

Tra

din

gsimulation

Overa

ll

UK

stock

market

1693-1

717

+-

--

-1

1718-1

742

+-

--

-1

1743-1

767

+-

--

-1

1768-1

792

+-

--

-1

1793-1

817

+-

--

-1

1818-1

842

+-

--

-1

1843-1

867

++

--

+3

1868-1

892

+-

--

-1

1893-1

917

+-

--

-1

1918-1

942

--

--

-1

1943-1

967

++

--

+3

1968-1

992

++

++

+5

1993-2

017

+-

--

+2

Canadia

nstock

market

1915-1

939

+-

--

-1

1940-1

964

+-

--

+2

1965-1

989

++

++

+5

1990-2

014

+-

--

+2

French

stock

market

1870-1

894

+-

--

-1

1895-1

919

+-

--

-1

1920-1

944

+-

--

-1

1945-1

969

+-

-+

+3

1970-1

994

++

+-

+4

1995-2

018

++

+-

+4

Germ

an

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

Italian

stock

market

1870-1

894

+-

--

-1

1895-1

919

+-

--

-1

1920-1

944

+-

--

-1

1945-1

969

+-

-+

+3

1970-1

994

++

+-

+4

1995-2

018

++

+-

+4

Japanese

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

1916-1

940

+-

--

-1

1941-1

965

+-

--

+2

1966-1

990

+-

--

-1

1991-2

015

+-

-+

+3

Swiss

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

Note+

meansthean

omalyis

present

and-meansthatit

isnotpresent

TheOverallcolumnsimply

counts

thenumber

of+

withahigher

number

indicatingstrongereviden

ceof

theanom

aly

9

5 Conclusion

In this paper we examined one of the most recognised calendar anomalies theHalloween effect in the US stock market According to the Halloween effect re-turns between November and April must have significantly stronger average growththan in other months This is impossible according to the EMH To achieve thiswe utilised average analysis the Students t-test and ANOVA the Mann-Whitneytest and the trading simulation approach The results revealed that in the USstock market the Halloween effect first appeared in the middle of the 20th centuryHowever nowadays the Halloween effect is less prevalent but continues to pro-vide opportunities to build a trading strategy which can beat the market Theseresults are well in line with other developed markets Therefore our results areinconsistent with the EMH An interesting observation from this study is that thepersistence of the Halloween effect in the various stock markets is consistent withthe Adaptive Market Hypothesis What is somewhat controversial is that the pe-riod in the middle of the 20th in which the Halloween effect was most persistentis in the literature thought of as the period when stock markets became moreefficient However this can be explained by the fact that unlike other the othercalendar anomalies the Halloween effect does not suffer from Murphyrsquos law

10

References

Afik Z amp Lahav Y (2015) A better autopilot than Sell-in-May 40 years inthe US market Journal of Asset Management 16 (1) 41ndash51

Arendas P Malacka V amp Schwarzova M (2018) A closer look at the Halloweeneffect The case of the Dow Jones Industrial Average International Journal ofFinancial Studies 6 (2) 42

Bouman S amp Jacobsen B (2002) The Halloween indicator rdquoSell in May andgo awayrdquo Another puzzle American Economic Review 92 (5) 1618ndash1635

Cao M amp Wei J (2005) Stock market returns A note on temperature anomalyJournal of Banking amp Finance 29 (6) 1559ndash1573

Dichtl H amp Drobetz W (2014) Are stock markets really so inefficient Thecase of the Halloween Indicator Finance Research Letters 11 (2) 112ndash121

Dichtl H amp Drobetz W (2015) Sell in May and go away Still good advice forinvestors International Review of Financial Analysis 38 29ndash43

Fama E F (1965) The behavior of stock-market prices The Journal of Business 38 (1) 34ndash105

Fama E F (1970) Efficient capital markets A review of theory and empiricalwork The Journal of Finance 25 (2) 383ndash417

Guo B Luo X amp Zhang Z (2014) Sell in May and go away Evidence fromChina Finance Research Letters 11 (4) 362ndash368

Haggard K S amp Witte H D (2010) The Halloween effect Trick or treatInternational Review of Financial Analysis 19 (5) 379 - 387

Jacobsen B amp Marquering W (2008) Is it the weather Journal of Banking ampFinance 32 (4) 526ndash540

Jacobsen B amp Visaltanachoti N (2009) The Halloween effect in US sectorsFinancial Review 44 (3) 437ndash459

Jacobsen B amp Zhang C Y (2018) The halloween indicator everywhere and allthe time Retrieved from httpsssrncomabstract=2154873

Jensen M (1978) Some anomalous evidence regarding market efficiency Journalof Financial Economics 6 (23 (1978)) 95ndash101

11

shows the persistence of the Halloween effect over time Table 1 shows this com-bined effect A higher overall score indicates that the Halloween effect was moreprevalent during that period Clearly Table 1 shows that the Halloween effect wasnot expressly persistence in the US stock market

Table 1 Halloween effect persistence in the US stock market

Period Averageanalysis

Studentst-test

ANOVA Mann-Whitney

test

Tradingsimula-tion

Overall

1791-1815 - - - - - 01816-1840 - - - - - 01841-1865 - - - - - 01866-1890 - - - - - 01891-1915 - - - - - 01916-1940 - - - - - 01941-1965 - - - - + 11966-1990 + - - - + 21991-2015 + - - + + 3

Note + means the anomaly is present and - means that it is not present The Overallcolumn simply counts the number of + with a higher number indicating stronger evidenceof the anomaly

An argument can be made that the Halloween effect is a recent phenomenonin the US stock market The Halloween effect in the US stock market was presentfrom the 1960s Interestingly the Halloween effect only became more and moreprevalent after the 1960s The Halloween effect also allowed for abnormal profitgeneration since the 1960s and as a result was detectable

4 Robustness

The persistence of the Halloween effect in the US stock market is well in line withother developed markets Table 2 shows that the persistence of the Halloweeneffect in the UK Canadian French German Italian Japanese and Swiss stockmarkets In the main the Halloween effect was absent in most of these marketsprior to the middle of the 20th century with the exception of the UK stock marketwhere the Halloween effect briefly appeared between 1843 and 1867 and in theJapanese stock market

7

Similar to the US stock market the persistence of the Halloween effect increasedwith time However in the German and Swiss stock markets the Halloween effectdisappeared in the last sub-period It is not clear if this suggests that the Halloweeneffect has disappeared indefinitely from these markets What is clear is that theHalloween effect is not omnipresent but has been more persistence since the middleof the middle of the 20th century

Figure 1 Halloween effect in developed stock markets

Note The scale is from 0 to 5 where 0 is total absence of anomaly and 5 is the mostconvincing presence of the anomaly

Figure 1 clearly confirms this point The average overall score for all stockmarkets including the US stock market increases substantially after 1917 andthen subsequently begins to slightly decline after 1992 Therefore our main resultis that the US stock market along with other developed markets is not in line withthe EMH due to the prevalence of the Halloween effect

As mentioned earlier clear explanations for the existence of the Halloween effectremain the subject of debate However the fact that the Halloween effect becamemore persistent during the same period in which significant improvements in thenews media available significant improvements in labour reforms- particularly withvacations and in data processing ability must be noted

8

Tab

le2

Persistence

oftheHalloweeneff

ectin

develop

edstock

market

Period

Avera

geanalysis

Stu

dents

t-te

stANOVA

Mann-W

hitney

test

Tra

din

gsimulation

Overa

ll

UK

stock

market

1693-1

717

+-

--

-1

1718-1

742

+-

--

-1

1743-1

767

+-

--

-1

1768-1

792

+-

--

-1

1793-1

817

+-

--

-1

1818-1

842

+-

--

-1

1843-1

867

++

--

+3

1868-1

892

+-

--

-1

1893-1

917

+-

--

-1

1918-1

942

--

--

-1

1943-1

967

++

--

+3

1968-1

992

++

++

+5

1993-2

017

+-

--

+2

Canadia

nstock

market

1915-1

939

+-

--

-1

1940-1

964

+-

--

+2

1965-1

989

++

++

+5

1990-2

014

+-

--

+2

French

stock

market

1870-1

894

+-

--

-1

1895-1

919

+-

--

-1

1920-1

944

+-

--

-1

1945-1

969

+-

-+

+3

1970-1

994

++

+-

+4

1995-2

018

++

+-

+4

Germ

an

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

Italian

stock

market

1870-1

894

+-

--

-1

1895-1

919

+-

--

-1

1920-1

944

+-

--

-1

1945-1

969

+-

-+

+3

1970-1

994

++

+-

+4

1995-2

018

++

+-

+4

Japanese

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

1916-1

940

+-

--

-1

1941-1

965

+-

--

+2

1966-1

990

+-

--

-1

1991-2

015

+-

-+

+3

Swiss

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

Note+

meansthean

omalyis

present

and-meansthatit

isnotpresent

TheOverallcolumnsimply

counts

thenumber

of+

withahigher

number

indicatingstrongereviden

ceof

theanom

aly

9

5 Conclusion

In this paper we examined one of the most recognised calendar anomalies theHalloween effect in the US stock market According to the Halloween effect re-turns between November and April must have significantly stronger average growththan in other months This is impossible according to the EMH To achieve thiswe utilised average analysis the Students t-test and ANOVA the Mann-Whitneytest and the trading simulation approach The results revealed that in the USstock market the Halloween effect first appeared in the middle of the 20th centuryHowever nowadays the Halloween effect is less prevalent but continues to pro-vide opportunities to build a trading strategy which can beat the market Theseresults are well in line with other developed markets Therefore our results areinconsistent with the EMH An interesting observation from this study is that thepersistence of the Halloween effect in the various stock markets is consistent withthe Adaptive Market Hypothesis What is somewhat controversial is that the pe-riod in the middle of the 20th in which the Halloween effect was most persistentis in the literature thought of as the period when stock markets became moreefficient However this can be explained by the fact that unlike other the othercalendar anomalies the Halloween effect does not suffer from Murphyrsquos law

10

References

Afik Z amp Lahav Y (2015) A better autopilot than Sell-in-May 40 years inthe US market Journal of Asset Management 16 (1) 41ndash51

Arendas P Malacka V amp Schwarzova M (2018) A closer look at the Halloweeneffect The case of the Dow Jones Industrial Average International Journal ofFinancial Studies 6 (2) 42

Bouman S amp Jacobsen B (2002) The Halloween indicator rdquoSell in May andgo awayrdquo Another puzzle American Economic Review 92 (5) 1618ndash1635

Cao M amp Wei J (2005) Stock market returns A note on temperature anomalyJournal of Banking amp Finance 29 (6) 1559ndash1573

Dichtl H amp Drobetz W (2014) Are stock markets really so inefficient Thecase of the Halloween Indicator Finance Research Letters 11 (2) 112ndash121

Dichtl H amp Drobetz W (2015) Sell in May and go away Still good advice forinvestors International Review of Financial Analysis 38 29ndash43

Fama E F (1965) The behavior of stock-market prices The Journal of Business 38 (1) 34ndash105

Fama E F (1970) Efficient capital markets A review of theory and empiricalwork The Journal of Finance 25 (2) 383ndash417

Guo B Luo X amp Zhang Z (2014) Sell in May and go away Evidence fromChina Finance Research Letters 11 (4) 362ndash368

Haggard K S amp Witte H D (2010) The Halloween effect Trick or treatInternational Review of Financial Analysis 19 (5) 379 - 387

Jacobsen B amp Marquering W (2008) Is it the weather Journal of Banking ampFinance 32 (4) 526ndash540

Jacobsen B amp Visaltanachoti N (2009) The Halloween effect in US sectorsFinancial Review 44 (3) 437ndash459

Jacobsen B amp Zhang C Y (2018) The halloween indicator everywhere and allthe time Retrieved from httpsssrncomabstract=2154873

Jensen M (1978) Some anomalous evidence regarding market efficiency Journalof Financial Economics 6 (23 (1978)) 95ndash101

11

Similar to the US stock market the persistence of the Halloween effect increasedwith time However in the German and Swiss stock markets the Halloween effectdisappeared in the last sub-period It is not clear if this suggests that the Halloweeneffect has disappeared indefinitely from these markets What is clear is that theHalloween effect is not omnipresent but has been more persistence since the middleof the middle of the 20th century

Figure 1 Halloween effect in developed stock markets

Note The scale is from 0 to 5 where 0 is total absence of anomaly and 5 is the mostconvincing presence of the anomaly

Figure 1 clearly confirms this point The average overall score for all stockmarkets including the US stock market increases substantially after 1917 andthen subsequently begins to slightly decline after 1992 Therefore our main resultis that the US stock market along with other developed markets is not in line withthe EMH due to the prevalence of the Halloween effect

As mentioned earlier clear explanations for the existence of the Halloween effectremain the subject of debate However the fact that the Halloween effect becamemore persistent during the same period in which significant improvements in thenews media available significant improvements in labour reforms- particularly withvacations and in data processing ability must be noted

8

Tab

le2

Persistence

oftheHalloweeneff

ectin

develop

edstock

market

Period

Avera

geanalysis

Stu

dents

t-te

stANOVA

Mann-W

hitney

test

Tra

din

gsimulation

Overa

ll

UK

stock

market

1693-1

717

+-

--

-1

1718-1

742

+-

--

-1

1743-1

767

+-

--

-1

1768-1

792

+-

--

-1

1793-1

817

+-

--

-1

1818-1

842

+-

--

-1

1843-1

867

++

--

+3

1868-1

892

+-

--

-1

1893-1

917

+-

--

-1

1918-1

942

--

--

-1

1943-1

967

++

--

+3

1968-1

992

++

++

+5

1993-2

017

+-

--

+2

Canadia

nstock

market

1915-1

939

+-

--

-1

1940-1

964

+-

--

+2

1965-1

989

++

++

+5

1990-2

014

+-

--

+2

French

stock

market

1870-1

894

+-

--

-1

1895-1

919

+-

--

-1

1920-1

944

+-

--

-1

1945-1

969

+-

-+

+3

1970-1

994

++

+-

+4

1995-2

018

++

+-

+4

Germ

an

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

Italian

stock

market

1870-1

894

+-

--

-1

1895-1

919

+-

--

-1

1920-1

944

+-

--

-1

1945-1

969

+-

-+

+3

1970-1

994

++

+-

+4

1995-2

018

++

+-

+4

Japanese

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

1916-1

940

+-

--

-1

1941-1

965

+-

--

+2

1966-1

990

+-

--

-1

1991-2

015

+-

-+

+3

Swiss

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

Note+

meansthean

omalyis

present

and-meansthatit

isnotpresent

TheOverallcolumnsimply

counts

thenumber

of+

withahigher

number

indicatingstrongereviden

ceof

theanom

aly

9

5 Conclusion

In this paper we examined one of the most recognised calendar anomalies theHalloween effect in the US stock market According to the Halloween effect re-turns between November and April must have significantly stronger average growththan in other months This is impossible according to the EMH To achieve thiswe utilised average analysis the Students t-test and ANOVA the Mann-Whitneytest and the trading simulation approach The results revealed that in the USstock market the Halloween effect first appeared in the middle of the 20th centuryHowever nowadays the Halloween effect is less prevalent but continues to pro-vide opportunities to build a trading strategy which can beat the market Theseresults are well in line with other developed markets Therefore our results areinconsistent with the EMH An interesting observation from this study is that thepersistence of the Halloween effect in the various stock markets is consistent withthe Adaptive Market Hypothesis What is somewhat controversial is that the pe-riod in the middle of the 20th in which the Halloween effect was most persistentis in the literature thought of as the period when stock markets became moreefficient However this can be explained by the fact that unlike other the othercalendar anomalies the Halloween effect does not suffer from Murphyrsquos law

10

References

Afik Z amp Lahav Y (2015) A better autopilot than Sell-in-May 40 years inthe US market Journal of Asset Management 16 (1) 41ndash51

Arendas P Malacka V amp Schwarzova M (2018) A closer look at the Halloweeneffect The case of the Dow Jones Industrial Average International Journal ofFinancial Studies 6 (2) 42

Bouman S amp Jacobsen B (2002) The Halloween indicator rdquoSell in May andgo awayrdquo Another puzzle American Economic Review 92 (5) 1618ndash1635

Cao M amp Wei J (2005) Stock market returns A note on temperature anomalyJournal of Banking amp Finance 29 (6) 1559ndash1573

Dichtl H amp Drobetz W (2014) Are stock markets really so inefficient Thecase of the Halloween Indicator Finance Research Letters 11 (2) 112ndash121

Dichtl H amp Drobetz W (2015) Sell in May and go away Still good advice forinvestors International Review of Financial Analysis 38 29ndash43

Fama E F (1965) The behavior of stock-market prices The Journal of Business 38 (1) 34ndash105

Fama E F (1970) Efficient capital markets A review of theory and empiricalwork The Journal of Finance 25 (2) 383ndash417

Guo B Luo X amp Zhang Z (2014) Sell in May and go away Evidence fromChina Finance Research Letters 11 (4) 362ndash368

Haggard K S amp Witte H D (2010) The Halloween effect Trick or treatInternational Review of Financial Analysis 19 (5) 379 - 387

Jacobsen B amp Marquering W (2008) Is it the weather Journal of Banking ampFinance 32 (4) 526ndash540

Jacobsen B amp Visaltanachoti N (2009) The Halloween effect in US sectorsFinancial Review 44 (3) 437ndash459

Jacobsen B amp Zhang C Y (2018) The halloween indicator everywhere and allthe time Retrieved from httpsssrncomabstract=2154873

Jensen M (1978) Some anomalous evidence regarding market efficiency Journalof Financial Economics 6 (23 (1978)) 95ndash101

11

Tab

le2

Persistence

oftheHalloweeneff

ectin

develop

edstock

market

Period

Avera

geanalysis

Stu

dents

t-te

stANOVA

Mann-W

hitney

test

Tra

din

gsimulation

Overa

ll

UK

stock

market

1693-1

717

+-

--

-1

1718-1

742

+-

--

-1

1743-1

767

+-

--

-1

1768-1

792

+-

--

-1

1793-1

817

+-

--

-1

1818-1

842

+-

--

-1

1843-1

867

++

--

+3

1868-1

892

+-

--

-1

1893-1

917

+-

--

-1

1918-1

942

--

--

-1

1943-1

967

++

--

+3

1968-1

992

++

++

+5

1993-2

017

+-

--

+2

Canadia

nstock

market

1915-1

939

+-

--

-1

1940-1

964

+-

--

+2

1965-1

989

++

++

+5

1990-2

014

+-

--

+2

French

stock

market

1870-1

894

+-

--

-1

1895-1

919

+-

--

-1

1920-1

944

+-

--

-1

1945-1

969

+-

-+

+3

1970-1

994

++

+-

+4

1995-2

018

++

+-

+4

Germ

an

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

Italian

stock

market

1870-1

894

+-

--

-1

1895-1

919

+-

--

-1

1920-1

944

+-

--

-1

1945-1

969

+-

-+

+3

1970-1

994

++

+-

+4

1995-2

018

++

+-

+4

Japanese

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

1916-1

940

+-

--

-1

1941-1

965

+-

--

+2

1966-1

990

+-

--

-1

1991-2

015

+-

-+

+3

Swiss

stock

market

1905-1

929

+-

--

-1

1930-1

954

--

--

-0

1955-1

979

+-

--

+2

1980-2

004

++

++

+5

2005-2

018

+-

--

-1

Note+

meansthean

omalyis

present

and-meansthatit

isnotpresent

TheOverallcolumnsimply

counts

thenumber

of+

withahigher

number

indicatingstrongereviden

ceof

theanom

aly

9

5 Conclusion

In this paper we examined one of the most recognised calendar anomalies theHalloween effect in the US stock market According to the Halloween effect re-turns between November and April must have significantly stronger average growththan in other months This is impossible according to the EMH To achieve thiswe utilised average analysis the Students t-test and ANOVA the Mann-Whitneytest and the trading simulation approach The results revealed that in the USstock market the Halloween effect first appeared in the middle of the 20th centuryHowever nowadays the Halloween effect is less prevalent but continues to pro-vide opportunities to build a trading strategy which can beat the market Theseresults are well in line with other developed markets Therefore our results areinconsistent with the EMH An interesting observation from this study is that thepersistence of the Halloween effect in the various stock markets is consistent withthe Adaptive Market Hypothesis What is somewhat controversial is that the pe-riod in the middle of the 20th in which the Halloween effect was most persistentis in the literature thought of as the period when stock markets became moreefficient However this can be explained by the fact that unlike other the othercalendar anomalies the Halloween effect does not suffer from Murphyrsquos law

10

References

Afik Z amp Lahav Y (2015) A better autopilot than Sell-in-May 40 years inthe US market Journal of Asset Management 16 (1) 41ndash51

Arendas P Malacka V amp Schwarzova M (2018) A closer look at the Halloweeneffect The case of the Dow Jones Industrial Average International Journal ofFinancial Studies 6 (2) 42

Bouman S amp Jacobsen B (2002) The Halloween indicator rdquoSell in May andgo awayrdquo Another puzzle American Economic Review 92 (5) 1618ndash1635

Cao M amp Wei J (2005) Stock market returns A note on temperature anomalyJournal of Banking amp Finance 29 (6) 1559ndash1573

Dichtl H amp Drobetz W (2014) Are stock markets really so inefficient Thecase of the Halloween Indicator Finance Research Letters 11 (2) 112ndash121

Dichtl H amp Drobetz W (2015) Sell in May and go away Still good advice forinvestors International Review of Financial Analysis 38 29ndash43

Fama E F (1965) The behavior of stock-market prices The Journal of Business 38 (1) 34ndash105

Fama E F (1970) Efficient capital markets A review of theory and empiricalwork The Journal of Finance 25 (2) 383ndash417

Guo B Luo X amp Zhang Z (2014) Sell in May and go away Evidence fromChina Finance Research Letters 11 (4) 362ndash368

Haggard K S amp Witte H D (2010) The Halloween effect Trick or treatInternational Review of Financial Analysis 19 (5) 379 - 387

Jacobsen B amp Marquering W (2008) Is it the weather Journal of Banking ampFinance 32 (4) 526ndash540

Jacobsen B amp Visaltanachoti N (2009) The Halloween effect in US sectorsFinancial Review 44 (3) 437ndash459

Jacobsen B amp Zhang C Y (2018) The halloween indicator everywhere and allthe time Retrieved from httpsssrncomabstract=2154873

Jensen M (1978) Some anomalous evidence regarding market efficiency Journalof Financial Economics 6 (23 (1978)) 95ndash101

11

5 Conclusion

In this paper we examined one of the most recognised calendar anomalies theHalloween effect in the US stock market According to the Halloween effect re-turns between November and April must have significantly stronger average growththan in other months This is impossible according to the EMH To achieve thiswe utilised average analysis the Students t-test and ANOVA the Mann-Whitneytest and the trading simulation approach The results revealed that in the USstock market the Halloween effect first appeared in the middle of the 20th centuryHowever nowadays the Halloween effect is less prevalent but continues to pro-vide opportunities to build a trading strategy which can beat the market Theseresults are well in line with other developed markets Therefore our results areinconsistent with the EMH An interesting observation from this study is that thepersistence of the Halloween effect in the various stock markets is consistent withthe Adaptive Market Hypothesis What is somewhat controversial is that the pe-riod in the middle of the 20th in which the Halloween effect was most persistentis in the literature thought of as the period when stock markets became moreefficient However this can be explained by the fact that unlike other the othercalendar anomalies the Halloween effect does not suffer from Murphyrsquos law

10

References

Afik Z amp Lahav Y (2015) A better autopilot than Sell-in-May 40 years inthe US market Journal of Asset Management 16 (1) 41ndash51

Arendas P Malacka V amp Schwarzova M (2018) A closer look at the Halloweeneffect The case of the Dow Jones Industrial Average International Journal ofFinancial Studies 6 (2) 42

Bouman S amp Jacobsen B (2002) The Halloween indicator rdquoSell in May andgo awayrdquo Another puzzle American Economic Review 92 (5) 1618ndash1635

Cao M amp Wei J (2005) Stock market returns A note on temperature anomalyJournal of Banking amp Finance 29 (6) 1559ndash1573

Dichtl H amp Drobetz W (2014) Are stock markets really so inefficient Thecase of the Halloween Indicator Finance Research Letters 11 (2) 112ndash121

Dichtl H amp Drobetz W (2015) Sell in May and go away Still good advice forinvestors International Review of Financial Analysis 38 29ndash43

Fama E F (1965) The behavior of stock-market prices The Journal of Business 38 (1) 34ndash105

Fama E F (1970) Efficient capital markets A review of theory and empiricalwork The Journal of Finance 25 (2) 383ndash417

Guo B Luo X amp Zhang Z (2014) Sell in May and go away Evidence fromChina Finance Research Letters 11 (4) 362ndash368

Haggard K S amp Witte H D (2010) The Halloween effect Trick or treatInternational Review of Financial Analysis 19 (5) 379 - 387

Jacobsen B amp Marquering W (2008) Is it the weather Journal of Banking ampFinance 32 (4) 526ndash540

Jacobsen B amp Visaltanachoti N (2009) The Halloween effect in US sectorsFinancial Review 44 (3) 437ndash459

Jacobsen B amp Zhang C Y (2018) The halloween indicator everywhere and allthe time Retrieved from httpsssrncomabstract=2154873

Jensen M (1978) Some anomalous evidence regarding market efficiency Journalof Financial Economics 6 (23 (1978)) 95ndash101

11

References

Afik Z amp Lahav Y (2015) A better autopilot than Sell-in-May 40 years inthe US market Journal of Asset Management 16 (1) 41ndash51

Arendas P Malacka V amp Schwarzova M (2018) A closer look at the Halloweeneffect The case of the Dow Jones Industrial Average International Journal ofFinancial Studies 6 (2) 42

Bouman S amp Jacobsen B (2002) The Halloween indicator rdquoSell in May andgo awayrdquo Another puzzle American Economic Review 92 (5) 1618ndash1635

Cao M amp Wei J (2005) Stock market returns A note on temperature anomalyJournal of Banking amp Finance 29 (6) 1559ndash1573

Dichtl H amp Drobetz W (2014) Are stock markets really so inefficient Thecase of the Halloween Indicator Finance Research Letters 11 (2) 112ndash121

Dichtl H amp Drobetz W (2015) Sell in May and go away Still good advice forinvestors International Review of Financial Analysis 38 29ndash43

Fama E F (1965) The behavior of stock-market prices The Journal of Business 38 (1) 34ndash105

Fama E F (1970) Efficient capital markets A review of theory and empiricalwork The Journal of Finance 25 (2) 383ndash417

Guo B Luo X amp Zhang Z (2014) Sell in May and go away Evidence fromChina Finance Research Letters 11 (4) 362ndash368

Haggard K S amp Witte H D (2010) The Halloween effect Trick or treatInternational Review of Financial Analysis 19 (5) 379 - 387

Jacobsen B amp Marquering W (2008) Is it the weather Journal of Banking ampFinance 32 (4) 526ndash540

Jacobsen B amp Visaltanachoti N (2009) The Halloween effect in US sectorsFinancial Review 44 (3) 437ndash459

Jacobsen B amp Zhang C Y (2018) The halloween indicator everywhere and allthe time Retrieved from httpsssrncomabstract=2154873

Jensen M (1978) Some anomalous evidence regarding market efficiency Journalof Financial Economics 6 (23 (1978)) 95ndash101

11

Kamstra M Kramer L amp Levi M (2003) Winter Blues A SAD stock marketcycle American Economic Review 93 (1) 324ndash343

Kelly P amp Meschke F (2010) Sentiment and stock returns The SAD anomalyrevisited Journal of Banking amp Finance 34 (6) 1308ndash1326

Lean H H (2011) The Halloween puzzle in selected Asian stock markets Inter-national Journal of Economics and Management 5 (1) 216ndash225

Lloyd R Zhang C amp Rydin S (2017) The Halloween Indicator is more a treatthan a trick Journal of Accounting and Finance 17 (6) 96ndash108

Lucey B M amp Zhao S (2008) Halloween or January Yet another puzzleInternational Review of Financial Analysis 17 (5) 1055ndash1069

Maberly E D amp Pierce R M (2003) The Halloween effect and Japanese equityprices Myth or exploitable anomaly Asia-Pacific Financial Markets 10 (4)319ndash334

Maberly E D amp Pierce R M (2004) Stock market efficiency withstands anotherchallenge Solving the rdquosell in May buy after Halloweenrdquo puzzle Econ JournalWatch 1 (1) 29

Schabek T amp Castro H (2017) Sell not only in May Seasonal effect on emergingand developed stock markets Dynamic Econometric Models 17 (1) 5ndash18

Schwert G W (2003) Chapter 15 anomalies and market efficiency In Financialmarkets and asset pricing (Vol 1 p 939 - 974) Elsevier

Siriopoulos C amp Giannopoulo P (2006) Market efficiency in the Greek stockexchange the Halloween effect Journal of Economics and Business 56 (2)75ndash88

12

Appendices

A Data

Table 3 Data

Country Stock Market Sample PeriodUS SampP 500 Index 1791-2015UK FTSE All Share Index 1693-2017Canada SampPTSX 300 Composite

Index1898-2018

France CAC All Tradable Index 1870-2018Germany CDAX Composite Index 1870-2018Italy Banca Commerciale Italiana

Index1905-2018

Japan Nikkei 225 1914-2013Switzerland Switzerland Composite Stock

Price Index1916-2015

B Halloween Effect in the US Stock Market

Table 4 Average returns in the US stock market

Period November to April May to October1791-1815 -041 -0041816-1840 -049 0321841-1865 214 -0491866-1890 056 0431891-1915 271 -0031916-1940 -175 2221941-1965 492 3221966-1990 440 0341991-2015 607 074

13

Figure 2 Average returns in the US stock market

Table 5 ANOVA test of the Halloween effect in the US stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1791-1815 003 087 405 not rejected not confirmed 0011816-1840 016 069 404 not rejected not confirmed 0041841-1865 042 052 404 not rejected not confirmed 0101866-1890 005 083 404 not rejected not confirmed 0011891-1915 089 035 404 not rejected not confirmed 0221916-1940 066 042 404 not rejected not confirmed 0161941-1965 046 050 404 not rejected not confirmed 0111966-1990 186 018 404 not rejected not confirmed 0461991-2015 333 007 404 not rejected not confirmed 082

14

Table 6 Mann-Whitney test of the Halloween effect in the US stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1791-1815 065 1 042 384 not rejected not confirmed 0171816-1840 104 1 031 384 not rejected not confirmed 0271841-1865 020 1 066 384 not rejected not confirmed 0051866-1890 000 1 095 384 not rejected not confirmed 0001891-1915 049 1 049 384 not rejected not confirmed 0131916-1940 132 1 025 384 not rejected not confirmed 0341941-1965 014 1 071 384 not rejected not confirmed 0041966-1990 111 1 029 384 not rejected not confirmed 0291991-2015 470 1 003 384 rejected confirmed 122

15

Table 7 T-test of the Halloween effect in the US stock market

Period Parameter November to April May to October1791-1815 Mean -041 -004

t-criterion -029Null hypothesis not rejected

Anomaly not detected1816-1840 Mean -049 032

t-criterion -046Null hypothesis not rejected

Anomaly not detected1841-1865 Mean 214 -049

t-criterion 061Null hypothesis not rejected

Anomaly not detected1866-1890 Mean 056 043

t-criterion 005Null hypothesis not rejected

Anomaly not detected1891-1915 Mean 271 -003

t-criterion 094Null hypothesis not rejected

Anomaly not detected1916-1940 Mean -175 222

t-criterion -079Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 492 322

t-criterion 060Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 440 034

t-criterion 137Null hypothesis not rejected

Anomaly not detected

16

1991-2015 Mean 607 074t-criterion 172

Null hypothesis not rejectedAnomaly not detected

Table 8 Trading simulation results of the Halloween effect in the US stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1791-1815 25 10 04 -1070 -043 - 048 failed1816-1840 26 10 038 -1329 -053 -044 failed1841-1865 26 12 046 5775 231 089 failed1866-1890 26 15 058 1501 060 037 failed1891-1915 26 15 058 7309 292 144 failed1916-1940 26 11 042 -4716 -189 -063 failed1941-1965 26 19 073 13284 531 261 passed1966-1990 26 15 058 11890 476 222 passed1991-2015 26 22 085 16399 656 361 passed

17

Figure 3 Trading simulation results of the Halloween effect in the US stock market

18

C Supplementary Appendices

C1 Halloween Effect in the UK Stock Market

Table 9 Average returns in the UK stock market

Period November to April May to October1693-1717 065 -1451718-1742 218 -3061743-1767 050 0331768-1792 077 -1341793-1817 053 -1111818-1842 204 -2661843-1867 227 -1281868-1892 070 0061893-1917 107 -1011918-1942 -014 1211943-1967 353 -6041968-1992 1116 -2271993-2017 452 -047

19

Figure 4 Average returns in the UK stock market

Table 10 ANOVA test of the Halloween effect in the UK stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1693-1717 041 053 404 not rejected not confirmed 0101718-1742 273 010 404 not rejected not confirmed 0681743-1767 001 092 404 not rejected not confirmed 0001768-1792 171 020 404 not rejected not confirmed 0421793-1817 066 042 404 not rejected not confirmed 0161818-1842 122 027 405 not rejected not confirmed 0301843-1867 399 005 405 not rejected not confirmed 0991868-1892 025 062 404 not rejected not confirmed 0061893-1917 137 025 405 not rejected not confirmed 0341918-1942 137 025 405 not rejected not confirmed 0341943-1967 137 025 405 not rejected not confirmed 0341968-1992 886 000 404 rejected confirmed 2191993-2017 367 006 404 not rejected not confirmed 091

20

Table 11 Mann-Whitney test of the Halloween effect in the UK stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1693-1717 019 1 066 384 not rejected not confirmed 0051718-1742 336 1 007 384 not rejected not confirmed 0881743-1767 000 1 099 384 not rejected not confirmed 0001768-1792 322 1 007 384 not rejected not confirmed 0841793-1817 104 1 031 384 not rejected not confirmed 0271818-1842 256 1 011 384 not rejected not confirmed 0671843-1867 339 1 007 384 not rejected not confirmed 0881868-1892 001 1 093 384 not rejected not confirmed 0001893-1917 293 1 009 384 not rejected not confirmed 0761918-1942 013 1 072 384 not rejected not confirmed 0031943-1967 063 1 043 384 not rejected not confirmed 0161968-1992 920 1 000 384 rejected confirmed 2401993-2017 200 1 016 384 not rejected not confirmed 052

21

Table 12 T-test of the Halloween effect in the UK stock market

Period Parameter November to April May to October1693-1717 Mean 065 -145

t-criterion 062Null hypothesis not rejected

Anomaly not detected1718-1742 Mean 218 -306

t-criterion 159Null hypothesis not rejected

Anomaly not detected1743-1767 Mean 050 033

t-criterion 010Null hypothesis not rejected

Anomaly not detected1768-1792 Mean 077 -134

t-criterion 126Null hypothesis not rejected

Anomaly not detected1793-1817 Mean 053 -111

t-criterion 078Null hypothesis not rejected

Anomaly not detected1818-1842 Mean 204 -266

t-criterion 108Null hypothesis not rejected

Anomaly not detected1843-1867 Mean 227 -128

t-criterion 192Null hypothesis rejected

Anomaly detected1868-1892 Mean 070 006

t-criterion 048Null hypothesis not rejected

Anomaly not detected1893-1917 Mean 107 -101

t-criterion 163Null hypothesis not rejected

Anomaly not detected

22

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Appendices

A Data

Table 3 Data

Country Stock Market Sample PeriodUS SampP 500 Index 1791-2015UK FTSE All Share Index 1693-2017Canada SampPTSX 300 Composite

Index1898-2018

France CAC All Tradable Index 1870-2018Germany CDAX Composite Index 1870-2018Italy Banca Commerciale Italiana

Index1905-2018

Japan Nikkei 225 1914-2013Switzerland Switzerland Composite Stock

Price Index1916-2015

B Halloween Effect in the US Stock Market

Table 4 Average returns in the US stock market

Period November to April May to October1791-1815 -041 -0041816-1840 -049 0321841-1865 214 -0491866-1890 056 0431891-1915 271 -0031916-1940 -175 2221941-1965 492 3221966-1990 440 0341991-2015 607 074

13

Figure 2 Average returns in the US stock market

Table 5 ANOVA test of the Halloween effect in the US stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1791-1815 003 087 405 not rejected not confirmed 0011816-1840 016 069 404 not rejected not confirmed 0041841-1865 042 052 404 not rejected not confirmed 0101866-1890 005 083 404 not rejected not confirmed 0011891-1915 089 035 404 not rejected not confirmed 0221916-1940 066 042 404 not rejected not confirmed 0161941-1965 046 050 404 not rejected not confirmed 0111966-1990 186 018 404 not rejected not confirmed 0461991-2015 333 007 404 not rejected not confirmed 082

14

Table 6 Mann-Whitney test of the Halloween effect in the US stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1791-1815 065 1 042 384 not rejected not confirmed 0171816-1840 104 1 031 384 not rejected not confirmed 0271841-1865 020 1 066 384 not rejected not confirmed 0051866-1890 000 1 095 384 not rejected not confirmed 0001891-1915 049 1 049 384 not rejected not confirmed 0131916-1940 132 1 025 384 not rejected not confirmed 0341941-1965 014 1 071 384 not rejected not confirmed 0041966-1990 111 1 029 384 not rejected not confirmed 0291991-2015 470 1 003 384 rejected confirmed 122

15

Table 7 T-test of the Halloween effect in the US stock market

Period Parameter November to April May to October1791-1815 Mean -041 -004

t-criterion -029Null hypothesis not rejected

Anomaly not detected1816-1840 Mean -049 032

t-criterion -046Null hypothesis not rejected

Anomaly not detected1841-1865 Mean 214 -049

t-criterion 061Null hypothesis not rejected

Anomaly not detected1866-1890 Mean 056 043

t-criterion 005Null hypothesis not rejected

Anomaly not detected1891-1915 Mean 271 -003

t-criterion 094Null hypothesis not rejected

Anomaly not detected1916-1940 Mean -175 222

t-criterion -079Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 492 322

t-criterion 060Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 440 034

t-criterion 137Null hypothesis not rejected

Anomaly not detected

16

1991-2015 Mean 607 074t-criterion 172

Null hypothesis not rejectedAnomaly not detected

Table 8 Trading simulation results of the Halloween effect in the US stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1791-1815 25 10 04 -1070 -043 - 048 failed1816-1840 26 10 038 -1329 -053 -044 failed1841-1865 26 12 046 5775 231 089 failed1866-1890 26 15 058 1501 060 037 failed1891-1915 26 15 058 7309 292 144 failed1916-1940 26 11 042 -4716 -189 -063 failed1941-1965 26 19 073 13284 531 261 passed1966-1990 26 15 058 11890 476 222 passed1991-2015 26 22 085 16399 656 361 passed

17

Figure 3 Trading simulation results of the Halloween effect in the US stock market

18

C Supplementary Appendices

C1 Halloween Effect in the UK Stock Market

Table 9 Average returns in the UK stock market

Period November to April May to October1693-1717 065 -1451718-1742 218 -3061743-1767 050 0331768-1792 077 -1341793-1817 053 -1111818-1842 204 -2661843-1867 227 -1281868-1892 070 0061893-1917 107 -1011918-1942 -014 1211943-1967 353 -6041968-1992 1116 -2271993-2017 452 -047

19

Figure 4 Average returns in the UK stock market

Table 10 ANOVA test of the Halloween effect in the UK stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1693-1717 041 053 404 not rejected not confirmed 0101718-1742 273 010 404 not rejected not confirmed 0681743-1767 001 092 404 not rejected not confirmed 0001768-1792 171 020 404 not rejected not confirmed 0421793-1817 066 042 404 not rejected not confirmed 0161818-1842 122 027 405 not rejected not confirmed 0301843-1867 399 005 405 not rejected not confirmed 0991868-1892 025 062 404 not rejected not confirmed 0061893-1917 137 025 405 not rejected not confirmed 0341918-1942 137 025 405 not rejected not confirmed 0341943-1967 137 025 405 not rejected not confirmed 0341968-1992 886 000 404 rejected confirmed 2191993-2017 367 006 404 not rejected not confirmed 091

20

Table 11 Mann-Whitney test of the Halloween effect in the UK stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1693-1717 019 1 066 384 not rejected not confirmed 0051718-1742 336 1 007 384 not rejected not confirmed 0881743-1767 000 1 099 384 not rejected not confirmed 0001768-1792 322 1 007 384 not rejected not confirmed 0841793-1817 104 1 031 384 not rejected not confirmed 0271818-1842 256 1 011 384 not rejected not confirmed 0671843-1867 339 1 007 384 not rejected not confirmed 0881868-1892 001 1 093 384 not rejected not confirmed 0001893-1917 293 1 009 384 not rejected not confirmed 0761918-1942 013 1 072 384 not rejected not confirmed 0031943-1967 063 1 043 384 not rejected not confirmed 0161968-1992 920 1 000 384 rejected confirmed 2401993-2017 200 1 016 384 not rejected not confirmed 052

21

Table 12 T-test of the Halloween effect in the UK stock market

Period Parameter November to April May to October1693-1717 Mean 065 -145

t-criterion 062Null hypothesis not rejected

Anomaly not detected1718-1742 Mean 218 -306

t-criterion 159Null hypothesis not rejected

Anomaly not detected1743-1767 Mean 050 033

t-criterion 010Null hypothesis not rejected

Anomaly not detected1768-1792 Mean 077 -134

t-criterion 126Null hypothesis not rejected

Anomaly not detected1793-1817 Mean 053 -111

t-criterion 078Null hypothesis not rejected

Anomaly not detected1818-1842 Mean 204 -266

t-criterion 108Null hypothesis not rejected

Anomaly not detected1843-1867 Mean 227 -128

t-criterion 192Null hypothesis rejected

Anomaly detected1868-1892 Mean 070 006

t-criterion 048Null hypothesis not rejected

Anomaly not detected1893-1917 Mean 107 -101

t-criterion 163Null hypothesis not rejected

Anomaly not detected

22

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Figure 2 Average returns in the US stock market

Table 5 ANOVA test of the Halloween effect in the US stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1791-1815 003 087 405 not rejected not confirmed 0011816-1840 016 069 404 not rejected not confirmed 0041841-1865 042 052 404 not rejected not confirmed 0101866-1890 005 083 404 not rejected not confirmed 0011891-1915 089 035 404 not rejected not confirmed 0221916-1940 066 042 404 not rejected not confirmed 0161941-1965 046 050 404 not rejected not confirmed 0111966-1990 186 018 404 not rejected not confirmed 0461991-2015 333 007 404 not rejected not confirmed 082

14

Table 6 Mann-Whitney test of the Halloween effect in the US stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1791-1815 065 1 042 384 not rejected not confirmed 0171816-1840 104 1 031 384 not rejected not confirmed 0271841-1865 020 1 066 384 not rejected not confirmed 0051866-1890 000 1 095 384 not rejected not confirmed 0001891-1915 049 1 049 384 not rejected not confirmed 0131916-1940 132 1 025 384 not rejected not confirmed 0341941-1965 014 1 071 384 not rejected not confirmed 0041966-1990 111 1 029 384 not rejected not confirmed 0291991-2015 470 1 003 384 rejected confirmed 122

15

Table 7 T-test of the Halloween effect in the US stock market

Period Parameter November to April May to October1791-1815 Mean -041 -004

t-criterion -029Null hypothesis not rejected

Anomaly not detected1816-1840 Mean -049 032

t-criterion -046Null hypothesis not rejected

Anomaly not detected1841-1865 Mean 214 -049

t-criterion 061Null hypothesis not rejected

Anomaly not detected1866-1890 Mean 056 043

t-criterion 005Null hypothesis not rejected

Anomaly not detected1891-1915 Mean 271 -003

t-criterion 094Null hypothesis not rejected

Anomaly not detected1916-1940 Mean -175 222

t-criterion -079Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 492 322

t-criterion 060Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 440 034

t-criterion 137Null hypothesis not rejected

Anomaly not detected

16

1991-2015 Mean 607 074t-criterion 172

Null hypothesis not rejectedAnomaly not detected

Table 8 Trading simulation results of the Halloween effect in the US stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1791-1815 25 10 04 -1070 -043 - 048 failed1816-1840 26 10 038 -1329 -053 -044 failed1841-1865 26 12 046 5775 231 089 failed1866-1890 26 15 058 1501 060 037 failed1891-1915 26 15 058 7309 292 144 failed1916-1940 26 11 042 -4716 -189 -063 failed1941-1965 26 19 073 13284 531 261 passed1966-1990 26 15 058 11890 476 222 passed1991-2015 26 22 085 16399 656 361 passed

17

Figure 3 Trading simulation results of the Halloween effect in the US stock market

18

C Supplementary Appendices

C1 Halloween Effect in the UK Stock Market

Table 9 Average returns in the UK stock market

Period November to April May to October1693-1717 065 -1451718-1742 218 -3061743-1767 050 0331768-1792 077 -1341793-1817 053 -1111818-1842 204 -2661843-1867 227 -1281868-1892 070 0061893-1917 107 -1011918-1942 -014 1211943-1967 353 -6041968-1992 1116 -2271993-2017 452 -047

19

Figure 4 Average returns in the UK stock market

Table 10 ANOVA test of the Halloween effect in the UK stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1693-1717 041 053 404 not rejected not confirmed 0101718-1742 273 010 404 not rejected not confirmed 0681743-1767 001 092 404 not rejected not confirmed 0001768-1792 171 020 404 not rejected not confirmed 0421793-1817 066 042 404 not rejected not confirmed 0161818-1842 122 027 405 not rejected not confirmed 0301843-1867 399 005 405 not rejected not confirmed 0991868-1892 025 062 404 not rejected not confirmed 0061893-1917 137 025 405 not rejected not confirmed 0341918-1942 137 025 405 not rejected not confirmed 0341943-1967 137 025 405 not rejected not confirmed 0341968-1992 886 000 404 rejected confirmed 2191993-2017 367 006 404 not rejected not confirmed 091

20

Table 11 Mann-Whitney test of the Halloween effect in the UK stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1693-1717 019 1 066 384 not rejected not confirmed 0051718-1742 336 1 007 384 not rejected not confirmed 0881743-1767 000 1 099 384 not rejected not confirmed 0001768-1792 322 1 007 384 not rejected not confirmed 0841793-1817 104 1 031 384 not rejected not confirmed 0271818-1842 256 1 011 384 not rejected not confirmed 0671843-1867 339 1 007 384 not rejected not confirmed 0881868-1892 001 1 093 384 not rejected not confirmed 0001893-1917 293 1 009 384 not rejected not confirmed 0761918-1942 013 1 072 384 not rejected not confirmed 0031943-1967 063 1 043 384 not rejected not confirmed 0161968-1992 920 1 000 384 rejected confirmed 2401993-2017 200 1 016 384 not rejected not confirmed 052

21

Table 12 T-test of the Halloween effect in the UK stock market

Period Parameter November to April May to October1693-1717 Mean 065 -145

t-criterion 062Null hypothesis not rejected

Anomaly not detected1718-1742 Mean 218 -306

t-criterion 159Null hypothesis not rejected

Anomaly not detected1743-1767 Mean 050 033

t-criterion 010Null hypothesis not rejected

Anomaly not detected1768-1792 Mean 077 -134

t-criterion 126Null hypothesis not rejected

Anomaly not detected1793-1817 Mean 053 -111

t-criterion 078Null hypothesis not rejected

Anomaly not detected1818-1842 Mean 204 -266

t-criterion 108Null hypothesis not rejected

Anomaly not detected1843-1867 Mean 227 -128

t-criterion 192Null hypothesis rejected

Anomaly detected1868-1892 Mean 070 006

t-criterion 048Null hypothesis not rejected

Anomaly not detected1893-1917 Mean 107 -101

t-criterion 163Null hypothesis not rejected

Anomaly not detected

22

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 6 Mann-Whitney test of the Halloween effect in the US stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1791-1815 065 1 042 384 not rejected not confirmed 0171816-1840 104 1 031 384 not rejected not confirmed 0271841-1865 020 1 066 384 not rejected not confirmed 0051866-1890 000 1 095 384 not rejected not confirmed 0001891-1915 049 1 049 384 not rejected not confirmed 0131916-1940 132 1 025 384 not rejected not confirmed 0341941-1965 014 1 071 384 not rejected not confirmed 0041966-1990 111 1 029 384 not rejected not confirmed 0291991-2015 470 1 003 384 rejected confirmed 122

15

Table 7 T-test of the Halloween effect in the US stock market

Period Parameter November to April May to October1791-1815 Mean -041 -004

t-criterion -029Null hypothesis not rejected

Anomaly not detected1816-1840 Mean -049 032

t-criterion -046Null hypothesis not rejected

Anomaly not detected1841-1865 Mean 214 -049

t-criterion 061Null hypothesis not rejected

Anomaly not detected1866-1890 Mean 056 043

t-criterion 005Null hypothesis not rejected

Anomaly not detected1891-1915 Mean 271 -003

t-criterion 094Null hypothesis not rejected

Anomaly not detected1916-1940 Mean -175 222

t-criterion -079Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 492 322

t-criterion 060Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 440 034

t-criterion 137Null hypothesis not rejected

Anomaly not detected

16

1991-2015 Mean 607 074t-criterion 172

Null hypothesis not rejectedAnomaly not detected

Table 8 Trading simulation results of the Halloween effect in the US stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1791-1815 25 10 04 -1070 -043 - 048 failed1816-1840 26 10 038 -1329 -053 -044 failed1841-1865 26 12 046 5775 231 089 failed1866-1890 26 15 058 1501 060 037 failed1891-1915 26 15 058 7309 292 144 failed1916-1940 26 11 042 -4716 -189 -063 failed1941-1965 26 19 073 13284 531 261 passed1966-1990 26 15 058 11890 476 222 passed1991-2015 26 22 085 16399 656 361 passed

17

Figure 3 Trading simulation results of the Halloween effect in the US stock market

18

C Supplementary Appendices

C1 Halloween Effect in the UK Stock Market

Table 9 Average returns in the UK stock market

Period November to April May to October1693-1717 065 -1451718-1742 218 -3061743-1767 050 0331768-1792 077 -1341793-1817 053 -1111818-1842 204 -2661843-1867 227 -1281868-1892 070 0061893-1917 107 -1011918-1942 -014 1211943-1967 353 -6041968-1992 1116 -2271993-2017 452 -047

19

Figure 4 Average returns in the UK stock market

Table 10 ANOVA test of the Halloween effect in the UK stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1693-1717 041 053 404 not rejected not confirmed 0101718-1742 273 010 404 not rejected not confirmed 0681743-1767 001 092 404 not rejected not confirmed 0001768-1792 171 020 404 not rejected not confirmed 0421793-1817 066 042 404 not rejected not confirmed 0161818-1842 122 027 405 not rejected not confirmed 0301843-1867 399 005 405 not rejected not confirmed 0991868-1892 025 062 404 not rejected not confirmed 0061893-1917 137 025 405 not rejected not confirmed 0341918-1942 137 025 405 not rejected not confirmed 0341943-1967 137 025 405 not rejected not confirmed 0341968-1992 886 000 404 rejected confirmed 2191993-2017 367 006 404 not rejected not confirmed 091

20

Table 11 Mann-Whitney test of the Halloween effect in the UK stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1693-1717 019 1 066 384 not rejected not confirmed 0051718-1742 336 1 007 384 not rejected not confirmed 0881743-1767 000 1 099 384 not rejected not confirmed 0001768-1792 322 1 007 384 not rejected not confirmed 0841793-1817 104 1 031 384 not rejected not confirmed 0271818-1842 256 1 011 384 not rejected not confirmed 0671843-1867 339 1 007 384 not rejected not confirmed 0881868-1892 001 1 093 384 not rejected not confirmed 0001893-1917 293 1 009 384 not rejected not confirmed 0761918-1942 013 1 072 384 not rejected not confirmed 0031943-1967 063 1 043 384 not rejected not confirmed 0161968-1992 920 1 000 384 rejected confirmed 2401993-2017 200 1 016 384 not rejected not confirmed 052

21

Table 12 T-test of the Halloween effect in the UK stock market

Period Parameter November to April May to October1693-1717 Mean 065 -145

t-criterion 062Null hypothesis not rejected

Anomaly not detected1718-1742 Mean 218 -306

t-criterion 159Null hypothesis not rejected

Anomaly not detected1743-1767 Mean 050 033

t-criterion 010Null hypothesis not rejected

Anomaly not detected1768-1792 Mean 077 -134

t-criterion 126Null hypothesis not rejected

Anomaly not detected1793-1817 Mean 053 -111

t-criterion 078Null hypothesis not rejected

Anomaly not detected1818-1842 Mean 204 -266

t-criterion 108Null hypothesis not rejected

Anomaly not detected1843-1867 Mean 227 -128

t-criterion 192Null hypothesis rejected

Anomaly detected1868-1892 Mean 070 006

t-criterion 048Null hypothesis not rejected

Anomaly not detected1893-1917 Mean 107 -101

t-criterion 163Null hypothesis not rejected

Anomaly not detected

22

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 7 T-test of the Halloween effect in the US stock market

Period Parameter November to April May to October1791-1815 Mean -041 -004

t-criterion -029Null hypothesis not rejected

Anomaly not detected1816-1840 Mean -049 032

t-criterion -046Null hypothesis not rejected

Anomaly not detected1841-1865 Mean 214 -049

t-criterion 061Null hypothesis not rejected

Anomaly not detected1866-1890 Mean 056 043

t-criterion 005Null hypothesis not rejected

Anomaly not detected1891-1915 Mean 271 -003

t-criterion 094Null hypothesis not rejected

Anomaly not detected1916-1940 Mean -175 222

t-criterion -079Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 492 322

t-criterion 060Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 440 034

t-criterion 137Null hypothesis not rejected

Anomaly not detected

16

1991-2015 Mean 607 074t-criterion 172

Null hypothesis not rejectedAnomaly not detected

Table 8 Trading simulation results of the Halloween effect in the US stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1791-1815 25 10 04 -1070 -043 - 048 failed1816-1840 26 10 038 -1329 -053 -044 failed1841-1865 26 12 046 5775 231 089 failed1866-1890 26 15 058 1501 060 037 failed1891-1915 26 15 058 7309 292 144 failed1916-1940 26 11 042 -4716 -189 -063 failed1941-1965 26 19 073 13284 531 261 passed1966-1990 26 15 058 11890 476 222 passed1991-2015 26 22 085 16399 656 361 passed

17

Figure 3 Trading simulation results of the Halloween effect in the US stock market

18

C Supplementary Appendices

C1 Halloween Effect in the UK Stock Market

Table 9 Average returns in the UK stock market

Period November to April May to October1693-1717 065 -1451718-1742 218 -3061743-1767 050 0331768-1792 077 -1341793-1817 053 -1111818-1842 204 -2661843-1867 227 -1281868-1892 070 0061893-1917 107 -1011918-1942 -014 1211943-1967 353 -6041968-1992 1116 -2271993-2017 452 -047

19

Figure 4 Average returns in the UK stock market

Table 10 ANOVA test of the Halloween effect in the UK stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1693-1717 041 053 404 not rejected not confirmed 0101718-1742 273 010 404 not rejected not confirmed 0681743-1767 001 092 404 not rejected not confirmed 0001768-1792 171 020 404 not rejected not confirmed 0421793-1817 066 042 404 not rejected not confirmed 0161818-1842 122 027 405 not rejected not confirmed 0301843-1867 399 005 405 not rejected not confirmed 0991868-1892 025 062 404 not rejected not confirmed 0061893-1917 137 025 405 not rejected not confirmed 0341918-1942 137 025 405 not rejected not confirmed 0341943-1967 137 025 405 not rejected not confirmed 0341968-1992 886 000 404 rejected confirmed 2191993-2017 367 006 404 not rejected not confirmed 091

20

Table 11 Mann-Whitney test of the Halloween effect in the UK stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1693-1717 019 1 066 384 not rejected not confirmed 0051718-1742 336 1 007 384 not rejected not confirmed 0881743-1767 000 1 099 384 not rejected not confirmed 0001768-1792 322 1 007 384 not rejected not confirmed 0841793-1817 104 1 031 384 not rejected not confirmed 0271818-1842 256 1 011 384 not rejected not confirmed 0671843-1867 339 1 007 384 not rejected not confirmed 0881868-1892 001 1 093 384 not rejected not confirmed 0001893-1917 293 1 009 384 not rejected not confirmed 0761918-1942 013 1 072 384 not rejected not confirmed 0031943-1967 063 1 043 384 not rejected not confirmed 0161968-1992 920 1 000 384 rejected confirmed 2401993-2017 200 1 016 384 not rejected not confirmed 052

21

Table 12 T-test of the Halloween effect in the UK stock market

Period Parameter November to April May to October1693-1717 Mean 065 -145

t-criterion 062Null hypothesis not rejected

Anomaly not detected1718-1742 Mean 218 -306

t-criterion 159Null hypothesis not rejected

Anomaly not detected1743-1767 Mean 050 033

t-criterion 010Null hypothesis not rejected

Anomaly not detected1768-1792 Mean 077 -134

t-criterion 126Null hypothesis not rejected

Anomaly not detected1793-1817 Mean 053 -111

t-criterion 078Null hypothesis not rejected

Anomaly not detected1818-1842 Mean 204 -266

t-criterion 108Null hypothesis not rejected

Anomaly not detected1843-1867 Mean 227 -128

t-criterion 192Null hypothesis rejected

Anomaly detected1868-1892 Mean 070 006

t-criterion 048Null hypothesis not rejected

Anomaly not detected1893-1917 Mean 107 -101

t-criterion 163Null hypothesis not rejected

Anomaly not detected

22

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

1991-2015 Mean 607 074t-criterion 172

Null hypothesis not rejectedAnomaly not detected

Table 8 Trading simulation results of the Halloween effect in the US stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1791-1815 25 10 04 -1070 -043 - 048 failed1816-1840 26 10 038 -1329 -053 -044 failed1841-1865 26 12 046 5775 231 089 failed1866-1890 26 15 058 1501 060 037 failed1891-1915 26 15 058 7309 292 144 failed1916-1940 26 11 042 -4716 -189 -063 failed1941-1965 26 19 073 13284 531 261 passed1966-1990 26 15 058 11890 476 222 passed1991-2015 26 22 085 16399 656 361 passed

17

Figure 3 Trading simulation results of the Halloween effect in the US stock market

18

C Supplementary Appendices

C1 Halloween Effect in the UK Stock Market

Table 9 Average returns in the UK stock market

Period November to April May to October1693-1717 065 -1451718-1742 218 -3061743-1767 050 0331768-1792 077 -1341793-1817 053 -1111818-1842 204 -2661843-1867 227 -1281868-1892 070 0061893-1917 107 -1011918-1942 -014 1211943-1967 353 -6041968-1992 1116 -2271993-2017 452 -047

19

Figure 4 Average returns in the UK stock market

Table 10 ANOVA test of the Halloween effect in the UK stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1693-1717 041 053 404 not rejected not confirmed 0101718-1742 273 010 404 not rejected not confirmed 0681743-1767 001 092 404 not rejected not confirmed 0001768-1792 171 020 404 not rejected not confirmed 0421793-1817 066 042 404 not rejected not confirmed 0161818-1842 122 027 405 not rejected not confirmed 0301843-1867 399 005 405 not rejected not confirmed 0991868-1892 025 062 404 not rejected not confirmed 0061893-1917 137 025 405 not rejected not confirmed 0341918-1942 137 025 405 not rejected not confirmed 0341943-1967 137 025 405 not rejected not confirmed 0341968-1992 886 000 404 rejected confirmed 2191993-2017 367 006 404 not rejected not confirmed 091

20

Table 11 Mann-Whitney test of the Halloween effect in the UK stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1693-1717 019 1 066 384 not rejected not confirmed 0051718-1742 336 1 007 384 not rejected not confirmed 0881743-1767 000 1 099 384 not rejected not confirmed 0001768-1792 322 1 007 384 not rejected not confirmed 0841793-1817 104 1 031 384 not rejected not confirmed 0271818-1842 256 1 011 384 not rejected not confirmed 0671843-1867 339 1 007 384 not rejected not confirmed 0881868-1892 001 1 093 384 not rejected not confirmed 0001893-1917 293 1 009 384 not rejected not confirmed 0761918-1942 013 1 072 384 not rejected not confirmed 0031943-1967 063 1 043 384 not rejected not confirmed 0161968-1992 920 1 000 384 rejected confirmed 2401993-2017 200 1 016 384 not rejected not confirmed 052

21

Table 12 T-test of the Halloween effect in the UK stock market

Period Parameter November to April May to October1693-1717 Mean 065 -145

t-criterion 062Null hypothesis not rejected

Anomaly not detected1718-1742 Mean 218 -306

t-criterion 159Null hypothesis not rejected

Anomaly not detected1743-1767 Mean 050 033

t-criterion 010Null hypothesis not rejected

Anomaly not detected1768-1792 Mean 077 -134

t-criterion 126Null hypothesis not rejected

Anomaly not detected1793-1817 Mean 053 -111

t-criterion 078Null hypothesis not rejected

Anomaly not detected1818-1842 Mean 204 -266

t-criterion 108Null hypothesis not rejected

Anomaly not detected1843-1867 Mean 227 -128

t-criterion 192Null hypothesis rejected

Anomaly detected1868-1892 Mean 070 006

t-criterion 048Null hypothesis not rejected

Anomaly not detected1893-1917 Mean 107 -101

t-criterion 163Null hypothesis not rejected

Anomaly not detected

22

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Figure 3 Trading simulation results of the Halloween effect in the US stock market

18

C Supplementary Appendices

C1 Halloween Effect in the UK Stock Market

Table 9 Average returns in the UK stock market

Period November to April May to October1693-1717 065 -1451718-1742 218 -3061743-1767 050 0331768-1792 077 -1341793-1817 053 -1111818-1842 204 -2661843-1867 227 -1281868-1892 070 0061893-1917 107 -1011918-1942 -014 1211943-1967 353 -6041968-1992 1116 -2271993-2017 452 -047

19

Figure 4 Average returns in the UK stock market

Table 10 ANOVA test of the Halloween effect in the UK stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1693-1717 041 053 404 not rejected not confirmed 0101718-1742 273 010 404 not rejected not confirmed 0681743-1767 001 092 404 not rejected not confirmed 0001768-1792 171 020 404 not rejected not confirmed 0421793-1817 066 042 404 not rejected not confirmed 0161818-1842 122 027 405 not rejected not confirmed 0301843-1867 399 005 405 not rejected not confirmed 0991868-1892 025 062 404 not rejected not confirmed 0061893-1917 137 025 405 not rejected not confirmed 0341918-1942 137 025 405 not rejected not confirmed 0341943-1967 137 025 405 not rejected not confirmed 0341968-1992 886 000 404 rejected confirmed 2191993-2017 367 006 404 not rejected not confirmed 091

20

Table 11 Mann-Whitney test of the Halloween effect in the UK stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1693-1717 019 1 066 384 not rejected not confirmed 0051718-1742 336 1 007 384 not rejected not confirmed 0881743-1767 000 1 099 384 not rejected not confirmed 0001768-1792 322 1 007 384 not rejected not confirmed 0841793-1817 104 1 031 384 not rejected not confirmed 0271818-1842 256 1 011 384 not rejected not confirmed 0671843-1867 339 1 007 384 not rejected not confirmed 0881868-1892 001 1 093 384 not rejected not confirmed 0001893-1917 293 1 009 384 not rejected not confirmed 0761918-1942 013 1 072 384 not rejected not confirmed 0031943-1967 063 1 043 384 not rejected not confirmed 0161968-1992 920 1 000 384 rejected confirmed 2401993-2017 200 1 016 384 not rejected not confirmed 052

21

Table 12 T-test of the Halloween effect in the UK stock market

Period Parameter November to April May to October1693-1717 Mean 065 -145

t-criterion 062Null hypothesis not rejected

Anomaly not detected1718-1742 Mean 218 -306

t-criterion 159Null hypothesis not rejected

Anomaly not detected1743-1767 Mean 050 033

t-criterion 010Null hypothesis not rejected

Anomaly not detected1768-1792 Mean 077 -134

t-criterion 126Null hypothesis not rejected

Anomaly not detected1793-1817 Mean 053 -111

t-criterion 078Null hypothesis not rejected

Anomaly not detected1818-1842 Mean 204 -266

t-criterion 108Null hypothesis not rejected

Anomaly not detected1843-1867 Mean 227 -128

t-criterion 192Null hypothesis rejected

Anomaly detected1868-1892 Mean 070 006

t-criterion 048Null hypothesis not rejected

Anomaly not detected1893-1917 Mean 107 -101

t-criterion 163Null hypothesis not rejected

Anomaly not detected

22

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

C Supplementary Appendices

C1 Halloween Effect in the UK Stock Market

Table 9 Average returns in the UK stock market

Period November to April May to October1693-1717 065 -1451718-1742 218 -3061743-1767 050 0331768-1792 077 -1341793-1817 053 -1111818-1842 204 -2661843-1867 227 -1281868-1892 070 0061893-1917 107 -1011918-1942 -014 1211943-1967 353 -6041968-1992 1116 -2271993-2017 452 -047

19

Figure 4 Average returns in the UK stock market

Table 10 ANOVA test of the Halloween effect in the UK stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1693-1717 041 053 404 not rejected not confirmed 0101718-1742 273 010 404 not rejected not confirmed 0681743-1767 001 092 404 not rejected not confirmed 0001768-1792 171 020 404 not rejected not confirmed 0421793-1817 066 042 404 not rejected not confirmed 0161818-1842 122 027 405 not rejected not confirmed 0301843-1867 399 005 405 not rejected not confirmed 0991868-1892 025 062 404 not rejected not confirmed 0061893-1917 137 025 405 not rejected not confirmed 0341918-1942 137 025 405 not rejected not confirmed 0341943-1967 137 025 405 not rejected not confirmed 0341968-1992 886 000 404 rejected confirmed 2191993-2017 367 006 404 not rejected not confirmed 091

20

Table 11 Mann-Whitney test of the Halloween effect in the UK stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1693-1717 019 1 066 384 not rejected not confirmed 0051718-1742 336 1 007 384 not rejected not confirmed 0881743-1767 000 1 099 384 not rejected not confirmed 0001768-1792 322 1 007 384 not rejected not confirmed 0841793-1817 104 1 031 384 not rejected not confirmed 0271818-1842 256 1 011 384 not rejected not confirmed 0671843-1867 339 1 007 384 not rejected not confirmed 0881868-1892 001 1 093 384 not rejected not confirmed 0001893-1917 293 1 009 384 not rejected not confirmed 0761918-1942 013 1 072 384 not rejected not confirmed 0031943-1967 063 1 043 384 not rejected not confirmed 0161968-1992 920 1 000 384 rejected confirmed 2401993-2017 200 1 016 384 not rejected not confirmed 052

21

Table 12 T-test of the Halloween effect in the UK stock market

Period Parameter November to April May to October1693-1717 Mean 065 -145

t-criterion 062Null hypothesis not rejected

Anomaly not detected1718-1742 Mean 218 -306

t-criterion 159Null hypothesis not rejected

Anomaly not detected1743-1767 Mean 050 033

t-criterion 010Null hypothesis not rejected

Anomaly not detected1768-1792 Mean 077 -134

t-criterion 126Null hypothesis not rejected

Anomaly not detected1793-1817 Mean 053 -111

t-criterion 078Null hypothesis not rejected

Anomaly not detected1818-1842 Mean 204 -266

t-criterion 108Null hypothesis not rejected

Anomaly not detected1843-1867 Mean 227 -128

t-criterion 192Null hypothesis rejected

Anomaly detected1868-1892 Mean 070 006

t-criterion 048Null hypothesis not rejected

Anomaly not detected1893-1917 Mean 107 -101

t-criterion 163Null hypothesis not rejected

Anomaly not detected

22

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Figure 4 Average returns in the UK stock market

Table 10 ANOVA test of the Halloween effect in the UK stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1693-1717 041 053 404 not rejected not confirmed 0101718-1742 273 010 404 not rejected not confirmed 0681743-1767 001 092 404 not rejected not confirmed 0001768-1792 171 020 404 not rejected not confirmed 0421793-1817 066 042 404 not rejected not confirmed 0161818-1842 122 027 405 not rejected not confirmed 0301843-1867 399 005 405 not rejected not confirmed 0991868-1892 025 062 404 not rejected not confirmed 0061893-1917 137 025 405 not rejected not confirmed 0341918-1942 137 025 405 not rejected not confirmed 0341943-1967 137 025 405 not rejected not confirmed 0341968-1992 886 000 404 rejected confirmed 2191993-2017 367 006 404 not rejected not confirmed 091

20

Table 11 Mann-Whitney test of the Halloween effect in the UK stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1693-1717 019 1 066 384 not rejected not confirmed 0051718-1742 336 1 007 384 not rejected not confirmed 0881743-1767 000 1 099 384 not rejected not confirmed 0001768-1792 322 1 007 384 not rejected not confirmed 0841793-1817 104 1 031 384 not rejected not confirmed 0271818-1842 256 1 011 384 not rejected not confirmed 0671843-1867 339 1 007 384 not rejected not confirmed 0881868-1892 001 1 093 384 not rejected not confirmed 0001893-1917 293 1 009 384 not rejected not confirmed 0761918-1942 013 1 072 384 not rejected not confirmed 0031943-1967 063 1 043 384 not rejected not confirmed 0161968-1992 920 1 000 384 rejected confirmed 2401993-2017 200 1 016 384 not rejected not confirmed 052

21

Table 12 T-test of the Halloween effect in the UK stock market

Period Parameter November to April May to October1693-1717 Mean 065 -145

t-criterion 062Null hypothesis not rejected

Anomaly not detected1718-1742 Mean 218 -306

t-criterion 159Null hypothesis not rejected

Anomaly not detected1743-1767 Mean 050 033

t-criterion 010Null hypothesis not rejected

Anomaly not detected1768-1792 Mean 077 -134

t-criterion 126Null hypothesis not rejected

Anomaly not detected1793-1817 Mean 053 -111

t-criterion 078Null hypothesis not rejected

Anomaly not detected1818-1842 Mean 204 -266

t-criterion 108Null hypothesis not rejected

Anomaly not detected1843-1867 Mean 227 -128

t-criterion 192Null hypothesis rejected

Anomaly detected1868-1892 Mean 070 006

t-criterion 048Null hypothesis not rejected

Anomaly not detected1893-1917 Mean 107 -101

t-criterion 163Null hypothesis not rejected

Anomaly not detected

22

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 11 Mann-Whitney test of the Halloween effect in the UK stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1693-1717 019 1 066 384 not rejected not confirmed 0051718-1742 336 1 007 384 not rejected not confirmed 0881743-1767 000 1 099 384 not rejected not confirmed 0001768-1792 322 1 007 384 not rejected not confirmed 0841793-1817 104 1 031 384 not rejected not confirmed 0271818-1842 256 1 011 384 not rejected not confirmed 0671843-1867 339 1 007 384 not rejected not confirmed 0881868-1892 001 1 093 384 not rejected not confirmed 0001893-1917 293 1 009 384 not rejected not confirmed 0761918-1942 013 1 072 384 not rejected not confirmed 0031943-1967 063 1 043 384 not rejected not confirmed 0161968-1992 920 1 000 384 rejected confirmed 2401993-2017 200 1 016 384 not rejected not confirmed 052

21

Table 12 T-test of the Halloween effect in the UK stock market

Period Parameter November to April May to October1693-1717 Mean 065 -145

t-criterion 062Null hypothesis not rejected

Anomaly not detected1718-1742 Mean 218 -306

t-criterion 159Null hypothesis not rejected

Anomaly not detected1743-1767 Mean 050 033

t-criterion 010Null hypothesis not rejected

Anomaly not detected1768-1792 Mean 077 -134

t-criterion 126Null hypothesis not rejected

Anomaly not detected1793-1817 Mean 053 -111

t-criterion 078Null hypothesis not rejected

Anomaly not detected1818-1842 Mean 204 -266

t-criterion 108Null hypothesis not rejected

Anomaly not detected1843-1867 Mean 227 -128

t-criterion 192Null hypothesis rejected

Anomaly detected1868-1892 Mean 070 006

t-criterion 048Null hypothesis not rejected

Anomaly not detected1893-1917 Mean 107 -101

t-criterion 163Null hypothesis not rejected

Anomaly not detected

22

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 12 T-test of the Halloween effect in the UK stock market

Period Parameter November to April May to October1693-1717 Mean 065 -145

t-criterion 062Null hypothesis not rejected

Anomaly not detected1718-1742 Mean 218 -306

t-criterion 159Null hypothesis not rejected

Anomaly not detected1743-1767 Mean 050 033

t-criterion 010Null hypothesis not rejected

Anomaly not detected1768-1792 Mean 077 -134

t-criterion 126Null hypothesis not rejected

Anomaly not detected1793-1817 Mean 053 -111

t-criterion 078Null hypothesis not rejected

Anomaly not detected1818-1842 Mean 204 -266

t-criterion 108Null hypothesis not rejected

Anomaly not detected1843-1867 Mean 227 -128

t-criterion 192Null hypothesis rejected

Anomaly detected1868-1892 Mean 070 006

t-criterion 048Null hypothesis not rejected

Anomaly not detected1893-1917 Mean 107 -101

t-criterion 163Null hypothesis not rejected

Anomaly not detected

22

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

1918-1942 Mean -014 121t-criterion -047

Null hypothesis not rejectedAnomaly not detected

1943-1967 Mean 353 -604t-criterion 208

Null hypothesis rejectedAnomaly detected

1968-1992 Mean 1116 -227t-criterion 290

Null hypothesis rejectedAnomaly detected

1993-2017 Mean 452 -047t-criterion 179

Null hypothesis not rejectedAnomaly not detected

Table 13 Trading simulation results of the Halloween effect in the UK stock market

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1693-1717 25 14 056 1684 067 035 failed1718-1742 25 15 06 5677 227 085 failed1743-1767 25 15 06 1295 052 048 failed1768-1792 25 15 06 1997 080 065 failed1793-1817 25 15 06 1385 055 054 failed1818-1842 25 15 06 5306 212 053 failed1843-1867 24 15 063 5665 227 188 passed1868-1892 25 11 044 1810 072 067 failed1893-1917 24 15 063 2664 107 148 failed1918-1942 24 13 054 -341 -014 -008 failed1943-1967 26 22 085 9258 370 264 passed1968-1992 26 20 077 30119 1205 344 passed1993-2017 26 20 077 12215 489 350 passed

23

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Figure 5 Trading simulation results of the Halloween effect in the UK stock market

24

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

C2 Halloween Effect in the Canadian Stock Market

Table 14 Average returns in the Canadian stock market

Period November to April May to October1915-1939 242 -0281940-1964 472 0771965-1989 721 -1571990-2014 468 016

Figure 6 Average returns in the Canadian stock market

25

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 15 ANOVA test of the Halloween effect in the Canadian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1915-1939 066 042 404 not rejected not confirmed 0161940-1964 188 018 404 not rejected not confirmed 0461965-1989 884 000 404 rejected confirmed 2191990-2014 229 014 404 not rejected not confirmed 057

Table 16 Mann-Whitney test of the Halloween effect in the Canadian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1915-1939 005 1 082 384 not rejected not confirmed 0011940-1964 104 1 031 384 not rejected not confirmed 0271965-1989 842 1 000 384 rejected confirmed 2191990-2014 132 1 025 384 not rejected not confirmed 034

26

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 17 T-test of the Halloween effect in the Canadian stock market

Period Parameter November to April May to October1915-1939 Mean 242 -028

t-criterion 072Null hypothesis not rejected

Anomaly not detected1940-1964 Mean 472 077

t-criterion 129Null hypothesis not rejected

Anomaly not detected1965-1989 Mean 721 -157

t-criterion 283Null hypothesis rejected

Anomaly detected1990-2014 Mean 468 016

t-criterion 142Null hypothesis not rejected

Anomaly not detected

Table 18 Trading simulation results of the Halloween effect in the Canadian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1915-1939 26 15 058 6535 654 134 failed1940-1964 26 18 069 12751 128 242 passed1965-1989 26 21 081 19463 1946 350 passed1990-2014 26 20 077 12624 1262 256 passed

27

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Figure 7 Trading simulation results of the Halloween effect in the Canadian stockmarket

28

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

C3 Halloween Effect in the French Stock Market

Table 19 Average returns in the French stock market

Period November to April May to October1898-1922 245 0141923-1947 642 2851948-1972 493 1701973-1997 1063 -2511998-2018 619 -260

Figure 8 Average returns in the French stock market

29

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 20 ANOVA test of the Halloween effect in the French stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1898-1922 074 039 404 not rejected not confirmed 0181923-1947 081 037 404 not rejected not confirmed 0201948-1972 111 030 404 not rejected not confirmed 0281973-1997 949 000 404 rejected confirmed 2351998-2018 475 004 408 rejected confirmed 116

Table 21 Mann-Whitney test of the Halloween effect in the French stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1898-1922 104 1 031 384 not rejected not confirmed 0271923-1947 049 1 049 384 not rejected not confirmed 0131948-1972 082 1 037 384 not rejected not confirmed 0211973-1997 820 1 000 384 rejected confirmed 2141998-2018 361 1 006 384 not rejected not confirmed 094

30

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 22 T-test of the Halloween effect in the French stock market

Period Parameter November to April May to October1898-1922 Mean 245 014

t-criterion 089Null hypothesis not rejected

Anomaly not detected1923-1947 Mean 642 285

t-criterion 080Null hypothesis not rejected

Anomaly not detected1948-1972 Mean 493 170

t-criterion 089Null hypothesis not rejected

Anomaly not detected1973-1997 Mean 1063 -251

t-criterion 300Null hypothesis rejected

Anomaly detected1998-2018 Mean 619 -260

t-criterion 208Null hypothesis rejected

Anomaly detected

31

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 23 Trading simulation results of the Halloween effect in the French stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1898-1922 26 16 062 6620 66 170 failed1923-1947 26 16 062 17332 1733 222 passed1948-1972 26 16 062 13300 1330 189 passed1973-1997 26 22 085 28691 2869 385 passed1998-2018 21 15 071 13615 1362 254 passed

Figure 9 Trading simulation results of the Halloween effect in the French stockmarket

32

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

C4 Halloween Effect in the German Stock Market

Table 24 Average returns in the German stock market

Period November to April May to October1870-1894 -038 1981895-1919 -279 -2811920-1944 852 -4391945-1969 539 0801970-1994 560 -1581995-2018 727 -210

Figure 10 Average returns in the German stock market

33

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 25 ANOVA test of the Halloween effect in the German stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1870-1894 079 038 405 not rejected not confirmed 0201895-1919 029 059 404 not rejected not confirmed 0071920-1944 167 020 404 not rejected not confirmed 0411945-1969 045 050 404 not rejected not confirmed 0111970-1994 487 003 404 rejected confirmed 1211995-2018 541 002 405 rejected confirmed 134

Table 26 Mann-Whitney test of the Halloween effect in the German stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1870-1894 074 1 039 384 not rejected not confirmed 0191895-1919 028 1 060 384 not rejected not confirmed 0071920-1944 320 1 007 384 not rejected not confirmed 0831945-1969 000 1 095 384 rejected confirmed 0001970-1994 546 1 002 384 not rejected not confirmed 1421995-2018 478 1 003 384 not rejected not confirmed 124

34

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 27 T-test of the Halloween effect in the German stock market

Period Parameter November to April May to October1870-1894 Mean -038 198

t-criterion -085Null hypothesis not rejected

Anomaly not detected1895-1919 Mean -279 -281

t-criterion 000Null hypothesis not rejected

Anomaly not detected1920-1944 Mean 852 -439

t-criterion 122Null hypothesis not rejected

Anomaly not detected1945-1969 Mean 539 080

t-criterion 063Null hypothesis not rejected

Anomaly not detected1970-1994 Mean 560 -158

t-criterion 210Null hypothesis rejected

Anomaly detected1995-2018 Mean 727 -210

t-criterion 223Null hypothesis rejected

Anomaly detected

35

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 28 Trading simulation results of the Halloween effect in the German stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1870-1894 25 14 056 -993 -040 - 022 failed1895-1919 26 17 065 -7523 -301 -093 failed1920-1944 26 18 069 23008 920 170 failed1945-1969 26 17 065 14551 582 277 passed1970-1994 26 19 073 15121 605 272 passed1995-2018 24 17 071 18178 727 301 passed

Figure 11 Trading simulation results of the Halloween effect in the German stockmarket

36

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

C5 Halloween Effect in the Italian Stock Market

Table 29 Average returns in the Italian stock market

Period November to April May to October1905-1929 143 -3371930-1954 635 6321955-1979 197 -1111980-2004 1409 -3542005-2018 292 -390

Figure 12 Average returns in the Italian stock market

37

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 30 ANOVA test of the Halloween effect in the Italian stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1905-1929 246 012 404 not rejected not confirmed 0611930-1954 000 099 404 not rejected not confirmed 0001955-1979 078 038 404 not rejected not confirmed 0191980-2004 1055 000 404 rejected confirmed 2612005-2018 175 020 423 not rejected not confirmed 042

Table 31 Mann-Whitney test of the Halloween effect in the Italian stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1905-1929 211 1 015 384 not rejected not confirmed 0551930-1954 141 1 024 384 not rejected not confirmed 0371955-1979 024 1 062 384 not rejected not confirmed 0061980-2004 1138 1 000 384 rejected confirmed 2962005-2018 166 1 020 384 not rejected not confirmed 043

38

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 32 T-test of the Halloween effect in the Italian stock market

Period Parameter November to April May to October1905-1929 Mean 143 -337

t-criterion 146Null hypothesis not rejected

Anomaly not detected1930-1954 Mean 635 632

t-criterion 000Null hypothesis not rejected

Anomaly not detected1955-1979 Mean 197 -111

t-criterion 079Null hypothesis not rejected

Anomaly not detected1980-2004 Mean 1409 -354

t-criterion 312Null hypothesis rejected

Anomaly detected2005-2018 Mean 292 -390

t-criterion 124Null hypothesis not rejected

Anomaly not detected

39

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 33 Trading simulation results of the Halloween effect in the Italian stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1905-1929 26 12 046 3857 386 076 failed1930-1954 26 18 069 17149 1715 119 failed1955-1979 26 13 05 5319 532 204 passed1980-2004 26 23 088 38040 3804 454 passed2005-2018 14 9 064 4376 438 078 failed

Figure 13 Trading simulation results of the Halloween effect in the Italian stockmarket

40

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

C6 Halloween Effect in the Japanese Stock Market

Table 34 Average returns in the Japanese stock market

Period November to April May to October1914-1938 309 -1621939-1963 1120 2191964-1988 962 1951989-2013 353 -604

Figure 14 Average returns in the Japanese stock market

41

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 35 ANOVA test of the Halloween effect in the Japanese stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1914-1938 153 022 405 not rejected not confirmed 0381939-1963 237 013 404 not rejected not confirmed 0591964-1988 750 001 404 rejected confirmed 1861989-2013 419 005 404 rejected confirmed 104

Table 36 Mann-Whitney test of the Halloween effect in the Japanese stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1914-1938 250 1 011 384 not rejected not confirmed 0651939-1963 141 1 024 384 not rejected not confirmed 0371964-1988 686 1 001 384 not rejected not confirmed 1791989-2013 486 1 003 384 rejected confirmed 127

42

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 37 T-test of the Halloween effect in the Japanese stock market

Period Parameter November to April May to October1914-1938 Mean 309 -162

t-criterion 114Null hypothesis not rejected

Anomaly not detected1939-1963 Mean 1120 219

t-criterion 138Null hypothesis not rejected

Anomaly not detected1964-1988 Mean 962 195

t-criterion 266Null hypothesis rejected

Anomaly detected1989-2013 Mean 353 -604

t-criterion 208Null hypothesis rejected

Anomaly detected

Table 38 Trading simulation results of the Halloween effect in the Japanese stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1914-1938 25 15 06 8043 322 112 failed1939-1963 26 19 073 30247 1210 220 passed1964-1988 26 20 077 25962 1038 485 passed1989-2013 26 18 069 9536 381 116 failed

43

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Figure 15 Trading simulation results of the Halloween effect in the Japanese stockmarket

44

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

C7 Halloween Effect in the Swiss Stock Market

Table 39 Average returns in the Swiss stock market

Period November to April May to October1916-1940 142 -2681941-1965 380 1181966-1990 381 -0911991-2015 605 018

Figure 16 Average returns in the Swiss stock market during

45

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 40 ANOVA test of the Halloween effect in the Swiss stock market

Period F p-value

F crit-ical

Null hypothesis Anomaly Anovamulti-plier

1916-1940 137 025 405 not rejected not confirmed 0341941-1965 094 034 404 not rejected not confirmed 0231966-1990 241 013 404 not rejected not confirmed 0601991-2015 372 006 404 not rejected not confirmed 092

Table 41 Mann-Whitney test of the Halloween effect in the Swiss stock market

Period AdjH

df p-value

Critvalue

Null hypothesis Anomaly Kruskallmul-ti-plier

1916-1940 154 1 021 384 not rejected not confirmed 0401941-1965 051 1 047 384 not rejected not confirmed 0131966-1990 179 1 018 384 not rejected not confirmed 0471991-2015 288 1 009 384 rejected confirmed 075

46

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Table 42 T-test of the Halloween effect in the Swiss stock market

Period Parameter November to April May to October1916-1940 Mean 142 -268

t-criterion 113Null hypothesis not rejected

Anomaly not detected1941-1965 Mean 380 118

t-criterion 084Null hypothesis not rejected

Anomaly not detected1966-1990 Mean 381 -091

t-criterion 145Null hypothesis not rejected

Anomaly not detected1991-2015 Mean 605 018

t-criterion 170Null hypothesis not rejected

Anomaly not detected

Table 43 Trading simulation results of the Halloween effect in the Swiss stockmarket

Period Numberof

tradesunits

Numberof suc-cessfultradesunit

Numberof suc-cessfultrades

Profit

Profit peryear

t-test Result

1916-1940 24 15 063 3543 354 078 failed1941-1965 26 18 069 10261 1026 250 passed1966-1990 26 15 058 10292 1029 164 failed1991-2015 26 21 081 16338 1634 292 passed

47

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48

Figure 17 Trading simulation results of the Halloween effect in the Swiss stockmarket

48