J. of Multi. Fin. Manag. 27 (2014) 11–21

Contents lists available at ScienceDirect

Journal of Multinational FinancialManagement

journal homepage: www.elsevier.com/locate/econbase

Tax revenues convergence across ASEAN, Pacificand Oceania countries: Evidence from clubconvergence

Nicholas Apergisa,∗, Arusha Coorayb

a Department of Economics, Curtin University, Perth, Australiab School of Economics, University of Wollongong, NSW 2552, Australia

a r t i c l e i n f o

Article history:Received 3 June 2014Accepted 5 June 2014Available online 17 June 2014

JEL classification:H20C33

Keywords:Tax revenuesASEANAsian PacificOceania countriesClub convergence approach

a b s t r a c t

The goal of the present paper is to investigate the degree of con-vergence in tax revenues for a panel of 11 ASEAN, Asia Pacific andOceania countries spanning the period 1990–2012. We apply themethodology of Phillips and Sul (2007) to various categories of taxrevenues to assess the presence of convergence clubs. We considerfour alternative categories of tax revenues. Overall, the results donot support the hypothesis that all countries converge to a singleequilibrium state in such tax revenues.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

The ASEAN encompasses 10 Southeast Asian nations – Brunei, Cambodia, Indonesia, Lao, Malaysia,Myanmar, Philippines, Singapore, Thailand and Vietnam. The ASEAN is currently in the process of

∗ Corresponding author. Tel.: +61 892664034.E-mail addresses: Nick.Apergis@curtin.edu.au (N. Apergis), arusha@uow.edu.au (A. Cooray).

http://dx.doi.org/10.1016/j.mulfin.2014.06.0071042-444X/© 2014 Elsevier B.V. All rights reserved.

12 N. Apergis, A. Cooray / J. of Multi. Fin. Manag. 27 (2014) 11–21

launching a single common market by 2015 through increased economic integration. In order toachieve this objective, the ASEAN member nations have established mechanisms through which tostrengthen the operation of its prevailing economic initiatives including the ASEAN Free Trade Area(AFTA), the ASEAN Framework Agreement on Services (AFAS) and the ASEAN Investment Area (AIA)(Jogarajan, 2011). The ASEAN countries planned to complete a network of bilateral tax treaties acrossall member nations by 2010. Although significant progress has been made towards achieving thisobjective, the process is still not complete. Some argue that an ASEAN multilateral tax treaty may bebetter for promoting intra-regional investments (ADBI, 2009). Taxation (among other factors, such astransparency, supply chains, labour markets) has the potential to either be a facilitator or an impedi-ment to greater integration and economic growth across ASEAN and Asia Pacific and Oceania countries,since taxation is associated with the likely increased investments into the region.

Against this backdrop, the goal of this study is to investigate, for the first time, tax policy con-vergence among ASEAN countries, as greater integration between the tax policies of these countrieswould accelerate the process of achieving a single common market. We examine tax policy conver-gence for income–profits–capital gains, social security contributions, property and general taxes ongoods and services. The ASEAN group is also exploring the means through which economic integrationcan be increased amongst its members and the Asia Pacific and Oceania including economies, such as,Japan, South Korea, China, Australia and New Zealand. Therefore, tax policy convergence between theASEAN and these countries is also examined. The empirical results are expecting to shed more light toquestions, such as whether should we expect countries to search for profitable (tax specializations, liketax havens which offer special tax treatments for a specific number of transactions. Additionally, thetask of tax convergence is of great importance for the ASEAN and Pacific Rim area, since the countriesmust face the trade-off between the national sovereignty and the potential of an economic unifiedmarket. The search over the effects of globalization on such state sovereignty has been taken up in thestrands of the literature of tax convergence, tax competition and tax harmonization (Reuven, 2000).Tax convergence is closely related to tax competition, since governments are pressured to reduce taxesto survive the tax competition.

The literature uses several alternative approaches to identify whether and when convergenceoccurs. Initial empirical tests of the convergence hypothesis considered ˇ-convergence. Withoutadditional control variables, the test considered absolute convergence, whereas with additionalcontrol variables, the test examined conditional convergence. The regressions used to test for ˇ-convergence are generally the log-linearized solutions to a non-stochastic model with an additive errorterm.

An alternative view of convergence, �-convergence, argues that a group of coun-tries/sectors/regions converge when the cross-section variance of the variable under considerationdeclines across time. As noted by Bliss (1999, 2000), however, the underlying assumption of anevolving data distribution introduces difficulties in the interpretation of the test distribution underthe null. Moreover, the rejection of the �-convergence hypothesis does not necessarily mean thatthey do not converge; the presence of transitional dynamics in the data could lead to the rejection ofthe null hypothesis of �-convergence.

Critics of ˇ-convergence argue that if countries/sectors/regions converge to a common equilibriumwith identical internal structures, then the dispersion of the variable under study should disappear inthe long-run as all converge to the same long-run path. If, however, they converge to convergence clubsor to their own unique equilibrium, the dispersion of this variable will not approach zero (Miller andUpadhyay, 2002). Moreover, in the latter case of specific equilibrium, the movements of the dispersionwill depend on the initial distribution of the variable under investigation relative to their final long-run outcomes. Overall, these two approaches suffer not only from specific estimation deficienciesassociated with the time series used (Caporale et al., 2009).

Finally, the use of cointegration and unit-root tests for determining convergence are subject toa number of serious drawbacks. First, these tests fail to detect convergence when, more than oneequilibria exist. Second, if the countries do converge, but the data available to the econometricianreflect a time period in which transitional dynamics prevail, cointegration and unit-root tests may not‘catch’ the tendency to converge. Thus, to study the issue of convergence requires that the researchermodel both transitional dynamics and long-run behaviour together in a consistent framework.

N. Apergis, A. Cooray / J. of Multi. Fin. Manag. 27 (2014) 11–21 13

Unfortunately, standard existing testing methodologies for convergence fail to account for bothregularities and, thus, cannot suitably test real economic convergence.

Therefore, an additional novelty of this paper stems from the implementation of the new method-ology of panel convergence testing, recommended by Phillips and Sul (2007). The philosophy of themethodological approach is based on the club convergence hypothesis, which considers that certaincountries, states, sectors or regions that belong to a club move from disequilibrium positions to theirclub-specific steady-state positions. This methodology has several appealing characteristics. To beginwith, no specific assumptions concerning the stationarity of the variable of interest and/or the exist-ence of common factors are necessary. Nevertheless, this convergence test could be interpreted asan asymptotic cointegration test without suffering from the small sample problems of unit root andcointegration testing. In addition, the methodology is based on a quite general form of a non-lineartime varying factor model.

Our results suggest evidence of club divergence in tax revenues with three to four clubs dependingon the type of tax investigated. This study will allow policy makers to determine the countries in whichtax systems have converged and help implement measures to promote greater integration betweentax systems of countries which have not converged. Tax convergence permits increasing the efficiencyof public policies and prevents member states from being exploited due to excessive taxes (Fuest andHuber, 1999), member states are able to compete strategically for mobile factors (Winner, 2005), andmember states are able to prevent a race to the bottom due to tax harmonization (Winner, 2005).

The rest of the paper is organized as follows. Section 2 reviews the recent empirical literature ontax revenues convergence. Section 3 presents the new methodology employed. Section 4 reports anddiscusses the results of the empirical analysis, while Section 5 summarizes the paper and offers somepolicy implications.

2. Literature review

The focus of much of the literature on tax convergence has been on the European Union (EU) andOECD nations. Empirical studies on tax convergence employing cross-sectional data have examinedbeta and sigma convergence (Esteve et al., 2000; Sosvilla-Rivero et al., 2001). Studies using time seriesdata have used mainly unit root tests (Blot and Serranito, 2006; Delgado and Presno, 2011) and studiesbased on panel data have used random effects estimation (Bretschger and Hettich, 2002). Delgado andPresno (2011) investigate tax policy convergence (for income and profits, social security contributions,property, general taxes on goods and services, taxes on specific goods and services) in the EU over theperiod 1965–2005 taking as reference Germany, the United Kingdom and the European average. Usingtime series methodology, they find little evidence of tax convergence across the countries in the sampleunder investigation. Blot and Serranito (2006) investigate convergence of fiscal policies among EMUmembers using unit root tests. Given that fiscal policy convergence was on the agenda of the Maastrichttreaty, they consider the likelihood of a break at this point. Convergence however, is not supported forall policy indicators. Evidence suggests that for those indicators convergence is supported, convergencecommenced prior to the Maastricht treaty. Kocenda et al. (2008) similarly examine fiscal convergenceof new EU members with the original EU15. Employing time series methods of fiscal convergenceallowing for structural breaks, they show significant heterogeneity in fiscal convergence.

Bretschger and Hettich (2002) investigate the impact of globalisation on corporate taxes for a panelof 14 OECD countries over the period 1967–1996 using a random effects model. They find evidence ofa negative and significant impact of globalisation on corporate taxes. They additionally find evidenceof an increase in labour taxes and social expenditures due to globalisation.

In a different strand of the literature, that exemplifies the ˇ-and �-convergence approaches,Sosvilla-Rivero et al. (2001) investigate the degree of beta and sigma convergence in the tax bur-den for 14 EU countries over the period 1965–1995. They find some evidence in favour of convergencein the sub-period 1967–1974, but evidence of divergence from 1974 to 1984 and evidence of con-vergence again beyond 1974. Similar conclusions are reached by Esteve et al. (2000) for fiscal policyconvergence in the EU over the period 1967–1994. Both beta and sigma convergence in tax policyindicate divergence over the period 1967–1979 and convergence over the period 1979–1994. Exam-ining if international tax competition leads to a shift of tax burden from mobile to immobile tax bases

14 N. Apergis, A. Cooray / J. of Multi. Fin. Manag. 27 (2014) 11–21

in 23 OECD economies, in particular, for small open economies over the period 1965–2000, Winner(2005) finds that capital mobility has a negative effect on capital tax burden and a positive impact onlabour tax burden. Country size is found to have a positive effect with small open economies impos-ing lower capital and labour taxes compared to larger ones. He finds evidence of an increase in taxcompetition rather than convergence between the countries since the mid 1980s. However, criticson ˇ-convergence question the adequacy of ˇ-convergence regressions. Binder and Pesaran (1999)show that ˇ-convergence, even when used to study the growth path of a given economy towards itsown steady state, can collapse in the case of a stochastic technological progress, while Durlauf et al.(2005) note that a negative coefficient (ˇ) on initial income in a cross-section framework could simplyimply that economies converge to their own different steady states. At the same time, critics of �-convergence argue that it provides a necessary, but not sufficient, ingredient for observing reductionsin any variable dispersion (Quah, 1993). If countries converge to a common equilibrium with sharedglobal technologies and identical internal structures, then the dispersion of income should disappearin the long-run as all countries converge to the same real per capita income. If, however, countriesconverge to convergence clubs or to their own unique equilibrium, the dispersion of real per capitaincome will not approach zero (Miller and Upadhyay, 2002).

In a study of corporate tax convergence using cross-sectional data and Ordinary Least Squares(OLS), Slemrod (2004) finds evidence that the corporate tax rate is independent of a country’s revenuerequirements. Slemrod (2004) however, observes a strong link between the top individual rate andthe top statutory corporate rate. Vintila and Tibulca (2012) investigate tax convergence in the EU usingcluster analysis. They on the contrary, detect a notable trend towards convergence among EU memberstates.

Devereux et al. (2002) and Devereux et al. (2008) similarly find evidence of tax competition ratherthan convergence across the EU and OECD countries. Devereux et al. (2002) examine the effect oftaxes on corporate income in the EU and G7 nations over a period of two decades. They concludethat effective tax rates on marginal investment have remained relatively unchanged but that thoseon more profitable investments have fallen due to tax-cutting and base-broadening restructuring.Devereux et al. (2008) in a study of whether OECD countries compete with each other over corpo-rate taxes to attract investment, find evidence of countries competing for the statutory tax rate, theeffective average tax rate and the effective marginal tax rate, but in particular, over the statutory taxrate and effective average tax rate. They state that this is due to locational choices by multinationalfirms.

Sorensen (2000, 2001, 2004) examines fiscal competition for mobile capital and its effect onresource allocation, income distribution and social welfare within a general equilibrium framework.He argues in favour of a fully harmonised corporate tax within the EU, observing the differencesbetween global and regional tax coordination. Fuest and Huber (1999) similarly argue that a coordi-nation of regional policies can help internalizing positive externalities of domestic policies. Theoreticalmodels of international taxation have also been put forward by Edwards and Keen (1996), Keen andMarchand (1997), Fuest and Huber (1999), Becker and Fuest (2010), Marchand et al. (2003), Kanburand Keen (1993), Razin and Sadka (1991), among others. The study of Becker and Fuest (2010), forinstance, shows that coordination of investment in transport cost cutting infrastructures within unioncountries, increases welfare and reduces tax competition. Marchand et al. (2003) argue that partialcoordination of withholding taxes on interest income can be welfare reducing as opposed to no coor-dination at all. Keen and Marchand (1997) show that welfare can be increased through a coordinateddecrease in the provision of local public inputs and an increase in the public provision of local publicgoods, and Kanbur and Keen (1993) find that smaller countries lose from coordination of tax ratesbetween those set in the non-cooperative equilibrium, but both small and large countries gain fromthe enforcement of a minimum tax.

To the best of our knowledge, there are no studies which have empirically examined tax con-vergence across the ASEAN and Asia Pacific and Oceania nations. Additionally, studies have also notemployed the Phillips and Sul (2007) convergence formulation to investigate tax policy convergence.Therefore, we extend upon the literature by empirically investigating tax policy convergence acrossthe ASEAN and the Asia Pacific and Oceania nations using this particular convergence methodologicalapproach.

N. Apergis, A. Cooray / J. of Multi. Fin. Manag. 27 (2014) 11–21 15

3. Econometric methodology

In this section, we outline the methodology proposed by Phillips and Sul (2007) (henceforth PS)to test for convergence in a panel of countries. Let us have panel data for a variable Xit, where i = 1,..., N and t = 1, ..., T, with N, T the number of countries and the sample size, respectively. Often Xit isdecomposed into two components, one systematic, git, and one transitory ait

Xit = git + ait (1)

PS transforms (1) in a way that common and idiosyncratic components in the panel are separated.Specifically,

Xit =(

git + ait

�t

)�t = ıit�t, for all i, t (2)

In this way, the variable of interest, Xit, is decomposed in two components, one common, �t, and oneidiosyncratic, ıit, both of which are time varying. This formulation enables testing for convergence bytesting whether the factor loadings ıit converge. To do so, PS define the relative transition parameter,hit, as

hit = Xit

(1/N)∑

Xit= ıit

(1/N)∑

ıit(3)

which measures the loading coefficient ıit in relation to the panel and, as such, the transition path forthe economy i relative to the panel average. The relative transition curves depict the relative transitioncoefficients hit, calculated from Eq. (3).

In the context of macroeconomic data, and since our interest is on the long-run behaviour, we firstremove the business cycle component of Xit by employing the Hodrick–Prescott (1997) filter. The onlyinput required is a smoothing parameter determined mainly by the frequency of the data.1 Havingextracted the trend component from the series denoted as Xit , we calculate the estimated transitionpaths as hit = Xit/(1/N)

∑Xit . Next, we construct the cross-sectional variation ratio H1/Ht, where

Ht = 1N

N∑i=1

(hit − 1)2

PS show that the transition distance Ht has a limiting form of

Ht∼ A

L(t)2t2˛as t → ∞

where A is a positive constant, L(t) is a slowly varying function, such as log(t + 1) and denotes thespeed of convergence. To test for the null hypothesis of convergence,

H0 : ıi = ı and ≥ 0

against the alternative

HA : ıi /= ı for all i, or ≺ 0

PS run the following log t regression:

log(

H1

Ht

)− 2 log L(t) = c + b log t + ut (4)

where L(t) = log(t + 1). The standard errors of the estimates are calculated using a heteroskedasticity andautocorrelation consistent (HAC) estimator for the long-run variance of the residuals. In this study, weemploy the quadratic spectral kernel and determine the bandwidth. By employing the conventional

1 In our application with quarterly data, the smoothing parameter �, is set equal to 1600.

16 N. Apergis, A. Cooray / J. of Multi. Fin. Manag. 27 (2014) 11–21

Table 1Total tax revenues.

Countries t-stat b coefficient

Full sample Australia, China, Indonesia, Japan, Malaysia,Myanmar, New Zealand, Philippines, Republicof Korea, Singapore, Thailand

−0.537 −9.342

1st club Australia, New Zealand, Singapore −0.155 −1.3752nd club Indonesia, Malaysia, Myanmar, Philippines,

Republic of Korea, Thailand2.327 2.482

Non-converging China, Japan

t-statistic tb the null hypothesis of convergence is rejected if tb < −1.65. In practice, this regression isrun after a fraction of the sample is removed. PS recommend starting the regression at some pointt = [rT], where [rT] is the integer part of rT, and r = 0.3.2

This null hypothesis implies relative convergence (conditional convergence) rather than absoluteconvergence (convergence in level). If we change the null hypothesis to ≥ 1, which is equivalent tob ≥ 2, we can test for absolute convergence. Given that rejection of the null for the panel as a wholedoes not imply the absence of club convergence, PS go one step beyond and develop an algorithm forclub convergence. We next briefly outline the basic steps of the respective algorithm.

4. Empirical analysis and results

4.1. Data

Quarterly data on total tax revenues and tax revenue classifications on income–profits–capitalgains, social security contributions, property, and general taxes on goods and services – all measuredas percentages of their corresponding GDP – are obtained from the Datastream database for a numberof ASEAN and Asia Pacific and Oceania countries, they are in local currency and spanning the period1990–2012. All tax revenues and the GDP are measured at 2000 constant prices. The countries includedin the sample are Australia, China, Indonesia, Japan, Malaysia, Myanmar, New Zealand, Philippines,Republic of Korea, Singapore and Thailand. Taxes on income–profits–capital gains are levied on theactual or presumptive net income of individuals, on the profits of corporations and enterprises, and oncapital gains, whether realized or not, on land, securities, and other assets. Social contributions includesocial security contributions by employees, employers, and self-employed individuals plus actual orimputed contributions to social insurance schemes operated by governments. Taxes on goods andservices include general sales and turnover or value added taxes, selective excises on goods, selectivetaxes on services, taxes on the use of goods or property, taxes on extraction and production of minerals,and profits of fiscal monopolies.

4.2. Empirical results

Table 1 reports the results for total tax revenues. We reject the null hypothesis of full convergencefor this variable at the 5% significance level. The results of the club convergence algorithm indicatethe presence of two clubs plus a non-converging group. Three countries (Australia, New Zealand,Singapore) form the first club, whereas the second club includes six countries (Indonesia, Malaysia,Myanmar, Philippines, the Republic of Korea and Thailand). China and Japan do not converge at all.

Table 2 reports the results for tax revenues on income, profits and capital gains. We reject thenull hypothesis of full convergence for this variable at the 5% significance level. The results of theclub convergence algorithm indicate the presence of three clubs with three countries, Australia, NewZealand, Singapore, in the first, Indonesia, Malaysia and Myanmar in the second club, and five countries

2 Extensive Monte Carlo simulations conducted by Phillips and Sul (2007) show that r = 0.3 is satisfactory in terms of bothsize and power.

N. Apergis, A. Cooray / J. of Multi. Fin. Manag. 27 (2014) 11–21 17

Table 2Tax revenues on income, profits and capital gains.

Countries t-stat b coefficient

Full sample Australia, China, Indonesia, Japan, Malaysia,Myanmar, New Zealand, Philippines, Republicof Korea, Singapore, Thailand

−0.792 −9.672

1st club Australia, New Zealand, Singapore 0.136 0.5732nd club Indonesia, Malaysia, Myanmar 0.371 3.3283rd club China, Japan, Philippines, Republic of Korea,

Thailand0.564 1.893

Table 3Taxes revenues on social security contributions.

Countries t-stat b coefficient

Full sample Australia, China, Indonesia, Japan, Malaysia,Myanmar, New Zealand, Philippines, Republicof Korea, Singapore, Thailand

−2.782 −6.548

1st club Australia, New Zealand, Republic of Korea,Singapore

2.518 3.509

2nd club Indonesia, Malaysia 1.472 3.8743rd club Japan, Philippines, Thailand 1.815 2.326Non-converging China, Myanmar

(China, Japan, Philippines, the Republic of Korea and Thailand) in the third. The strongest speed ofconvergence is between countries in the second group, Indonesia, Malaysia and Myanmar with b =3.328. Compared to the total tax revenues variable, we see more dispersion of countries with more inthe third group.

Table 3 reports the results for tax revenues on social security contributions. We reject again thenull hypothesis of full convergence at the 5% significance level. The results of the club convergencealgorithm indicate the presence of three clubs with four countries (Australia, New Zealand, Singaporeand the Republic of Korea) in the first club, two countries (Indonesia and Malaysia) in the second,and three countries (Japan, Philippines and Thailand) in the third. China and Myanmar form a non-converging group. The speed of convergence for tax revenues on social security contributions is highestbetween member countries in the second group, Indonesia and Malaysia with b = 3.874.

The last two tables (Tables 4 and 5) report the convergence findings for tax revenues on propertyand tax revenues on general taxes on goods and services, respectively. In particular, Table 4 displays thepresence of four converging clubs plus a non-converging group in terms of tax revenues on property.Australia and New Zealand form the first club, Indonesia, Malaysia and Myanmar the second, theRepublic of Korea and the Philippines the third, and Japan, Singapore and Thailand the fourth. Chinafalls into the non-converging group. These results document high dispersion across the countriesincluded in the sample under investigation. The highest speed of convergence is recorded betweenAustralia and New Zealand (b = 3.782). Finally, similar results are obtained in Table 5 that showsdivergence in terms of tax revenues on general taxes on goods and services. Table 5 indicates the

Table 4Tax revenues on property.

Countries t-stat b coefficient

Full sample Australia, China, Indonesia, Japan, Malaysia,Myanmar, New Zealand, Philippines, Republicof Korea, Singapore, Thailand

−0.584 −21.328

1st club Australia, New Zealand 0.471 3.7822nd club Indonesia, Malaysia, Myanmar 0.152 2.6733rd club Philippines, Republic of Korea 0.094 0.7934th club Japan, Singapore, Thailand 1.683 3.861Non-converging China

18 N. Apergis, A. Cooray / J. of Multi. Fin. Manag. 27 (2014) 11–21

Table 5Tax revenues on general taxes on goods and services.

Countries t-stat b coefficient

Full sample Australia, China, Indonesia, Japan, Malaysia,Myanmar, New Zealand, Philippines, Republicof Korea, Singapore, Thailand

−1.384 −23.906

1st club Australia, New Zealand 0.086 3.6722nd club Indonesia, Malaysia −0.037 −0.19413rd club Japan, Republic of Korea, Singapore 0.493 1.7854th club Philippines, Thailand 0.381 0.852Non-converging Myanmar, China

presence of four clubs with Australia and New Zealand in the first, Indonesia and Malaysia in thesecond, Japan, the Republic of Korea and Singapore in the third and the Philippines and Thailand in thefourth. This time the non-convergence group, in addition to China, contains the country of Myanmar.The highest speed of convergence for general taxes on goods and services is between Australia andNew Zealand (b = 3.672).

The results for all tax revenue groups are summarized in Table 6. There is strong evidence ofconvergence in all tax revenues between Australia and New Zealand. This is not surprising given theirgeographical proximity and similar economic conditions. Singapore forms part of this club with regardto tax revenues on income–profits–capital gains and social security contributions. There is also evi-dence of tax convergence between Indonesia and Malaysia. These two countries are also geographicallyclosely located and have similar cultural and economic backgrounds. Additionally, the taxable income(chargeable income/employment income less exemptions and deductions) of Indonesia, Myanmarand Malaysia are similar (National Tax Research Journal (NTRC, 2007). The tax revenues of Thailandand the Philippines also appear to move together. This perhaps is because the tax base of employmentincome, that is, the annual gross income less deductions of the Philippines is similar to the assessableincome (gross employment income less personal reliefs and standard/itemized deductions) tax baseof Thailand (NTRC, 2007).

Table 6Summary of results.

Tax revenues

Clubs Income, profitsand capitalgains

Social securitycontributions

Property Goods andservices

Total

1st club Australia, NewZealand,Singapore

Australia, NewZealand,Singapore,Republic ofKorea

Australia, NewZealand

Australia, NewZealand

Australia, NewZealand,Singapore

2nd club Indonesia,Malaysia,Myanmar

Indonesia,Malaysia

Indonesia,Malaysia,Myanmar

Indonesia,Malaysia

Indonesia,Malaysia,Myanmar,Philippines,Republic ofKorea, Thailand

3rd club China, Japan,Philippines,Republic ofKorea, Thailand

Japan,Philippines,Thailand

Philippines,Republic ofKorea

Japan, Republicof Korea,Singapore

-

4th club – – Japan,Singapore,Thailand

Philippines,Thailand

–

Non-converging – China,Myanmar

China China,Myanmar

China, Japan

N. Apergis, A. Cooray / J. of Multi. Fin. Manag. 27 (2014) 11–21 19

Fig. 1. Total tax revenues: relative transition curves of convergence clubs.

There is little evidence to suggest that China’s tax revenues will converge with those of the ASEANin the near future. Japan and the Republic of Korea show only weak evidence of convergence. Japan’stax revenues appear to be most closely aligned with those of Thailand and the Republic of Korea’s withthat of the Philippines. Overall, although we observe evidence that supports club divergence acrossthe ASEAN and the Asia Pacific and Oceania countries and in terms of disaggregated tax revenues, thisdivergence does not seem substantial when it comes to the total tax revenues, denoting that a specificmechanism in tax collection can generate tendencies for convergence. The empirical findings revealthat the tax policies remain national. Thus, policy makers, in the search of equilibrium between theneed for higher tax convergence and the need to maintain or increase welfare levels with increaseddemands for social protection, seem to have opted for deviating from an hypothetical common path inthe tax issues. In other words, the countries in our sample still do have some freedom in determiningthe level of taxes they can impose and, thus, the level of public services that they wish to provide. Inthis manner, these countries they should decide whether they will go for further convergence or theywill have to engage in a considerable amount of cooperation and harmonization of their tax systems.The critical question arising is whether the absence of tax convergence can generate some additionaleffects on the future of the area integration. In addition, these diverging findings confirm that theautonomy of the countries and differentiated economic structures and policies preferences about thedegree of public intervention prevails over the idea of potentially higher harmonization in the taxfield.

4.3. Robustness tests

Fig. 1 depicts the relative transition curves for total tax revenues of each convergence club. Visualinspection of these curves enables us to gain some insight on the outcomes of the testing methodologyand monitor the convergence of total tax revenues for each club, relatively to the sample average. Inparticular, the transition curves report a graphical picture about the tendency of the cluster partici-pants to converge or diverge from above or below 1, which is the convergence path reference pointduring the period under study. Both transition curves show that at 2012 (the end point of our sample)the tendency is towards convergence, though curves 1 seems to approach convergence from aboveand curve 2 seems to approach convergence from below.

Finally, Phillips and Sul (2009) argue that their convergence club methodology tends to overes-timate the number of clubs than their true number. To avoid this overdetermination, they run the

Table 7Convergence club classification (total tax revenues).

Club Tests of club merging

1 Club 1 + 2 = −0.083*

2 (−6.71)

* Statistical significant at the 5% level, while it rejects the null hypothesis of convergence. The figure in parenthesis denotest-statistics.

20 N. Apergis, A. Cooray / J. of Multi. Fin. Manag. 27 (2014) 11–21

algorithm across the sub-clubs to assess whether any evidence exists in support of merging clubs intolarger clubs. The results of the new converging tests are reported in Table 7. The empirical findingsdisplay that for the two sub-clubs there is no evidence to support mergers of the original clubs.

5. Conclusions

This paper investigated the issue of tax convergence across a sample of 11 ASEAN and Asia Pacificand Oceania countries spanning the period 1990–2012. To serve this objective, the novel method-ology of Phillips and Sul (2007) was used. This methodology used a non-linear factor model with acommon and an idiosyncratic component – both time-varying, which allowed for technical progressheterogeneity across these countries. In terms of total tax revenues, the empirical findings suggestedthat the 11 ASEAN and Asia Pacific and Oceania countries under study did not form a homogeneousconvergence club. In addition to the total tax revenues variables, we also used four different disaggre-gated tax revenues variables. We found no evidence supporting full convergence across all countriesin our sample. Rather, we discovered club divergence with a number of different clubs, depending onthe definition of the variable measured.

Club divergence could suggest that differences in tax revenues between clubs could exist over aperiod of time. The potential adverse effects of a non-harmonized tax system include tax competitionand harmful bidding wars between nations leading to a race to the bottom. The ASEAN and Asia Pacificand Oceania countries could follow the example of other regions which have been successful in taxpolicy cooperation. For example, Nordic tax cooperation is strongly aligned with the region’s overallcooperation objectives. The treaty on administrative assistance among these countries specifies therequirements for countries with respect to the exchange of information and tax collections (Jogarajan,2011). These countries could follow a similar practice to increase transparency and accelerate theprocess of tax cooperation. Similarly, ensuring that tax policies are tightly aligned with the objectivesof these countries would permit maximizing the welfare gains from tax coordination.

As future extensions of this empirical research, we can tackle the interaction between tax conver-gence and income convergence, since the combination of a number of economic perspectives, i.e. taxrevenues and income, could help to comprehend better both the diversity and the clustering insidethe region under study. Finally, future empirical attempts could link tax convergence issues to a num-ber of economic, political and institutional characteristics that are highly possible to influence the taxstructure in these economies.

References

ADBI, 2009. Second ADBI regional tax forum intensive course on international tax treaties for selected Asian countries. In:Executive Summary of Tax Forum Proceedings, Tokyo.

Becker, J., Fuest, C., 2010. EU regional policy and tax competition. Eur. Econ. Rev. 54, 150–161.Binder, M., Pesaran, M.H., 1999. Stochastic growth models and their econometric implications. J. Econ. Growth 4, 173–183.Bliss, C., 1999. Galton’s fallacy and economic convergence. Oxf. Econ. Pap. 51, 4–14.Bliss, C., 2000. Galton’s fallacy and economic convergence: a reply to Cannon and Duck. Oxf. Econ. Pap. 52, 420–422.Blot, C., Serranito, F., 2006. Convergence of fiscal policies in EMU: a unit-root test analysis with structural breaks. Appl. Econ.

Lett. 13, 211–216.Bretschger, L., Hettich, F., 2002. Globalization, capital mobility and tax competition: theory and evidence for OECD countries.

Eur. J. Political Econ. 18, 695–716.Caporale, G.M., Erdogan, B., Kuzin, V., 2009. Testing for convergence in stock markets: A non-linear factor approach. Discussion

Paper No. 932. Deutsches Institut fur Wirtschaftsforschung, Berlin.Delgado, F., Presno, M.J., 2011. Convergence of fiscal pressure in the EU: a time series approach. Appl. Econ. 43, 4257–4267.Devereux, M.P., Lockwood, B., Redoano, M., 2008. Do countries compete over corporate tax rates? J. Public Econ. 92, 1210–1235.Devereux, M.P., Griffith, R., Klemm, A., 2002. Corporate income tax reforms and international tax competition. Econ. Policy 35,

450–495.Durlauf, S.N., Johnson, P.A., Temple, R.W., 2005. Growth econometrics. In: Aghion, P., Durlauf, S.N. (Eds.), Handbook of Economic

Growth, Volume 1. Amsterdam, North Holland.Edwards, J., Keen, M., 1996. Tax competition and leviathan. Eur. Econ. Rev. 40, 113–134.Esteve, V., Sosvilla-Rivero, S., Tamarit, C., 2000. Convergence in fiscal pressure across EU countries. Appl. Econ. Lett. 7, 117–123.Fuest, C., Huber, B., 1999. Can tax coordination work? FinanzArchiv/Public Fin. Anal. 56, 443–458.Jogarajan, S., 2011. A multilateral tax treaty for ASEAN: lessons for the Andean, Caribbean, Nordic and South Asian nations.

Asian J. Compar. Law 6, 145–166.

N. Apergis, A. Cooray / J. of Multi. Fin. Manag. 27 (2014) 11–21 21

Kanbur, R., Keen, M., 1993. Jeux sans frontier: tax competition and tax coordination when countries differ in size. Am. Econ.Rev. 83, 877–892.

Keen, M., Marchand, M., 1997. Fiscal competition and the pattern of public spending. J. Public Econ. 66, 33–53.Kocenda, E., Kutan, A.M., Yigit, T.M., 2008. Fiscal convergence in the European Union. N. Am. J. Econ. Fin. 19, 319–330.Marchand, M., Pestieau, P., Sato, M., 2003. Can partial fiscal coordination be welfare worsening? A model of tax competition. J.

Urban Econ. 64, 451–458.Miller, S.M., Upadhyay, M.P., 2002. Total factor productivity and the convergence hypothesis. J. Macroecon. 24, 267–286.NTRC, 2007. Summary significant features of the income tax structure among ASEAN member countries. Natl. Tax Res. J. 19,

1–44.Phillips, P.C.B., Sul, D., 2009. Economic transition and growth. J. Appl. Econ. 24, 1153–1185.Phillips, P.C.B., Sul, D., 2007. Transition modeling and econometric convergence tests. Econometrica 75, 1771–1855.Quah, D., 1993. Galton’s fallacy and tests of the convergence hypothesis. Scand. J. Econ. 95, 427–443.Reuven, S.A.Y., 2000. Globalization, tax competition, and the fiscal crises of the welfare state. Harv. Law Rev. 113, 1573–1580.Slemrod, J., 2004. Are corporate tax rates, or countries, converging? J. Public Econ. 88, 1169–1186.Sorensen, P.B., 2004. International tax coordination: regionalism versus globalism. J. Public Econ. 88, 1187–1214.Sorensen, P.B., 2001. Tax coordination in the European Union: what are the issues? Swed. Econ. Policy Rev. 8, 143–195.Sorensen, P.B., 2000. The case for international tax co-ordination reconsidered. Econ. Policy 15, 429–472.Sosvilla-Rivero, S., Galindo, M.A., Meseguer, A.J., 2001. Tax burden convergence in Europe. Estudios de Economia Aplicada 17,

183–191.Vintila, G., Tibulca, I.L., 2012. Fiscal pressure and tax mix convergence in the European Union. J. East. Eur. Res. Bus. Econ.,

http://dx.doi.org/10.5171/2012.295339, 29533.Winner, H., 2005. Has tax competition emerged in OECD countries? Evidence from panel data. Int. Tax Public Fin. 12, 667–687.