1 Does hedging add value? Evidence from The Netherlands Master Thesis August, 2018 Abstract This study examines the corporate use of financial derivatives and firm value. Prior research concerning the value of hedging is mainly focused on the U.S. due to data availability. This paper aims to find additional empirical evidence by using hand-collected hedging data of Dutch firms. Univariate tests and multivariate regression analyses are carried out with panel data methodology including generalized least squares and fixed effects methods. The sample includes all public non-financial Dutch firms with non-missing data during the period of 2012 to 2017. Unlike previous studies, weak evidence is found of the existence of a hedging premium. The results imply that nonfinancial Dutch firms can increase their value by hedging, but the impact is close to zero. Hence, in the case of Dutch firms, hedging does not create shareholder value. Student: Yirong Lo (10753974) MSc. Finance (track: Quantitative Finance) Supervisor: Derya Güler, MSc.
42
Embed
Does hedging add value? Evidence from The Netherlands
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Does hedging add value?
Evidence from The Netherlands
Master Thesis August, 2018
Abstract
This study examines the corporate use of financial derivatives and firm value. Prior research
concerning the value of hedging is mainly focused on the U.S. due to data availability. This paper
aims to find additional empirical evidence by using hand-collected hedging data of Dutch firms.
Univariate tests and multivariate regression analyses are carried out with panel data
methodology including generalized least squares and fixed effects methods. The sample includes
all public non-financial Dutch firms with non-missing data during the period of 2012 to 2017.
Unlike previous studies, weak evidence is found of the existence of a hedging premium. The
results imply that nonfinancial Dutch firms can increase their value by hedging, but the impact is
close to zero. Hence, in the case of Dutch firms, hedging does not create shareholder value.
This document is written by student Yirong Lo who declares to take full responsibility for
the contents of this document.
I declare that the text and the work presented in this document are original and that no
sources other than those mentioned in the text and its references have been used in
creating it.
The Faculty of Economics and Business is responsible solely for the supervision of
completion of the work, not for the contents.
Acknowledgements
To mark the end of a meaningful graduate experience, I would like to take this moment
to express my gratitude for the ample opportunities that were given to me in the past four
years at the University of Amsterdam.
A notable highlight of my academic journey was an exchange semester in Singapore,
where I encountered the challenging field of Risk Management for the first time. This
exposure has ultimately led me to pursue this specific Master track, this thesis topic, and
perhaps the direction of a professional career in the future.
I would especially like to thank everyone who was involved in the completion of my
thesis. A special word of thanks goes to my supervisor, Derya Güler, for her
encouragement since the beginning of my thesis process. Furthermore, I would like to
thank colleagues of team Risk Solutions at KAS BANK N.V. for their flexibility in allowing
me to gain practical experience next to writing this thesis. Lastly, I am most grateful for
the continuous support from my friends and family.
3
Table of Contents
Statement of Originality .............................................................................................................................. 2
4 Data .......................................................................................................................................................... 19
4.1 Sources and Collection Procedure ....................................................................................... 19
(27%). The distribution of hedging over the years can be found in Appendix 3, p.37.
In line with previous findings, firms use currency derivatives the most. Currency
risk being the most hedged could be explained by the fact that The Netherlands has an
open economy and depend heavily on foreign trade with a current account surplus. This
is consistent with the findings of Bodnar et al. (2003). The largest industries in The
Netherlands by the number of observations are Manufacturing (276) and Services (150).
The smallest number of observations are found in Mining (18) and Wholesale Trade (18).
Moreover, firms in the Utilities segment are all hedgers.
21
Table 1. Sample Overview The sample includes all publicly-listed non-financial Dutch firms with non-missing data during the years 2012 to 2017. A firm is classified as hedger if it reported hedging interest rate, currency, and/or commodity price risk with derivatives in that year. Non-hedgers are firms that do not make use of any financial derivatives for hedging purposes in that year. Industries are classified according to 2-digit SIC codes.
Hedgers Non-Hedgers N Currency (%) Interest rate (%) Commodity price (%) (%) By industry:
Note that multiple types of risk can be hedged simultaneously. An example to interpret the table: 40.31% of the total sample are non-hedgers. This means that 59.69% are hedgers. From the total amount of hedgers, 80.34% hedge currency risk. Thus, approximately 48% of the total sample uses currency derivatives for hedging purposes (0.5969×0.8034=0.4795).
4.3 Descriptive Statistics
Summary statistics of relevant variables are shown in Table 2. Firms in the sample
have a mean (median) market cap of €4,264 (€404) million and mean (median) value of
assets of €9,983 (€625) million. This is similar to the samples in previous studies focusing
on the US, Spanish, and French markets, except that small firms have not been excluded
which is similar to Ayturk who studied the Turkish market. The average D/E ratio is 68%,
34% issue dividends, average profit margin of -0.35%. This is also similar to the sample
in previous studies. Additionally, it can be seen that median values are smaller than mean
values. For example, the relatively large difference in mean and median of total assets
indicates that there are more smaller firms in the sample than larger firms. The same is
true for total sales and market cap. This skewness is corrected for by log-transformation
of the variables.
Out of the 588 observations, 351 are classified as hedgers. Total notional
derivatives amounts are recorded for 434 observations. Thus, the sample size is smaller
for the regression with the continuous hedging variable. On average, a Dutch firm hedges
€1.97 billion per year in total notional amounts. Hedgers that did not report notional
22
values are accepted as missing values. The low hedging medians are due to non-hedgers
(40%) having a notional derivatives value of 0, in line with the hedging dummies.
Table 2. Descriptive Statistics This table presents summary statistics that describe the sample, including all variables. The sample includes all public non-financial Dutch firms with non-missing data from 2012-2017. Firm characteristics are directly obtained from Compustat and DataStream. The dependent variable is firm value, measured in four ways as described on p.14. Detailed construction of the control variables are explained on p.15&16. Data of hedging variables are manually collected from annual reports. N Mean Median St. Dev. Min Max Firm characteristics Total assets (€ mln) 588 11,944.26 663.19 44815 0.007 565258 Total sales (€ mln) 588 12,816.72 793.81 81593.39 0 1295008 Market Cap (€ mln) 588 4419.06 405.85 9904.83 0.95 86096.44 Business Diversification (dummy) 588 0.6531 1 0.4764 0 1 Dependent variable FV1 588 0.1433 0.0629 0.9113 -1.8072 4.7885 FV2 588 0.4589 0.3303 0.7648 -1.5238 4.7958 FV3 588 0.1092 7.45e-9 0.9002 -1.9232 4.7861 Hedging variables Hedger dummy 588 0.5969 1 0.4909 0 1 Interest rate hedger dummy 588 0.4184 0 0.4937 0 1 Currency hedger dummy 588 0.4915 0 0.5004 0 1 Commodity price hedger dummy 588 0.1650 0 0.3715 0 1 Total Hedging (€ mln) 434 343.4272 0 10544.84 0 107909 Interest Rate Hedging (€ mln) 510 343.43 0 1232.93 0 12400 Currency Hedging (€ mln) 508 1,509.2 0 8696.50 0 99606 Commodity price hedging (€ mln) 540 12.333 0 77.62 0 765 Control variables Firm Size 588 6.3844 6.4971 2.9126 -4.9618 13.245 Profitability 588 -0.1289 0.0229 1.0232 -7.9061 1.7180 Leverage 588 0.3011 0.2172 0.6505 0 9.1660 Growth Opportunities 588 0.0796 0.0422 0.2023 0 3.9239 International Diversification 588 0.4604 0.5255 0.4107 0 1 Liquidity 588 0.6422 0.2198 1.4206 -0.6512 9.5911 Dividend dummy 588 0.3384 0 0.4736 0 1
The 1% extreme outliers of the dependent variables is winsorized by replacement.
23
5 Results
This section presents empirical results from the analyses that test the main hypothesis.
That is, whether hedging adds value to shareholders. Two approaches are used: 1.
univariate tests to compare hedgers with non-hedgers, and 2. Multivariate regression
analysis to test the effect of hedging on firm value.
5.1 Univariate Results
The main hypothesis that firms using derivatives for hedging are rewarded by
investors with higher valuation is tested here. Hedgers and non-hedgers are split using
the hedger dummy variable using four measures of firm value. A firm is classified as
hedger if it reported in that year to have used derivatives for hedging purposes. Firm
value is compared between hedgers and non-hedgers using two-sample t-tests of mean
values, as shown in Table 3.
Table 3. Comparison of Firm Value between Hedgers and Non-hedgers This table presents a univariate comparison of firm value (FV) between hedgers and non-hedgers with three different measures. The construction of the three measures of firm value are explained in detail on p.14. A firm is classified as hedger for a given year if it reports to have managed risk with the use of derivatives. The sample consists of 588 observations and includes all public non-financial Dutch firms in the period 2012-2017. Hedgers Non-Hedgers Difference t-value General hedging FV1 0.0422 0.2930 -0.2508*** -3.3010 FV2 0.3714 0.5886 -0.2172*** -3.4086 FV3 0.0496 0.3020 -0.2524*** -3.9798
Difference in the means are compared using t-tests. The 1% outliers of Firm Value are winsorized by replacement. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.
24
Table 3 reports differences in mean firm values and related t statistics obtained from
twelve t-tests. The hypothesis is tested separately for general hedging, currency hedging,
interest rate hedging, and commodity price hedging. Four types of hedging are examined,
each with three measures of firm value. General hedging represents firms that reported
to have used currency, interest rate, and/or commodity derivatives for hedging purposes
in a given year. Whereas currency hedging, interest rate hedging, and commodity price
hedging represent the corresponding type of hedging separately.
The univariate results suggest that hedgers have lower firm value in comparison
to non-hedgers. This is true for all three different measures of firm value and indicates
some robustness. The difference in firm value between firms that use either type of
hedging (General hedging) as opposed to firms that do not hedge at all is -0.25, -0.22, and
-0.25 for all three measures respectively. In economic terms4, this translates into value
differences of -22.18%, -19.52%, and -22.31% resulting in an average hedging discount
of -21.34%. This finding rejects the hypothesis at a 1% significance level and suggests
that hedgers are valued lower than non-hedgers. More specifically, using the same
calculation method, the average hedging discounts are -15.09% for currency hedging, -
12.68% for interest rate hedging, and -19.47% commodity hedging.
From the three types of hedging, commodity price hedging seems to be the most
negative. Commodity derivatives are also the least used for hedging, as shown in the
sample overview (Table 1). A plausible reason for this may be that commodities are often
part of a firm’s core business. Therefore, fluctuations in commodity prices would impact
the firm’s core business more than currency or interest rate changes would in general.
Another reason why the results of commodity price hedging deviate from currency and
interest rate hedging could be due to the small sample size. The sample records only 16%
of commodity hedgers. While roughly 50% and 40% is reported for currency and interest
hedgers.
Overall, the univariate findings are vastly inconsistent with numerous studies that
infer a valuation premium imposed by the use of derivatives. An explanation may be due
to the existence of large differences in structural characteristics as well as market
valuation of Dutch and U.S. firms. Instead, evidence of a hedging discount is in line with
4 Calculation of hedging discount e.g. (e-0.2508-1×100%) = -22.18%
25
Khediri (2010) and, to an extent, supports Jin & Jorion (2006) and Ayturk et al. (2016).
Khediri (2010) argues that a hedging discount exists due to investors’ perception of
corporate hedging decisions which may be linked to insider’s motives and their risk
aversion. Hence, investors value the derivatives use at a discount.
To further analyze the differences between hedgers and non-hedgers, an
additional comparison between hedgers and non-hedgers are drawn from other
variables (see Appendix 9, p.38). The results in A3 show that the characteristics of
hedgers and non-hedgers are significantly different within the sample. More specifically,
hedgers are larger in size; more profitable and diversified; and more likely to issue
dividends. Whereas non-hedgers are more leveraged, liquid, and have more growth
opportunities. Differences in these variables could also explain why hedgers are valued
lower by shareholders. So it is important to further investigate with regression analyses
in order to identify a causal relationship.
5.2 Regression Results
The findings in the univariate tests suggest that hedgers are valued lower by
investors than non-hedgers. In this paragraph, linear regressions are performed to
control for other variables that may have impact on firm value. To estimate the
coefficients of Equations 1 & 2, two panel regression models are used. Model 1 regresses
on the Hedging Dummy and estimates with generalized least squares method. Whereas
Model 2 regresses the Amount of Hedging with fixed firm effects method. Following
previous studies, the model includes various control variables and year dummies to fix
for time-specific effects as well.
For a clear appearance, one measure of firm value (FV1) as dependent variable is
presented here. The other two measures are presented in the robustness section later.
The number of observations for notional amounts of hedging (434) is lower than that of
hedging dummies (588). This is due to some firms that do not report notional values of
derivatives usage although they do report hedging activity. In such cases, the notional
value of derivative instruments are accepted as missing data. Additionally, for some
observation years, multiple types of hedging are executed, but not all notional values are
reported. Therefore, the number of observations in Model (2) for General Hedging is
lower while it represents the total sample. Regression results are indicated in Table 4.
26
Table 4. Effect of Hedging on Firm Value This table presents panel regression results. Model 1 is estimated with Generalized Least Squares method on a hedging dummy that equals 1 if the firm has reported the use of derivatives for hedging in a given year. While Model 2 regresses on the total notional amount of hedging reported in a given year with Fixed Effects method. Firm Value is calculated as the natural log of [(MV Equity + Preferred Stock + Total Debt)/BV Assets]. Construction of control variables are described on p. 16&17. Year dummies are included to control for time effects. General Hedging Currency Hedging Interest Rate Hedging Commodity Price Hedging
Dependent variable: Firm Value (1) (2) (1) (2) (1) (2) (1) (2) Hedging (Dummy) 0.0063 0.5171 0.0949 -0.0960 (0.01) (0.85) (0.23) (-0.14) Amount of Hedging (Notional) 7.87e-6*** 7.51e-6** 3.13e-5** -0.0006*** (2.74) (2.41) (2.37) (-3.16) Firm Size -0.3590*** -0.5667*** -0.3879*** -0.5610*** -0.3512*** -0.5569*** -0.3571*** -0.5509*** (-4.24) (-12.22) (-4.63) (-12.66) (-3.95) (2.37) (-4.07) (-12.57) Profitability -0.5982* -0.0804 -0.6289** -0.0948 -0.6188* -0.1011 -0.5973* -0.1126 (-1.77) (-0.38) (-2.20) (-0.45) (-1.92) (-0.47) (-1.87) (-0.53) Leverage 1.5263 0.1899*** 0.0879 0.1921*** 1.4899 0.1891*** 1.4696 0.1914*** (1.46) (8.53) (0.16) (8.45) (1.51) (7.74) (1.47) (7.88) Growth Opportunities -1.3369 0.2272 -1.1952 0.2710* -1.3934 0.2524 -1.3618 0.2025 (-1.38) (1.51) (-1.10) (1.78) (-1.51) (1.60) (-1.48) (1.30) International Diversification 0.9411* 0.3235** 1.4401*** 0.2967** 0.9763** 0.1811 0.9903** 0.1643 (1.89) (2.08) (4.15) (2.01) (2.31) (1.21) (2.12) (1.22) Liquidity 0.4009* -0.0331** 0.2706 -0.0312* 0.4128* -0.0259 0.4023* -0.0265 (1.81) (-2.11) (1.12) (-1.95) (1.94) (-1.51) (1.86) (-1.55) Dividend Dummy 0.9854*** 0.1315 0.7681 0.1288 0.9854*** 0.1303* 0.9783*** 0.1323* (3.62) (1.47) (2.07) (1.59) (3.56) (1.76) (3.13) (1.79) Year Dummies Yes Yes Yes Yes Yes Yes Yes Yes Fixed Effects No Yes No Yes No Yes No Yes Intercept 0.4840 3.4469*** 1.5615*** 3.5097*** 0.4253 3.5506 0.4922 3.4947 (0.76) (12.24) (3.34) (12.52) (0.71) (12.30) (0.95) (12.79) R2 0.2539 0.2329 0.2256 0.2154 N 588 434 588 508 588 510 588 540 In parentheses are robust t-statistics. All standard errors are clustered at the firm level. The 1% outliers of Firm Value and Profitability is winsorized by replacement. R2 values are the model’s Overall-R2 (weighted average of the Within-R2 and Between-R2). R2 from GLS estimation does not have the same interpretation. Hence, R2 of model (1) are not reported. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.
27
From the regression analysis, weak evidence is found of the existence of a hedging
premium for Dutch firms during the sample period. Overall, the coefficients are positive
for both the Hedging Dummy and the Amount of Hedging except when considering
commodity hedging separately.
The coefficients of the Hedging Dummy variable represent the change in firm value
when a firm hedges compared to a firm that does not hedge. Firm value increases by
0.63%5 when a firm hedges either currency, interest rate, and/or commodity price risk.
As for currency, interest rate, and commodity hedging, the effects are 67.72%, 9.95%, and
-9.15%, respectively. These results, however, are not significant and, thus, provide no
evidence that hedging increases firm value. Moreover, results from the Hedging Dummy
can be misleading as it does not measure the exact level of hedging. To solve this issue, a
continuous variable is included measuring the notional amounts of outstanding
derivatives.
When the Amount of Hedging is considered, the coefficients indicate a slight
increase in firm value following an increase in the extent of hedging in terms of notional
amounts. In other words, the greater the hedging position, the higher the firm value. The
coefficients are significant at a 1-5% level. However, the effects seem to be very small. For
example, hedging an additional €1 million of currency exposure increases firm value by
0.00075%. The results imply that nonfinancial Dutch firms can increase their value by
hedging, but the impact is extremely small.
Furthermore, some control variable have an effect on firm value as well.
Significant results are found for firm size, leverage, international diversification, liquidity,
and dividend dummy. The signs of these coefficients seem to be logical and as expected.
Leverage, International Diversification, and Dividend Dummy are positively related to
firm value, whereas, Firm Size and Profitability are negatively related to firm value. The
negative signs for size is in line with the findings of Chun et al. (1985). They explained
that smaller firms are riskier and captured the size effect in the risk premium that justifies
higher returns for smaller firms. As expected, dividends are positively associated with
firm value. Negative signs for profitability are counter-intuitive.
5 The coefficient of General Hedging in Model (1) is 0.0063. Thus, the effect on firm value is (e0.0063-1)x100%=0.63%.
28
6 Robustness & Discussion
Several additional tests are performed to examine the robustness of our results. We test
whether the results obtained from the estimation of Models (1) and (2) are robust to the
alternative measures of firm value (FV2 and FV3). Results of the robustness tests are
presented in Appendix 10, 11, and 12. The inference of our results does not change after
running the regression models another sixteen times with different measures. Thus, the
robustness checks support the earlier results that hedging does not increase firm value.
Overall, the results are largely inconsistent with earlier studies. Allayannis and
Weston (2001), Panaretou (2014), and Bua et al. (2015) found a statistically and
economically significant hedging premium of 5%, 14%-16%, and 1.53%, respectively.
The discrepancy in the results could be explained by the different country focus as well
as the different sampling period. On the contrary, our results are most in line with Guay
& Kothari (2003) who concluded that potential gains of hedging are positive but small.
Their interpretation is that either the observed increase in firm value is driven by other
risk management activities or that the correlation is spurious. In addition, the
insignificance supports the findings of Jin & Jorion (2006) who verified that hedging
reduces the firm’s stock price sensitivity but does not affect firm value for a sample of 119
US oil and gas producers. Another reasoning suggested by Tufano (1996) is that hedging
reflects managerial risk aversion and may actually harm firm value if risk management is
costly.
Our evidence raises doubts about the conclusions of existing literature. The small
increase in firm value documented in our research could indicate that the relationship is
indeed either driven by other risk management activities or that the results are spurious.
However, given the small number of firms in our sample, the lack of significant results
could be due to the relatively low power of the empirical tests.
29
7 Conclusion
This study examines the corporate use of financial derivatives and firm value in The
Netherlands. Univariate tests and multivariate regression analyses are carried out with
panel data methodology including generalized least squares and fixed effects methods.
The sample consists of 98 public non-financial Dutch firms during the period of 2012 to
2017. Total firm-year observations of 588 are segmented into currency, interest rate, and
commodity hedgers. Similar to U.S. and U.K. studies, the use of currency and interest rate
derivatives is more common than the use of commodity derivatives. More than half of the
sample – roughly 60% – reported the use of derivatives for hedging purposes. At the
same time, a Dutch firm hedges €1.97 billion per year in total notional amounts on
average. This supports the view that Dutch firms are widely exposed to currency and
interest rate risk. Therefore, Dutch firms have high incentives to hedge. The empirical
results, however, indicate that hedging does not add value for Dutch firms. On the one
hand, univariate results indicate an average hedging discount of -21.34% for any type of
hedger compared to a non-hedger. More specifically, the average hedging discounts are -
15.09% for currency hedging, -12.68% for interest rate hedging, and -19.47% commodity
hedging. On the other hand, regression results suggest a very small hedging premium of
0.0000787% significant at a 1% level for hedging in general. This translates into a mere
0.0000787% increase of firm value following an additional €1 million notional amount
in hedging instruments. The results are furthermore robust to different measures of firm
value. Conclusively, hedgers are valued significantly lower than non-hedgers and the
value-adding component of hedging is close to zero. Hedging, therefore, does not add
value in the case of Dutch firms.
7.1 Implications and Suggestions
The above empirical results are an extension of existing research. Since prior
studies are mostly concentrated on the U.S., the main contribution of this research is the
use of a unique dataset for The Netherlands that is not readily available to investigate the
value implications of hedging. The discrepancy between this research and numerous
existing studies suggests that there are major differences between Dutch and U.S. firms.
For instance, corporate governance may indirectly affect the valuation of derivatives use.
The research topic is relevant for risk management practices as well as shareholders.
Further research is therefore required to determine whether corporate hedging should
30
be considered an important component of a firm’s risk management policy and/or
whether corporate hedging is a value-enhancing activity. Finally, while this study
provides useful insight in the hedging activity of Dutch firms, the implications are sample
specific. Future research should focus on more generalizable out-of-sample analysis in
order to assess the value implications of corporate hedging.
7.2 Limitations
Shortcomings and limitations in this research are briefly discussed in this paragraph.
First, the greatest issue in empirical Corporate Finance research is endogeneity
(Roberts & Whited, 2012). Endogeneity is present when there is a correlation between
the independent variables and the error term in a regression. This can lead to biased and
inconsistent parameter estimates making the results unreliable. Moreover, Guay and
Kothari (2003), Jin and Jorian (2004), Bua et al. (2015), and Magee (2009) raised serious
endogeneity concerns regarding firm value and hedging. Linear regressions used in this
research assume that all variables in the model are strictly exogenous. An example of
endogeneity is reverse causality. For instance, we tested the effect of hedging on firm
value. But firm value may also affect the decision to hedge because higher valued firms
have more resources and perhaps have more incentivized to use derivatives.
Secondly, research regarding the use of derivative instruments of European firms
is restricted due to data availability. Despite the time and effort put in obtaining hedging
data, the data is prone to errors from manual collection. This affects the accuracy of the
results. Moreover, due to small number of observations, no size threshold has been
implemented in the selection of the sample firms for this research. The majority of related
papers, however, excluded small firms from their sample due to lower risk exposure and
less need for hedging. However, direct costs of insolvency are widely independent on firm
size and imply that small firms should have more incentives to hedge. Dolde (1993) finds
evidence that large firms use significantly more hedging instruments, but small firms
hedge to a greater extent. Although firm size is controlled for in the regression models,
including small firms might have driven contaminated the results.
31
References
Allayannis, G., and Weston, J.P. (2001). “The Use of Foreign Currency Derivatives and Firm
Market Value”, The Review of Financial Studies 14 (1), pp. 243-276.
Allayannis, G., Lel, U., and Miller, D.P. (2012). “The use of foreign currency derivatives,
corporate governance, and firm value around.” Journal of International Economics
87(1), pp. 65-79.
Ayturk, Y., Gurbuz, A.O., and Serhat, Y. (2016). “Corporate derivatives use and firm value:
Evidence from Turkey”, Borsa Istanbul Review 16 (2), pp. 108-120.
Bartram, S.M., Brown, G.W., and Conrad, J. (2011). “The effects of derivatives on firm risk
and value”, Journal of Financial and Quantitative Analysis 46 (2), pp. 967-999.
Belghitar, S.M., Clark, E., and Mefteh, S. (2013). “Foreign currency derivative use and
shareholder value”, International Review of Financial Analysis, pp. 283-293.
Bodnar, G.M., de Jong, A., and Macrae, V. (2003). “The Impact of Institutional Differences
on Derivatives Usage: A Comparative Study of U.S. and Dutch Firms.” European
Financial Management 9, pp. 271–297.
Búa, M.V., González, L.O., López, S.F., and Santomil, P.D. (2015). “Is value creation
consistent with currency hedging?” The European Journal of Finance 21, pp. 912-
945.
Carter, D., Rogers, D., and Simkins, B. (2006). “Does Fuel Hedging Make Economic Sense?
The Case of the U.S. Airline Industry.” Financial Management 35 (1), pp. 53–86.
A9 Diagnostic Test: Fixed Effects vs Random Effects (Hausman Test)
Prob>chi2 = 0.0000
= 165.70
chi2(8) = (b-B)'[(V_b-V_B)^(-1)](b-B)
Test: Ho: difference in coefficients not systematic
B = inconsistent under Ha, efficient under Ho; obtained from xtreg
b = consistent under Ho and Ha; obtained from xtreg
y5 -.1959221 -.1439524 -.0519697 .0045523
y4 -.0905756 -.0463021 -.0442734 .0045431
y3 -.1132793 -.0567675 -.0565118 .0071788
y2 -.3074691 -.1828513 -.1246177 .0106958
y1 -.3640398 -.2539525 -.1100873 .0104416
dividend .143761 .2183241 -.0745631 .0469698
liquidity -.0205842 -.0003652 -.020219 .0036709
internatio~v .0017024 .0029484 -.001246 .0005078
growth .0102226 .0215932 -.0113707 .0036533
leverage -.0693197 -.1092891 .0399694 .004248
roa .2843394 -.2059141 .4902535 .054055
size .2108194 .5141773 -.3033579 .0242115
hedger -.0042123 .183748 -.1879603 .0665942
fixed random Difference S.E.
(b) (B) (b-B) sqrt(diag(V_b-V_B))
Coefficients
39
A10 Univariate Results
Comparison of Hedgers and Non-hedgers This table contains t-test results comparing the mean values of hedgers and non-hedgers for each control variable. Constructions of all variables are explained in detail on p. 16-17.
Hedgers Non-Hedgers Difference Firm Size 7.59 4.51 3.08*** (14.77) Leverage 1.13 1.46 -0.33 (0.96) Profitability 0.03 -0.05 0.08*** (5.24) Growth opportunities 0.05 0.76 -0.71*** (-3.24) International Diversification 51.45 36.79 14.68*** (4.33) Liquidity 0.54 1.07 -0.53*** (3.12) Dividend dummy 0.46 0.16 0.30*** (7.96) Business segment diversification 0.75 0.54 0.21*** (5.47) N 356 238 In parentheses are t-statistics of the difference in mean values. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.
40
A11 Robustness Check: Different measures (FV2) This table presents panel regression results. Model 1 is estimated with Generalized Least Squares method on a hedging dummy that equals 1 if the firm has reported the use of derivatives for hedging in a given year. While Model 2 regresses on the total notional amount of hedging reported in a given year with Fixed Effects method. Firm Value is calculated as the natural log of [(BV Assets + MV Equity – BV Equity)/BV Assets]. Construction of control variables are described on p. 16&17, Year dummies are included to control for time effects. General Hedging Currency Hedging Interest Rate Hedging Commodity Price Hedging
Dependent variable: Firm Value (1) (2) (1) (2) (1) (2) (1) (2) Hedger (Dummy) 0.2568 0.6084 0.6539 -0.4167 (0.65) (1.27) (1.44) (-0.51) Hedging (Notional) 6.34e-6** 6.15e-6** 2.81e-5** -0.0002 (2.38) (2.04) (2.31) (-1.46) Firm Size -0.3772*** -0.4893*** -0.3890*** -0.4818*** -0.3905*** -0.4812*** -0.3879*** -0.4746*** (-4.31) (-12.03) (-4.64) (-12.09) (-4.56) (-11.53) (-4.25) (-11.63) Profitability -0.5110* -0.0858 -0.5130* -0.1025 -0.5326* -0.1040 -0.4291 -0.1178 (-1.84) (-0.47) (-1.89) (-0.57) (-1.95) (-0.57) (-1.54) (-0.64) Leverage 0.5736 0.1977*** 0.5504 0.1996*** 0.5771 0.1957*** 0.3782 0.1982*** (1.18) (7.85) (1.14) (7.63) (1.20) (7.04) (0.74) (7.09) Growth Opportunities 0.4131 0.2381* 0.2245 0.2695** 0.2956 0.2566* -0.9365 0.2182* (0.52) (1.94) (0.28) (2.16) (0.38) (1.97) (-0.85) (1.69) International Diversification 1.1131*** 0.2869** 1.1281*** 0.2606** 1.2205*** 0.1617 1.1770*** 0.1487 (3.27) (2.28) (3.52) (2.18) (3.60) (1.35) (3.14) (1.38) Liquidity -0.0362 -0.0336** -0.0178 -0.0319* -0.0357 -0.0279* 0.2705 -0.0281* (-0.20) (-2.11) (-0.10) (-1.96) (-0.20) (-1.67) (1.07) (-1.67) Dividend Dummy 0.1214* 0.1144* 0.1116* 0.4120 0.1152** (1.71) (1.76) (1.92) (0.74) (2.00) Year Dummies Yes Yes Yes Yes Yes Yes Yes Yes Fixed Effects No Yes No Yes No Yes No Yes Intercept 1.7074*** 3.2784*** 1.8346*** 3.3227*** 1.7829*** 3.3680*** 1.8386*** 3.3218 (4.23) (13.01) (4.80) (12.98) (4.63) (12.70) (3.60) (12.85) R2 0.2757 0.2555 0.2444 0.2318 N 588 434 588 508 588 510 588 540 In parentheses are robust t-statistics. All standard errors are clustered at the firm level. The 1% outliers of Firm Value and Profitability is winsorized by replacement. R2 values are the model’s Overall-R2 (weighted average of the Within-R2 and Between-R2). R2 from GLS estimation does not have the same interpretation. Hence, R2 of model (1) are not reported. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. Dividend dummy variable in Model (1) is omitted due to collinearity.
41
A12 Robustness Check: Different measures (FV3) This table presents panel regression results. Model 1 is estimated with Generalized Least Squares method on a hedging dummy that equals 1 if the firm has reported the use of derivatives for hedging in a given year. While Model 2 regresses on the total notional amount of hedging reported in a given year with Fixed Effects method. Firm Value is calculated as the natural log of firm value as calculated in FV1 minus the industry median firm value. Construction of control variables are described on p. 16&17, Year dummies are included to control for time effects. General Hedging Currency Hedging Interest Rate Hedging Commodity Price Hedging
Dependent variable: Industry-adjusted Firm Value (1) (2) (1) (2) (1) (2) (1) (2) Hedger (Dummy) -0.0659 -0.4208 0.6539 -0.4167 (-0.14) (-0.69) (1.44) (-0.51) Hedging (Notional) 8.17e-6*** 7.86e-6** 2.81e-5** -0.0002 (2.75) (2.42) (2.31) (-1.46) Firm Size -0.3531*** -0.5736*** -0.3626*** -0.4818*** -0.3905*** -0.4812*** -0.3879*** -0.4746*** (-4.14) (-12.42) (-4.35) (-12.09) (-4.56) (-11.53) (-4.25) (-11.63) Profitability -0.5863* -0.0736 -0.6408** -0.1025 -0.5326* -0.1040 -0.4291 -0.1178 (-1.72) (-0.35) (-2.08) (-0.57) (-1.95) (-0.57) (-1.54) (-0.64) Leverage 1.5440 0.1935*** 1.4196 0.1996*** 0.5771 0.1957*** 0.3782 0.1982*** (1.42) (8.11) (1.45) (7.63) (1.20) (7.04) (0.74) (7.09) Growth Opportunities -1.0370 0.2228 -0.9906 0.2695** 0.2956 0.2566* -0.9365 0.2182* (-1.33) (1.49) (-0.95) (2.16) (0.38) (1.97) (-0.85) (1.69) International Diversification 0.9057* 0.3304** 0.9370** 0.2606** 1.2205*** 0.1617 1.1770*** 0.1487 (1.73) (2.13) (2.32) (2.18) (3.60) (1.35) (3.14) (1.38) Liquidity 0.4126* -0.0331 0.3322 -0.0319* -0.0357 -0.0279* 0.2705 -0.0281* (1.81) (-2.12) (1.35) (-1.96) (-0.20) (-1.67) (1.07) (-1.67) Dividend Dummy 1.0653*** 0.1347 0.9328*** 0.1144* 0.1116* 0.4120 0.1152** (3.68) (1.50) (2.94) (1.76) (1.92) (0.74) (2.00) Year Dummies Yes Yes Yes Yes Yes Yes Yes Yes Fixed Effects No Yes No Yes No Yes No Yes Intercept 0.3517 3.4505 0.4407 3.3227*** 1.7829*** 3.3680*** 1.8386*** 3.3218 (0.52) (-3.72) (0.77) (12.98) (4.63) (12.70) (3.60) (12.85) R2 0.2653 0.2555 0.2444 0.2318 N 588 434 588 508 588 510 588 540 In parentheses are robust t-statistics. All standard errors are clustered at the firm level. The 1% outliers of Firm Value and Profitability is winsorized by replacement. R2 values are the model’s Overall-R2 (weighted average of the Within-R2 and Between-R2). R2 from GLS estimation does not have the same interpretation. Hence, R2 of model (1) are not reported.
***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.