gilje taillard risk management 05-09-2014

8/9/2019 Gilje Taillard Risk Management 05-09-2014

1/48

∗

†

‡

∗


2/48


3/48

R 2

R 2

R 2

R 2


4/48


5/48


6/48


7/48


8/48


9/48


10/48


11/48

R 2


12/48


13/48

yi,t

Post t CADummy i

Post t ∗CADummy i

yi,t = α + β 1 CADummy t + β 2 Post t + β 3 Post t ∗CADummy i + FirmFE i + ε i,t

β 3 Post t ∗

CADummy i

yi,t


14/48

yi,t

Post t CADummy i HighLev i

Post t ∗ CADummy i Post t ∗ HighLev i CADummy i ∗

HighLev i Post t ∗CADummy i ∗HighLev i


15/48

β 7


16/48


17/48


18/48


19/48


20/48

∆ = − 1


21/48


22/48


23/48


24/48


25/48


26/48


27/48


28/48


29/48


30/48


31/48


32/48

Figure 1:This figure separately plots the benchmark light oil prices for the treatment (Canadian) and control (U.S.) firms in this study. The Canadian benchmark is the Edmonton PAR and the U.S. benchmark is the West Texas Intermediate

(WTI). The pre-event year used in the main tests in this study is the year prior to the reduced R 2. The first major dislocation and drop in R 2 is highlighted in Q1 2012. The post-event period used to measure firm outcomes in this studyis the post-event year, running from Q2 2012 to Q1 2013.

$70

$80

$90

$100

$110

$120

1/1/2011 4/1/2011 7/1/2011 10/1/2011 1/1/2012 4/1/2012 7/1/2012 10/1/2012 1/1/2013 4/1/2013 7/1/2013 10/1/2013

U S D o l l a r / B a r r e l

Oil Price Benchmarks: Edmonton PAR vs. West Texas Intermediate (WTI)

Edmonton PAR WTI USD

Pre ‐Event Year R2 = 0.54 Post ‐Event Year R2 = 0.26Oil Price

DislocationR2 = 0.23

32


33/48

Figure 2:This figure separately plots the cumulative stock returns for the equalweighted portfolio of treatment (Canadian) andcontrol (U.S.) firms used in this study. The pre-event year used in the main tests in thisstudy is the year prior to the reduced correlation of oil prices, running from Q1 2011 to Q4 2011. The first major price dislocation and drop in correlation between Canadian and U.S. benchmark prices ishighlighted in Q1 2012. The post-event period used to measure firm outcomes in this study is the post-event year, running from Q2 2012 to Q1 2013.

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

30%

1/3/2011 4/3/2011 7/3/2011 10/3/2011 1/3/2012 4/3/2012 7/3/2012 10/3/2012 1/3/2013 4/3/2013 7/3/2013

S t o c k

R e t u r n

Stock Returns: Canadian vs. U.S. Oil Firms

Canada Average Stock Return US Average Stock Return

Oil PriceDislocation

Pre-Event Year Post-EventYear

33


34/48

Figure 3A:This figure separately plots the cumulative stock returns for the equal weighted portfolio of highly leveraged treatment (Canadian) and control (U.S.) firms used in this study. Highly leveraged firms are defined as firms with leverage above their peer median leverage as of Dec 31, 2011. The pre-event year used in the main tests in this study is the yea r prior to the reduced correlation of oil prices, running from Q12011 to Q4 2011. The first major dislocation and drop in correlation is highlighted in Q1 2012. The post-event period used to measure f irm outcomes in this study is the post-event year, running from Q2 2012to Q1 2013.

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

30%

1/3/2011 4/3/2011 7/3/2011 10/3/2011 1/3/2012 4/3/2012 7/3/2012 10/3/2012 1/3/2013 4/3/2013 7/3/2013

S t o c k

R e t u r n

Stock Returns: Canadian vs. U.S. Oil Firms among highly leveraged firms


Oil PriceDislocation

Pre-Event Year Post-EventYear

34


35/48

Figure 3B:This figure separately plots the cumulative stock returns for the equal weighted portfolio of low leverage treatment (Canadian) and control (U.S.) firms used in this study. Low leverage firms are defined asfirms with leverage below their peer median leverage as of Dec 31, 2011. The pre-event year used in the main tests in this study is the year prior to the reduced correlation of oil prices, running from Q1 2011to Q4 2011. The first major dislocation and drop in correlation is highlighted in Q1 2012. The post-event period used to measure firm outcomes in this study is the post-event year, running from Q2 2012 to Q12013.

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

30%

1/3/2011 4/3/2011 7/3/2011 10/3/2011 1/3/2012 4/3/2012 7/3/2012 10/3/2012 1/3/2013 4/3/2013 7/3/2013

S t o c k

R e t u r n

Stock Returns: Canadian vs. U.S. Oil Firms among low leverage firms


Oil PriceDislocationPre-Event Year Post-Event

Year

35


36/48

Table 1. Summary Statistics

Panel A: Treatment Firms (Canada)

Number of Firms 42

Number of Firm Quarters 336

Mean Median Standard Dev.

1620.86 435.91 2993.87

61% 58% 21%

1.28 1.02 0.87

Book Leverage 0.39 0.40 0.13

Profitability 0.03 0.03 0.02

Investment Intensity 0.07 0.05 0.05

Panel B: Control Firms (U.S.)

Number of Firms 37

Number of Firm Quarters 296

Mean Median Standard Dev.

3291.99 1666.65 4871.79

Light Oil Percentage 65% 63% 23%

1.51 1.38 0.64

Book Leverage 0.54 0.54 0.17Profitability 0.04 0.04 0.03

Investment Intensity 0.08 0.07 0.05

This table contains summary statistics on data for treatment (Canadian) and control (U.S.) firms used in this study.Our study covers the time period from Q1 2011 to Q1 2013; nine quarters in total. The summary statistics belowexclude the event quarter in our study, Q1 2012, as this quarter is excluded from our tests. The variable definitions areas follows. Tobin's Q is defined as the ratio of total assets plus market capitalization minus common equity minus

deferred taxes and investment tax credit (atq + prccq × cshoq − ceqq − txditcq) to total assets (atq). Firm size ismeasured by total assets (atq). Investment intensity is defined as quarterly capital expenditures (capxq) normalized bytotal assets (atq). Book leverage is defined as the ratio of total liabilities (ltq) to total assets (atq). Profitability isdefined as quarterly operating profits (oancfq) normalized by total assets (atq). Lastly, we compute the share of lightoil sales for each firm (Light Oil Percentage) by computing the proportion of light oil revenues to total revenues for the fiscal year ending on December 31st 2011.

Tobins q

Assets ($ Millions)

Light Oil Percentage

Assets ($ Millions)

Tobins q

36


37/48

Panel C: Comparison of firms in Q4 2011

Canadian Firm Comparison Q4 2011

High Leverage Low Leverage Difference p-value

1704.66 1426.38 278.28 0.76

0.59 0.62 -0.02 0.73

% of Production Hedged 0.47 0.34 0.13 0.19

1.47 1.21 0.26 0.28

Book Leverage 0.49 0.30 0.19 0.00

Profitability 0.20 0.16 0.04 0.20

Investment Intensity 0.08 0.07 0.00 0.75 N 21 21

U.S. Firm Comparison Q4 2011

High Leverage Low Leverage Difference p-value

2356.70 3749.43 -1392.73 0.36

Light Oil Percentage 0.59 0.70 -0.12 0.13

% of Production Hedged 0.50 0.66 -0.16 0.16

1.49 1.58 -0.09 0.65Book Leverage 0.66 0.40 0.25 0.00

Profitability 0.22 0.24 -0.02 0.65

Investment Intensity 0.10 0.10 -0.01 0.76

N 18 19

Assets ($ Millions)

Tobins q

This panel compares high and low leverage firms within the treatment (Canadian) and control (U.S.) groups used in this study. Thecomparison is made as of Q4 2011, the quarter prior to the event under study.

Assets ($ Millions)

Light Oil Percentage

Tobins q

37


38/48

Table 2. Hedging and Capital Expenditures

Dependent Variable = Capital Expenditures/Assets

All Firms High Leverage Low Leverage All Firms(1) (2) (7) (8)

(β1) Canada Dummy i

(β2) Post t -0.001 0.004 -0.007 -0.007[-0.20] [0.52] [-0.68] [-0.68]

(β3) Canada Dummy i * Post t -0.020** -0.039*** -0.001 -0.001[-2.13] [-2.80] [-0.10] [-0.10]

(β4) High Leverage i

(β5) High Leverage i * Post t 0.011[0.86]

(β6) High Leverage i * Canada Dummy i

(β7) High Leverage i * Canada Dummy i * Post t -0.038**[-2.08]

FirmFE i Yes Yes Yes Yes

R 2 Within 0.124 0.257 0.040 0.181 N - Total Firm Years 158 78 80 158

This table reports firm level regressions that measure the change in investment activity for treatment (Canadian) firms impacted by the withdrawalof effective hedging instruments due to a significant increase in basis risk. The dependent variable is capital expenditures scaled by beginning of quarter assets. Firm quarter level observations are aggregated into two separate time periods, one for the average investments in the four quarters

prior to the loss of effective hedging instruments (Q1 2011 to Q4 2011) and one for after (Q2 2012 to Q1 2013). The resulting dataset has two time

periods per firm, one for the time period before the loss of hedging instruments by treatment (Canadian) firms and one for the time period after theloss of hedging instruments by treatment (Canadian) firms. U.S. oil producers serve as the control group. The Canada dummy variable takes thevalue of one for treatment (Canadian) firms. The Post dummy takes a value of one for the time period after the event. High (respectively low)leverage firms are firms with above (respectively below) median book leverage as of Dec. 31 2011. The High Leverage dummy variable takes thevalue of one for firms with high leverage. Standard errors are clustered by firm, with t-statistics reported in brackets below the coefficient estimates,where * indicates significance at the 10% level, ** at the 5% level, and *** at the 1% level.

Pre-Period = [Q1 2011 to Q4 2011], Post-Period = [Q2 2012 to Q1 2013]

Absorbed by FirmFE i



I/K i,t = α + β 1CA Dummy i + β 2 Post t + β 3 Post t * CA Dummy i + FirmFE i + εi,t

I/K i,t = α + β 1CA Dummy i + β 2 Post t + β 3 Post t * CA Dummy i + β 4 High Leverage i

+ β 5 High Leverage i * Post t + β 6 High Leverage i * CA Dummy i + β 7 High Leverage i * CA Dummy i * Post t + FirmFE i + εi,t

38


39/48

Table 3. Hedging and Firm Value

Dependent Variable = Tobin's Q



(β2) Post t -0.351*** -0.262*** -0.435*** -0.435***

[-5.71] [-3.98] [-4.36] [-4.38]

(β3) Canada Dummy i * Post t 0.004 -0.297 0.299** 0.299**[0.03] [-1.32] [2.61] [2.63]









This table reports firm level regressions that measure the change in firm value for treatment (Canadian) firms impacted by the withdrawal of effective hedging instruments due to a significant increase in basis risk. Firm value is proxied by Tobin's Q. Firm quarter level observations areaggregated into two separate time periods, one for the average Tobin's Q in the four quarters prior to the loss of effective hedging instruments (Q12011 to Q4 2011) and one for after (Q2 2012 to Q1 2013). The resulting dataset has two time periods per firm, one for the time period before the

loss of hedging instruments by treatment (Canadian) firms and one for the time period after the loss of hedging instruments by treatment(Canadian) firms. U.S. oil producers serve as the control group. High (respectively low) leverage firms are firms with above (respectively below)median book leverage as of Dec. 31 2011. All indicator variables are defined in Table 2. Standard errors are clustered by firm, with t-statisticsreported in brackets below the coefficient estimates, where * indicates significance at the 10% level, ** at the 5% level, and *** at the 1% level.



Q i,t = α + β 1CA Dummy i + β 2 Post t + β 3 Post t * CA Dummy i + FirmFE i + εi,t

Q i,t = α + β 1CA Dummy i + β 2 Post t + β 3 Post t * CA Dummy i + β 4 High Leverage i+ β 5 High Leverage i * Post t + β 6 High Leverage i * CA Dummy i + β 7 High Leverage i * CA Dummy i * Post t + FirmFE i + ε i,t

39

T bl 4 H dgi g d St k R t


40/48

Table 4. Hedging and Stock Returns

Dependent Variable = Cumulative Stock Returns

All Firms High Leverage Low Leverage All Firms All Firms High Leverage Low Leverage All Firms(1) (2) (3) (4) (5) (6) (7) (8)

(β1) Canada Dummy i -11.316 -30.072** 7.377 7.377 -8.306 -27.276*** 10.665 10.665[-1.14] [-2.47] [0.50] [0.55] [-1.13] [-3.11] [1.05] [1.13]

(β2) HighLeverage i 2.330 -0.002[0.17] [-0.00]

(β3) High Leverage i * Canada Dummy i -37.449* -37.941***[-1.94] [-2.81]

R 2 Within 0.017 0.141 0.006 0.102 0.016 0.207 0.028 0.197 N - Total Firm Years 79 39 40 79 79 39 40 79

This table reports regressions that measure cumulative stock returns for treatment (Canadian) firms impacted by the withdrawal of effective hedging instruments due to a significant increase in basis risk.The dependent variable is the nominal cumulative stock return from January 1, 2012 (onset of event quarter) to June 30, 2012 and March 31st 2013. U.S. oil producers serve as the control group. High(respectively low) leverage firms are firms with above (respectively below) median book leverage as of Dec. 31 2011. All indicator variables are defined in Table 2. T-statistics are reported in brackets

below the coefficient estimates, where * indicates significance at the 10% level, ** at the 5% level, and *** at the 1% level.

Window = [Jan 1, 2012 to June 30, 2012] Window = [Jan 1, 2012 to Mar 31, 2013]

StockReturn i = α + β 1CA Dummy i + β 2 HighLeverage i + β 3CA Dummy i * HighLeverage i + εi

40

Table 5 Asset Acquisitions/Dispositions by Canadian Firms


41/48

Table 5. Asset Acquisitions/Dispositions by Canadian Firms

High Leverage Low Leverage Difference(1) 2012 Net Acquisitions/Dispositions -5.85% 0.77% -6.62%*

(2) 2012 and 2013 Net Acquisitions/Dispositions -6.31% 2.87% -9.18%*

Canadian firms

This table reports acquisition and disposition activity by Canadian light oil producers after the basis risk shock. Specifically, it reportsthe net acquisitions conducted by Canadian firms, scaled by assets: (Acquisitions t - Dispositions t)/Assets t-1. The reported acquisitionsand dispositions activity periods correspond to (1): Q1-Q4 2012 and (2): Q1 2012-Q3 2013 (up to most recent quarterly filing). Meanactivity is computed for high and low leverage Canadian oil producers separately. High (low) leverage firms are defined as having book leverage above (below) median book leverage as of Q4 2011, the quarter prior to the event under study. Differences in means arecomputed where * indicates significance at the 10% level, ** at the 5% level, and *** at the 1% level.

41

Table 6 Placebo Test: Hedging and Capital Expenditures


42/48

Table . Placebo Test: Hedging and Capital Expenditures




(β2) PostPlacebo t 0.032*** 0.025*** 0.037*** 0.037***[4.78] [2.80] [3.90] [3.93]

(β3) Canada Dummy i * PostPlacebo t -0.015 -0.006 -0.023 -0.023[-1.40] [-0.47] [-1.38] [-1.39]


(β5) High Leverage i * PostPlacebo t -0.012[-0.92]


(β7) High Leverage i * Canada Dummy i * PostPlacebo t 0.017[0.78]




This table reports firm level regressions which measure the change in investment activity for treatment (Canadian) firms in response to a placeboevent which occurs over the 9 quarters prior to the actual withdrawal of effective hedging instruments in Q1 2012. The dependent variable iscapital expenditures scaled by beginning of quarter assets. Firm quarter level observations are aggregated into two separate time periods, one for the average investment in the quarters prior to the placebo event (Q4 2009 to Q3 2010) and one for average investment in the quarters after the

placebo event (Q1 2011 to Q4 2011). The resulting dataset has two time periods for a firm, one for the time period before the placebo treatment andone for the time period after the placebo treatment. U.S. oil producers serve as the control group. High (respectively low) leverage firms are firmswith above (respectively below) median book leverage as of Oct 31, 2010. All indicator variables are defined in Table 2. The PostPlacebo indicator variable indicates a post-placebo event observation. Standard errors are clustered by firm, with t-statistics reported in brackets below the coefficientestimates, where * indicates significance at the 10% level, ** at the 5% level, and *** at the 1% level.




I/K i,t = α + β 1CA Dummy i + β 2 PostPlacebo t + β 3 PostPlacebo t * CA Dummy i + FirmFE i + εi,t

I/K i,t = α + β 1CA Dummy i + β 2 PostPlacebo t + β 3 PostPlacebo t * CA Dummy i + β 4 High Leverage i+ β 5 High Leverage i * PostPlacebo t + β 6 High Leverage i * CA Dummy i + β 7 High Leverage i * CA Dummy i * PostPlacebo t + FirmFE i + εi,t

42

Table 7 . Placebo Test: Hedging and Firm Value


43/48

g g




(β2) PostPlacebo t 0.092 0.055 0.130 0.130[1.21] [0.62] [1.04] [1.04]

(β3) Canada Dummy i * PostPlacebo t -0.270 -0.090 -0.436 -0.436[-1.40] [-0.33] [-1.59] [-1.60]


(β5) High Leverage i * PostPlacebo t -0.075[-0.49]


(β7) High Leverage i * Canada Dummy i * PostPlacebo t 0.346[0.90]




This table reports firm level regressions which measure the change in firm value for treatment (Canadian) firms in response to a placebo eventwhich occurs over the 9 quarters prior to the actual withdrawal of effective hedging instruments in Q1 2012. The dependent variable is Tobin's Q.Firm quarter level observations are aggregated into two separate time periods, one for the average Tobin's Q in the quarters prior to the placeboevent (Q4 2009 to Q3 2010) and one for average Tobin's Q in the quarters after the placebo event (Q1 2011 to Q4 2011). The resulting dataset has

two time periods for a firm, one for the time period before the placebo treatment and one for the time period after the placebo treatment. U.S. oil producers serve as the control group. High (respectively low) leverage firms are firms with above (respectively below) median book leverage as of Oct 31, 2010. All indicator variables are defined in Table 2. The PostPlacebo indicator variable indicates a post-placebo event observation.Standard errors are clustered by firm, with t-statistics reported in brackets below the coefficient estimates, where * indicates significance at the 10%level, ** at the 5% level, and *** at the 1% level.




Q i,t = α + β 1CA Dummy i + β 2 PostPlacebo t + β 3 Post t * CA Dummy i + FirmFE i + εi,t

Q i,t = α + β 1CA Dummy i + β 2 PostPlacebo t + β 3 PostPlacebo t * CA Dummy i + β 4 High Leverage i+ β 5 High Leverage i * PostPlacebo t + β 6 High Leverage i * CA Dummy i + β 7 High Leverage i * CA Dummy i * PostPlacebo t + FirmFE i + εi,t

43

Table 8. Placebo Test: Effect of Negative Macro Oil Price Shock and Capital Expenditures


44/48




(β2) PostPriceShock t -0.041*** -0.036* -0.045*** -0.045***[-3.99] [-2.00] [-4.18] [-4.21]

(β3) Canada Dummy i * PostPriceShock t 0.012 0.017 0.008 0.008

[0.99] [0.86] [0.51] [0.51](β4) High Leverage i

(β5) High Leverage i * PostPriceShock t 0.009[0.41]


(β7) High Leverage i * Canada Dummy i * PostPriceShock t 0.009

[0.36]FirmFE i Yes Yes Yes Yes


High Lev Canadian Firms Post vs. Low Lev Canadian Firms Post ( β5 + β7) 0.017[1.29]


This table reports firm level regressions which measure the change in investment activity for treatment (Canadian) firms in response to a negative macro oil priceshock which occurs in Q3 and Q4 2008. The dependent variable is capital expenditures scaled by beginning of quarter assets. Firm quarter level observationsare aggregated into two separate time periods, one for the average investment in the quarters prior to the negative macro oil price shock (Q3 2007 to Q2 2008)and one for average investment in the quarters after the negative macro oil price shock (Q1 2009 to Q4 2009). The resulting dataset has two time periods for afirm, one for the time period before the negative macro oil price shock and one for the time period after the negative macro oil price shock. U.S. oil producersserve as the control group. High (respectively low) leverage firms are firms with above (respectively below) median book leverage as of June 30, 2008. Allindicator variables are defined in Table 2. The PostPriceShock indicator variable indicates a post-macro oil price shock observation. Standard errors are clustered

by firm, with t-statistics reported in brackets below the coefficient estimates, where * indicates significance at the 10% level, ** at the 5% level, and *** at the1% level.




I/K i,t = α + β 1CA Dummy i + β 2 PostPriceShock t + β 3 PostPriceShock t * CA Dummy i + FirmFE i + εi,t

I/K i,t = α + β 1CA Dummy i + β 2 PostPriceShock t + β 3 PostPriceShock t * CA Dummy i + β 4 High Leverage i+ β 5 High Leverage i * PostPriceShock t + β 6 High Leverage i * CA Dummy i + β 7 High Leverage i * CA Dummy i * PostPriceShock t + FirmFE i + εi,t

44

Table 9. Placebo Test: Effect of Negative Macro Oil Price Shock and Firm Value


45/48




(β2) PostPriceShock t -0.742*** -0.468*** -1.017** -1.017**[-3.12] [-4.62] [-2.21] [-2.23]

(β3) Canada Dummy i * PostPriceShock t 0.372 0.197* 0.554 0.554

[1.50] [1.84] [1.15] [1.16](β4) High Leverage i

(β5) High Leverage i * PostPriceShock t 0.550[1.17]


(β7) High Leverage i * Canada Dummy i * PostPriceShock t -0.357

[-0.73]FirmFE i Yes Yes Yes Yes


High Lev Canadian Firms Post vs. Low Lev Canadian Firms Post ( β5 + β7) 0.193[1.42]


This table reports firm level regressions which measure the change in investment activity for treatment (Canadian) firms in response to a negative macro oil priceshock which occurs in Q3 and Q4 2008. The dependent variable is capital expenditures scaled by beginning of quarter assets. Firm quarter level observationsare aggregated into two separate time periods, one for the average investment in the quarters prior to the negative macro oil price shock (Q3 2007 to Q2 2008)and one for average investment in the quarters after the negative macro oil price shock (Q1 2009 to Q4 2009). The resulting dataset has two time periods for afirm, one for the time period before the negative macro oil price shock and one for the time period after the negative macro oil price shock. U.S. oil producersserve as the control group. High (respectively low) leverage firms are firms with above (respectively below) median book leverage as of June 30, 2010. Allindicator variables are defined in Table 2. The PostPriceShock indicator variable indicates a post-macro oil price shock observation. Standard errors are clustered

by firm, with t-statistics reported in brackets below the coefficient estimates, where * indicates significance at the 10% level, ** at the 5% level, and *** at the1% level.




Q i,t = α + β 1CA Dummy i + β 2 PostPriceShock t + β 3 PostPriceShock t * CA Dummy i + FirmFE i + εi,t

Q i,t = α + β 1CA Dummy i + β 2 PostPriceShock t + β 3 PostPriceShock t * CA Dummy i + β 4 High Leverage i+ β 5 High Leverage i * PostPriceShock t + β 6 High Leverage i * CA Dummy i + β 7 High Leverage i * CA Dummy i * PostPriceShock t + FirmFE i + εi,t

45

Table 10 . Placebo test: Effect of Negative Macro Shock and Leverage on Stock Price Return


46/48

Dependent Variable = Cumulative Stock ReturnsWindow = [July 1, 2008 to Dec 31, 2008] Window = [July 1, 2008 to Sep 30, 2009]

Canadian Firms Canadian Firms(1) (2)

(β1) High Leverage i 11.069 3.713[0.78] [0.30]

R 2 Within 0.019 0.003 N - Total Firm Years 33 33

This table reports regressions which measure the change in stock price of treatment (Canadian) firms during the 2008 financial crises and commodity price collapse. The dependent variable is gross stock return, relative to July 1, 2008 (onset of oil price crash). High leverage firms are firms with above median book leverage as of July 1, 2008, while low leverage firms are firms with below median leverage as of July 1, 2008. T-statistics are reported in brackets below the coefficient estimates, where * indicates significance at the 10% level, ** at the 5% level, and *** at the 1% level.

StockReturn i = α + β 1 HighLeverage i + εi

46

Table 11 . Hedging and Capital Expenditures (Similar Oil Prices in Pre and Post Period)Thi bl fi l l i h h h i i i i f (C di ) fi i d b h i hd l f


47/48


All Firms High Leverage Low Leverage All Firms


(β2) Post t 0.005 0.014 -0.002 -0.002

[0.63] [1.63] [-0.15] [-0.15](β3) Canada Dummy i * Post t -0.028** -0.053*** -0.004 -0.004

[-2.52] [-3.66] [-0.28] [-0.28]








This table reports firm level regressions that measure the change in investment activity for treatment (Canadian) firms impacted by the withdrawal of effective hedging instruments due to a significant increase in basis risk. The dependent variable is capital expenditures scaled by beginning of quarter assets. Firm quarter level observations are aggregated into two separate time periods, one for the average investments in four quarters prior to the loss of effective hedging instruments, (Q3 2010 to Q2 2011) and one for after (Q2 2012 to Q1 2013). This comparison is done such that theoil prices in the pre and post period are similar (1.7% difference). The resulting dataset has two time periods per firm, one for the time period beforethe loss of hedging instruments by treatment (Canadian) firms and one for the time period after the loss of hedging instruments by treatment(Canadian) firms. U.S. oil producers serve as the control group. The Canada dummy variable takes the value of one for treatment (Canadian) firms.The Post dummy takes a value of one for the time period after the event. High (respectively low) leverage firms are firms with above (respectively

below) median book leverage as of Dec. 31 2011. The High Leverage dummy variable takes the value of one for firms with high leverage. Standarderrors are clustered by firm, with t-statistics reported in brackets below the coefficient estimates, where * indicates significance at the 10% level, **at the 5% level, and *** at the 1% level.




I/K i,t = α + β 1CA Dummy i + β 2 Post t + β 3 Post t * CA Dummy i + FirmFE i + εi,t

I/K i,t = α + β 1CA Dummy i + β 2 Post t + β 3 Post t * CA Dummy i + β 4 High Leverage i

+ β 5 High Leverage i * Post t + β 6 High Leverage i * CA Dummy i + β 7 High Leverage i * CA Dummy i * Post t + FirmFE i + εi,t

47

Table 1 2 . Hedging and Firm Value (Similar Oil Prices in Pre and Post Period)This table reports firm level regressions that measure the change in firm value for treatment (Canadian) firms impacted by the withdrawal of


48/48


All Firms High Leverage Low Leverage All Firms


(β2) Post t -0.535*** -0.374*** -0.680*** -0.680***[-5.38] [-3.58] [-4.27] [-4.29]

(β3) Canada Dummy i * Post t -0.006 -0.436 0.407** 0.407**[-0.02] [-0.99] [2.35] [2.36]




(β7) High Leverage i * Canada Dummy i * Post t -0.843*[-1.79]




This table reports firm level regressions that measure the change in firm value for treatment (Canadian) firms impacted by the withdrawal of effective hedging instruments due to a significant increase in basis risk. Firm value is proxied by Tobin's Q. Firm quarter level observations areaggregated into two separate time periods, one for the average Tobin's Q in four quarters prior to the loss of effective hedging instruments (Q32010 to Q2 2011) and one for after (Q2 2012 to Q1 2013). This comparison is done such that the oil prices in the pre and post period are similar

(1.7% difference). The resulting dataset has two time periods per firm, one for the time period before the loss of hedging instruments by treatment(Canadian) firms and one for the time period after the loss of hedging instruments by treatment (Canadian) firms. U.S. oil producers serve as thecontrol group. High (respectively low) leverage firms are firms with above (respectively below) median book leverage as of Dec. 31 2011. Allindicator variables are defined in Table 2. Standard errors are clustered by firm, with t-statistics reported in brackets below the coefficientestimates, where * indicates significance at the 10% level, ** at the 5% level, and *** at the 1% level.




Q i,t = α + β 1CA Dummy i + β 2 Post t + β 3 Post t * CA Dummy i + FirmFE i + εi,t

Q i,t = α + β 1CA Dummy i + β 2 Post t + β 3 Post t * CA Dummy i + β 4 High Leverage i+ β 5 High Leverage i * Post t + β 6 High Leverage i * CA Dummy i + β 7 High Leverage i * CA Dummy i * Post t + FirmFE i + ε i,t

48

gilje taillard risk management 05-09-2014

Documents