Dutch firms
Author: Christiaan Bekker Student Number: 1436163
University of Groningen Faculty of Economics and Business Master’s Thesis, MSc Business Administration Specialization: Corporate Financial Management
Supervisor: Dr. P.P.M. Smid
Dutch firms
Abstract
The majority of capital structure studies so far have focused on either the pecking order theory or the static trade-off theory. Although there exists some consensus that these two theories partially explain capital structure developments within firms, empirical evidence does not yield unambiguous results about the validity of these theories. Even less in favour of these theories, recent research has indicated that capital structure changes can be explained by equity market timing. Equity market timing implies that firms prefer equity financing when the relative cost of equity is low and prefer debt otherwise. This study examines whether market timing provides empirical explanation for capital structures of Dutch listed firms for the time period 1998 till 2007. In doing so, I examine the effect of market timing on capital structure changes in the short run and in the long run, while controlling for the pecking order theory and static trade-off theory. The findings imply that Dutch firms indeed time the equity market in the short run, and that the static trade-off theory and, to a certain extent, the pecking order theory also provide empirical explanation for capital structure changes in the short run. In the long run however, I find no evidence of market timing behaviour. Tests over a longer time horizon indicate that firms’ financing decisions are consistent with the trade-off theory and also some (limited) evidence is found in favour of the pecking order theory. The findings indicate that, over a longer time period, overleveraged Dutch firms bring back their leverage towards a target debt ratio and more profitable firms prefer internally generated funds over debt.
JEL Code(s): C23, G32
PREFACE
This Thesis has been written as final part of my MSc Corporate Financial Management. After several interesting study years, this thesis is the last hurdle that I had to take for finishing my studies. I started my research in February 2009 and finished it in August 2009. The primary focus of this study is on equity market timing and the first and foremost goal of this thesis is to contribute to the academical literature. However, I also think that this thesis is of interest to people from the ‘field’ because it adresses real-world decisions that managers make: whether they issue additional capital at times when the stock market valuation of their firm is high relative to their intrinsic value.
I would like to thank my supervisor dr. Smid for giving me good advices, his patience and, maybe the most important aspect, for forcing me to completely dig into this topic. I also want to thank my family, friends and girlfiriend, for their continued support.
TABLE OF CONTENTS
1. Introduction ... 5
2. Literature Review ... 8
2.1 Pecking order theory ... 8
2.2 Static trade-off theory... 9
2.3 Market timing theory... 10
2.4 Empirical evidence ... 11
3. Model specification and methodology ... 16
3.1 Model Specification ... 17
3.2 Variables... 20
3.2.1 Dependent variables... 20
3.2.2 Independent variables ... 20
3.3 Hypotheses ... 24
4. Data and Summary Statistics... 26
5. Results ... 30
5.1 Market timing in the short run... 30
5.2 Market timing in the long run... 33
6. Conclusion... 36
7. References ... 39
APPENDIX 1: Variable definitions ... 42
APPENDIX 2: Industry Specification... 43
APPENDIX 3 :Correlation Table... 44
APPENDIX 4: Predicting the target debt ratio... 45
APPENDIX 5: Correlated Random Effects- Hausman Test ... 46
APPENDIX 5: Correlated Random Effects- Hausman Test ... 47
APPENDIX 6a: Alternative model specifications for short run regressions... 48
1. Introduction
About 50 years ago, Modigliani and Miller (1958) illustrated that under special circumstances, including perfect and frictionless capital markets and the absence of taxes, the value of a firm only depends on the level and risk of its future cash flows. According to this ‘irrelevance theory of capital structure’ there does not exist an optimal capital structure because a firms’ value can not be affected by its choices of financing.
One of the more intriguing challenges in corporate finance however, is to provide an explanation why some firms finance incremental investments with debt while other firms use equity. Over time two main theories have emerged that diverge from Modigliani and Millers’ (1958) ‘irrelevance theory of capital structure’: the pecking order theory and the static trade-off theory. The pecking order theory of capital structures (Myers, 1984; Myers and Maljuf, 1984) states that firms establish their capital structures by following a financing hierarchy: firms will firstly use internally generated funds, then issue debt and as a last resort firms will issue equity. This financing hierarchy results mainly from the problem of asymmetric information between managers and shareholders, which makes external funds more expensive than internally generated funds. The static trade-off theory (Myers, 1984) on the other hand, predicts that firms have an optimal capital structure by trading off the benefits and the costs of debt and equity financing, such as tax shields, financial distress, and agency costs.
(i.e. trade-off) behaviour. In so doing, they aim to obtain a better understanding of the effect of market timing on capital structures. They find that the market timing theory, pecking order theory and capital structure targeting all provide explanations for firms’ capital structures in the short run. In the long run however, firms move towards target debt ratios consistent with the trade-off theory of capital structures.
Up till now, existing studies that test the market timing theory have been predominantly US based. To the best of my knowledge, there is only one study that examines the market timing theory of capital structures for Dutch listed firms, executed by de Bie and de Haan (2007). They find evidence that firms with higher stock returns are more likely to issue equity than debt, which is consistent with the predictions of market timing behaviour. However, they do not find evidence that market timing has a persistent, long lasting effect on Dutch capital structures. Whereas de Bie and de Haan (2007) examine Dutch listed firms from 1983 till 1997, the aim of this study is to test whether the market timing theory of capital structures holds for Dutch listed firms for the period 1998 till 2007. It is interesting to test whether the market timing theory holds for this particular period because it is conceivable that the financing behaviour of Dutch firms has become different during this time period under investigation than in earlier time periods. In this period the Dutch financial system has become more market oriented. According to studies of De Nederlandsche Bank (De Nederlandsche Bank Quarterly Bulletin, September 2006), the volume of share issues increased strongly in the mid 1990s relative to previous years (see figure 1). Furthermore, the study of the Nederlandsche Bank (De Nederlandsche Bank Quarterly Bulletin, September 2006) also points out that the Dutch stock market has become more internationally open to foreign investors during this period. These facts indicate that equity financing has become a more important source of financing in the period under investigation in this study than in the study of de Bie and Haan (2007).
Figure 1: Public share issues and the level of the AEX-index from 1986 till 2004
source: De Nederlandsche Bank Quarterly Bulletin, September 2006
In the light of this, it is an interesting question whether firms are more inclined to time the equity market during this period than in earlier time periods. To this end, the main question of this study is:
Do Dutch listed firms time the equity market from the period 1998 till 2007?
If Dutch firms indeed time the equity market, firms with high stock market valuations will prefer equity financing over debt financing. Therefore, the most important subquestions that I examine to answer the research question are 1) whether past stock returns have a negative relation with leverage changes and 2) whether historical market-to-book ratios (interacted with firms’ capital needs in corresponding years) have a negative relation with leverage changes. In order to answer the main question I also control for the fact that the pecking order theory might provide empirical explanation for capital structure changes of Dutch firms. In line with predictions of the pecking order theory, I examine whether firms’ capital needs (i.e. financial deficits) are positively related to changes in leverage and whether profitability is negatively related to leverage changes. Finally, I also control for the static trade-off theory by examining the relationship between changes in leverage and deviations from optimal debt ratios. The last subquestion of this study is whether this relation is negative, as is predicted by the static trade-off theory.
I will adopt the methodology of Kayhan and Titman (2007) using accounting data of financial statements extracted from Datastream. The dependent variable is the change in leverage, both in market and book values, and the independent variables are proxies for market timing, pecking order and trade-off behaviour. Following Kayhan and Titman (2007) and de Bie and Haan (2007), I will run regressions for both a short and long time horizon. In the short run period I will examine changes in leverages over a four year period and in the long run I will examine changes in leverage over an eight year period.
2. Literature Review
In this section I will discuss the most important theories and empirical evidence about capital structures. Although the focus of this thesis is on the market timing theory, I will also discuss the pecking order theory and the static trade- off theory in order to position the market timing theory in existing capital structure theories.
2.1 Pecking order theory
The foundation of the pecking order theory goes back to Donaldson (1961), who was the first to observe that firms finance incremental investments according to a hierarchy: firms prefer internally generated funds over debt issues, and prefer debt issues over equity issues. Following the idea of Donaldson (1961), Myers (1984) and Myers and Maljuf (1984) called this the pecking order theory, where firms establish their capital structures by following this hierarchy.
They offer an explanation for this ‘pecking order’ of financing decisions from the perspective of asymmetric information. Insiders (managers) are supposed to know more about the value of a firm’s assets and growth opportunities than outside investors. When faced with an investment decision, managers are therefore willing to issue equity when a firm is overvalued but want to avoid this when the firm is undervalued. Investors are aware of the fact that firms tend to issue equity when their shares are overvalued. For this reason, they will anticipate by rationally adjusting the price they are willing to pay for new shares downwards, causing the new shares to be underpriced. Likewise, managers will be reluctant to undertake a new project if they have to finance this project with equity, especially when all benefits from this project will accrue to new shareholders. New investors will consider an equity issue as unfavorable information about the firm’s current value. This will reduce the price they are willing to pay for new shares. However, when new projects are mistakenly assessed by the market, leading to undervaluation of new shares, this will consequently transfer wealth from current shareholders to new shareholders. In order to avoid underpricing of new equity and to avoid distortions of investment decisions, managers prefer sources that are less exposed to asymmetric information. Therefore, they will firstly finance incremental investments with internally generated funds which are not exposed to asymmetric information problems. When internally generated funds are not enough to cover the investment, firms will issue low risk debt, for which asymmetric information problems are negligible, and subsequently issue risky debt. Firms will only issue equity as a last resort, when incremental investments could not be covered by internally generated funds, or debt.
2.2 Static trade-off theory
The static trade-off theory of Myers (1984), states that firms establish their capital structures by trading off the benefits and the costs of debt and equity financing. This trade-off will finally result in an optimal capital structure for each firm. The trade-off theory originates from the debate that started after the work of Miller and Modigliani and Miller (1963), which was a revised version of their earlier ‘irrelevance theory of capital structure’ (Modigliani and Miller, 1958). This theory, which incorporates income taxes, takes into account that interest payments associated with debt are tax deductible. This creates an advantage for debt financing because it serves to shield earnings from taxes. Under this circumstance, firm value is not independent of a firm’s capital structure because debt financing increases firm value. However, this does not imply that 100% debt financing maximizes firm value. Debt financing also has a main disadvantage that often will arise if a firm relies to heavily on debt: the cost of possible financial distress. This includes both direct costs from bankruptcy, such as legal costs, and indirect costs that arise if bankruptcy is avoided, but the efficiency of the way the firm is managed is reduced. This leads to a first trade-off, namely that firms’ capital structures will reflect an optimal leverage ratio that is the result of a trade-off of tax benefits and a higher risk of financial distress.
shareholders, do not undertake positive net present value projects because the benefits from these projects will accrue mainly to existing debt holders.
Taking into account agency costs in the static trade-off theory means that firms do not only assess the benefits of tax shields and costs of potential financial distress when they establish their capital structures. Firms also trade off benefits and costs of debt and equity financing from agency based considerations. Under the static trade-off theory many other factors have been identified that firms consider when choosing their optimal capital structure. However, tax benefits, the costs of financial distress and agency costs can be identified as the most important factors.
A very important assumption of the static trade-off theory is that firms have a target capital structure, which reflects their optimal capital structure. Firms might move away from their optimal capital structure in the short run, but they will gradually move back towards their target capital structure by adapting their financing policies in a way that will lead to this optimal level.
2.3 Market timing theory
The market timing theory states that capital structures are determined by fluctuations in share prices and market valuations. Firms issue shares when their shares are overvalued and repurchase shares if their shares are undervalued. Likewise, firms faced with financing decisions prefer debt when their share price is low and prefer equity issues otherwise.
The existence of market timing behaviour can be explained from two different theoretical views. The first theory takes as starting point that agents are rational and that there exists an informational asymmetry between firms’ managers and its shareholders. This informational asymmetry is time varying, because the degree of informational asymmetry is not fixed over time. For example, a release of an annual report or earnings statement inform shareholders about the firms’ performance and will temporarily decrease the amount of information asymmetry. Share price declines from equity issues will be lower when information asymmetry is relatively low and share prices will usually go up after positive information releases. Therefore, firms will issue equity immediately after a positive information release that reduces the information asymmetry between shareholders and the firms’ management. In doing so, firms create their own timing opportunities (Lucas and McDonald 1990; Korajczyk, Lucas and McDonald, 1992).
for managers’ perceptions of the shares to be mispriced, because they determine their notion of under-or overvaluation by comparing the current market to book ratio with market to book ratios in surrounding periods. A temporary high market to book ratio relative to surrounding periods is considered to be a short-term overvaluation, whereas a temporary low market to book ratio relative to surrounding periods is considered as a short-term undervaluation. According to the market timing theory, managers will exploit a short term overvaluation in the benefit of the company by issuing equity instead of debt when they are faced with a financing decision.
2.4 Empirical evidence
The empirical literature finds evidence for both the pecking order theory and the static-trade-off theory. Most of the empirical tests which have been constructed to examine whether the pecking order theory, or the static trade-off theory provide an explanation for firms’ capital structures, yield mixed results. Shyham-Sunder and Myers (1999) have constructed an empirical test that relates a firms’ financial deficit to its net debt issues. A firms’ financial deficit is the amount of capital that a firm comes short in a given year, and is therefore the amount of capital that a firm actually raises in that year. The pecking order theory predicts that firms will issue debt when they can not cover its investments or dividend payments with internally generated funds. For this reason, Shyham-Sunder and Myers (1999) argue that a firms’ financial deficit is expected to have a positive relation with leverage. Shyham-Sunder and Myers (1999) find evidence of this relationship and conclude that the pecking order theory is superior in explaining financing choices with regard to debt or equity. Based on a sample of 157 US firms with continuous data over the period 1971-1989, they find that, although some equity was issued during this period, debt issues considerably dominate equity issues. Frank and Goyal (2003) perform a similar test with a substantial larger sample. They examine the period 1971-1998 and do not require that continuous data is available for all firms. Based on their findings they conclude that, contrary to the predictions of the pecking order theory, equity issues considerably dominate the financing decisions of firms. Fama and French (2002) also test the pecking order theory against the static trade-off theory and find support for both theories of capital structure. Consistent with the pecking order theory, more profitable firms are less indebted. However, they also find that firms with more investment opportunities are less indebted. This is in line with the static trade-off theory because firms with more investment opportunities should need less discipline of debt to reduce agency problems.
subsequently they prefer share issues and lastly, bond issues. According to the authors, share issues in the Netherlands are preferred over bond issues due to the fact that the Dutch corporate bond market is relatively underdeveloped. De Jong and Veld (2001) study capital structures of Dutch firms for the period 1977 till 1996. Their findings are that Dutch firms establish their capital structures more in line with the static trade-off model. They do not find evidence for the pecking order theory.
be timed to conditions in the equity market, but they do not have a long lasting effect on capital structures. Debt transactions do not exhibit timing patterns that are likely to lead to a negative relationship between market-to-book ratios and leverage. Hovikimian (2003) also finds that past average market-to-book ratios have a significant effect on current investment decisions and financing. This implies that historical market to book ratios contains information about growth opportunities, which is not captured by current market-to-book ratios. Kayhan and Titman (2007) include market timing variables in regressions of conventional capital structure theories, that is, the pecking order theory and capital structure targeting (i.e. trade-off) behaviour. They split the market timing variable of Baker and Wurgler (2002) into two parts: a short term and a long term measure of market timing. They do this because they argue that the market timing measure of Baker and Wurgler is biased (I will elaborate on this in detail in the methodology section). Their first market timing measure, which they call Yearly Timing, is constructed by interacting past market-to-book ratios to financial deficits. This captures the idea of the timing measure of Baker and Wurgler (2002), by taking high values for firms that raised relatively much external finance, both equity and debt, when their market-to-book ratios were high. In line with Baker and Wurgler (2002), Kayhan and Titman (2007) argue that this variable must have a negative relation with leverage, because firms that raise capital when their market valuation is relatively high will, in line with the market timing theory, issue equity instead of debt. Their second timing measure, which they call Long Term Timing, is constructed by taking the product of average past financial deficits and the average of past market to book ratios. Kayhan and Titman (2007) argue that this measure has nothing to do with market timing but captures the idea that market to book ratios also proxy for firms’ growth opportunities. In this case firms can choose not to issue debt, for other reasons than market timing, but to maintain financial flexibility (Myers, 1976). Consistent with Baker and Wurglers’ (2002) findings, they find evidence that capital structures are driven by market timing (along with pecking order and trade-off behaviour) in the short run. In the long run however, they do not find evidence that market timing has an effect capital structure changes. They conclude that firms in the long run move towards target debt ratios consistent with the trade-off theory of capital structures.
timing behaviour. However, they also do not find evidence that market timing has a persistent, long lasting effect on Dutch capital structures. Table 1 provides a table of the most relevant empirical literature for this particular study.
Table 1
Overview of previous empirical evidence on capital structures.
Authors Key Question Data Methodology Main Conclusion
1. Shyham-Sunder & Myers (1998) Testing Pecking Order Theory vs Static Trade-off Theory Panel of 119 US firms using balance sheet data from 1971-1998 Continuus data is not required.
Empirical test that relates net debt issues to the financial deficit and to deviations of the current debt ratio from the target debt ratio. OLS-regressions are employed.
Pecking Order Theory offers better explanation in debt-equity choice than Trade-off Theory. 2. de Jong and Veld (2001) Debt-equity choice by examining Static Trade-off Theory and adverse selection model. 110 issues of public and private
seasoned equity and 137 public issues of straight debt made by Dutch firms from 1977 till 1996.
Logit analysis for the debt-equity choice. Logistic regressions in which the dependent variable takes the value of one for equity issues and zero for debt issues. Debt-equity choice is better explained by Static Trade-off model, than by adverse selection model. 3. Baker and Wurgler (2002) Which theory explains capital structure developments: Pecking Order,Static Trade-off, or Market Timing.
US IPO firms from 1968 till 1999.
Testing the relation between equity issues and external-finance-weighted average market-to-book ratios, current market-to-book ratios, asset
tangibility, profitability and firm size by means of OLS.
Market Timing has a persistent effect on capital structure changes. 4. de Haan and Hinloopen (2003) Testing Pecking Order and Trade-off measures. Also, determining firms' preference in their financing hierarchy. Panel of 150 Dutch firms from 1984 till 1997.
First, a multinomial logit model is estimated to test predictions of the Pecking Order theory and Static Trade-off theory. Then,a Probit model is estimated to determine which financing hierarchy fits the data best.
Dutch firms follow more or less a Pecking Order hierarchy: 1)internal finance,2) bank loans, 3) share issues, 4) bond issues.
Authors Key Question Data Methodology Main Conclusion 5. Frank and Goyal (2003) Testing Pecking Order Theory. Panel of 768 US firms using balance sheet data from 1971-1998 Continuus data is not required.
Empirical test that relates net debt issues to the financial deficit. OLS-regressions on panel data.
Pecking order Theory rejected. Net equity issues track the financial deficit more closely than net debt issues.
6.Kayhan and Titman (2007 Examining the Market Timing Theory,while controlling for Pecking Order financing and Trade-off theory. US firms from 1971till 2002.
Market Timing is examined by testing the relation between stock returns yearly timing, long term timing, profitability, financial deficits and deviation from target leverage with debt ratios. First stage: a Tobit regression to determine the target debt ratio. In the second stage GLS with bootstrapped standard errors.
In the short run, Market Timing and financial deficits influence capital structure changes most. In the long run, firms move back to a target debt ratio, consistent with the Static Trade-off theory.
7.de Haan and de Bie (2007) Testing the Market Timing Theory, while controlling for Pecking Order Theory, and Static Trade-off Theory. Sample f 135 Dutch listed firms and subsample of 37 IPO-companies from 1983 till 1997
Same methodology as Kayhan and Titman (2007). See previous box.
3. Model specification and methodology
3.1 Model Specification
In order to test the market timing theory in the short run, I examine how proxy variables for the market timing theory, pecking order theory and capital structure targeting taken over a 4 year period from years t-4 till year t affect changes in leverages over a four year period between year t and t-4. In the long run I examine how proxy variables for the market timing theory, pecking order theory and capital structure targeting over the 4 year period from year t-8 till t-4 affect changes in leverage over the entire eight year period between year t and t-8. Although Kayhan and Titman (2007) examine changes in leverage over 5 years and 10 years respectively, I follow de Bie and Haan (2007) who also use 4 years and 8 years in order to save observations1.
The empirical analysis involves two stages of regression estimations. In the first stage I predict the target leverage (as I will discuss in section 3.2.2.2) by regressing the debt ratio on a set of variables that have been found to be the most important determinants of leverage in previous studies. I lag the variables one period in order to reduce possible endogeneity problems (Rajan and Zingales, 1995; Kayhan and Titman, 2007; de Bie and Haan, 2007). The regression is estimated with a Tobit regression to take into account that the target debt ratio is restricted to lie between 0 and 1, further details are provided in appendix 4. Subsequently, I use the residuals of this regression estimation as ‘deviation from the target leverage’ variable (DEVTAR) in the further analysis. In stage 2 I include this variable, constructed in stage 1, and estimate the regression model together with our other variables.
Stage 1: estimate the target leverage
t t
t t
t
t PROF M B Size Tang IndustryDummy
L
1 1
2 / 1
3 1
4 1
5
1) I construct the ‘deviation from target leverage’ measure DevTar at t-4 by taking the difference between the actual debt ratio and the predicted value of the debt ratio:4 4 4
ˆ
t
t tL
L
DevTar
2)Where the dependent variable Lis estimated for book leverage and market leverage, PROF stands for profitability, M /Bis market-to-book ratio, Size is firm size and Tang is tangibility,DevTaris
the deviation from target leverage andLˆis the predicted target leverage.
1
Stage 2: estimate the regression model. it it it it it it it it it it it it it it it it mmy IndustryDu PROF dumFD FD DevTar LT YT n Stockretur L L
8 4 , 7 4 , 6 4 , 5 4 4 4 , 3 4 , 2 4 , 1 4 3)Where i denotes the individual firm and t denotes the year,
is an industry specific constant,Lit Lit4is the dependent variable and is the change in leverage (both market leverage and book leverage) between year t and year t-4, Stockreturnit,it4is the 4 year cumulative stock log returnbetween year t and year t-4, YTit,it4is the covariance between the market to book ratio and the
financial deficit from year t to year t-4, LTit,it4is the product of the average market to book ratio and
the average financial deficit (FD) between year t and year t-4, DevTarit4 is the difference between the actual debt ratio and the target debt ratio at year t-4, where the target debt ratio is approximated by the predicted value of the debt ratio (see appendix 4), FDit,it4is the total external financing from year
t till year t-4, dumFDit,it4 is a dummy variable that interacts with the financial deficit and takes the
value of one when FD is positive and zero when FD is negative, PROFit,it4 is the 4 year cumulative
profitability between year t and year t-4 and
itis the error term.In order to test whether market timing has a persistent effect on capital structure changes I also estimate whether the proxy variables over the 4 year period from t-8 till t-4 affect changes in leverage over the entire eight year period between year t and t-8. Again the empirical analysis involves two stages, where I estimate the target leverage and construct the ‘deviation from target leverage’ variable
(DevtTar) in stage 1 and I estimate the complete equation in stage 2.
Stage 1: estimate the target leverage
t t
t t
t
t PROF M B Size Tang IndustryDummy
L
1 1
2 / 1
3 1
4 1
5
, 4) I construct the ‘deviation from target leverage’ measure DevTarat t-8 by taking the difference between the actual debt ratio and the predicted value of the debt ratio:Where the dependent variable Lis estimated for book leverage and market leverage, PROF stands for profitability, M /Bis market-to-book ratio, Size is firm size and Tang is tangibility, DevTaris
the deviation from target leverage andLˆis the predicted target leverage.
Stage 2: estimate the regression model.
it it it it it it it it it it it it it it it it mmy IndustryDu PROF dumFD FD DevTar LT YT n Stockretur L L
8 8 , 4 7 8 , 4 6 8 , 4 5 8 4 8 , 4 3 8 , 4 2 8 , 4 1 8 6)Where i denotes the individual firm and t denotes the year,
is an industry specific constant.Lit Lit8 is the dependent variable and is the change in leverage (both market leverage and book leverage) between year t and year t-8,Stockreturnit it4, 8is the 4 year cumulative stock logreturn between year t-4 and year t-8, YTit it4, 8is the covariance between the market to book ratio and
the financial deficit from year t-4 to year t-8,LTit it4, 8is the product of the average market to book
ratio and the average financial deficit (FD) between year t-4 and year t-8, DevTarit8 is the difference between the actual debt ratio and the target debt ratio at year t-8, where the target debt ratio is approximated by the predicted value of the debt ratio (see appendix 4), FDit it4, 8is the total financial
deficit between year t-4 and year t-8, dumFDit it4, 8 is a dummy variable that interacts with the
financial deficit and takes the value of one when FD is positive and zero when FD is negative,
8 ,
4
it it
PROF is the 4 year cumulative profitability between year t-4 and year t-8 and
itis the error term.consistent fixed effects estimator (Brooks, 2002). Appendix 4 presents the outcomes of these test and shows that the use of random effects is justified, since I cannot reject that random effects estimation is appropriate in all the regressions. To correct for biases from using multiple observations and moving lagged variables, I use White standard errors robust to heteroscedasticity and adjusted for correlations of error terms across observations.
3.2 Variables
3.2.1 Dependent variables
The dependent variable of this study is the change in leverage during the period 1998 till 2007. As discussed, I follow de Bie and Haan (2007) by examining leverage changes over a 4 year period for the short run regressions and by examining leverage changes over an 8 year period for the long run regressions. Leverage is measured in market values (market leverage) and in book values (book leverage). I take these two measures of leverage because most capital structure theories have not specified explicitly which leverage measure should be taken, and most empirical studies about capital structures, also Kayhan and Titman (2007), use both measures.
3.2.2 Independent variables
3.2.2.1 Market Timing
As we haven seen in section 3.1, the first proxy to account for market timing is the stock return, measured as the cumulative 4 year log return on the stock between year t and t4. The stock return variable is included to account for the direct impact of stock market valuation on changes in the debt ratio. First, there is a direct and mechanical effect of stock prices on market leverage. Since market leverage is expressed in market values, it may go down when stock prices rise, even if no equity is issued. Second, as stated by Welch (2004), positive stock returns are expected to have a negative relation with market leverage, if firms do not rebalance their debt ratio to a target ratio after periods of stock price increases or decreases. If we would find, next to market leverage, also a negative relationship between cumulative log returns and book leverage this provides additional evidence that firms issue equity when their market valuation is relatively high.
The second proxy for market timing is based on the market timing measure as constructed by Baker and Wurgler (2002). They construct a variable which they call the external-finance-weighted average market to book ratio (EFWAMB ) and is defined as follows2:
2
s t s t r r s t
M
B
FD
FD
EFWAMB
(
/
)
1 1 1 1
7)Where FD stands for the financial deficit of the firm, M /Bis the market-to-book ratio and the suffix
s
and rdenote time. The financial deficit,FD, is the amount of external capital that a firm actually raises in a given year and is defined as net debt issues plus net equity issues.In equation 7, accumulates firms’ financial deficits over a number of past years. In the study of Baker and Wurgler (2002), they accumulate all the years since the initial public offering (IPO) of the firm. A firm at time t has an external-finance-weighted average market-to-book ratio, which is a weighted average of past market to book ratios beginning at the year of the initial public offering (IPO) till yeart1. The weight for a particular year is the financial deficit of that year divided by the total financial deficit that has accumulated since the initial public offering (IPO). This variable is expected to have a negative relation with leverage because high market to book ratios should correspond with higher equity issues than debt issues. As we read in the literature review, Baker and Wurgler (2002) find evidence for this negative relationship between leverage and their external-finance-weighted average market-to-book ratio. However, Kayhan and Titman (2007) criticize the external-finance-weighted average market-to-book ratio of Baker and Wurgler (2002). They split the timing measure of Baker in Wurgler (2002) into two components to support their critical comment3:
)
/
(
)
/
,
cov(
)
/
(
1 0 1 0 1M
B
FD
B
M
FD
B
M
FD
FD
EFWAMB
s t s t r r s t
8)The first term of the new equation is the covariance between the financial deficit and the market to book ratio, scaled by the average financial deficit. This term captures the idea of Baker and Wurgler (2002) that a firm which raises external capital when its valuation is relatively high is likely to decrease its debt ratio. The intuition is that managers will issue equity when the firm is temporarily overvalued, where the concept of under-or overvaluation is determined by a firms’ current market to book ratio relative to surrounding years. The second term is a firms’ average market to book ratio. Kayhan and Titman (2007) argue that this term has nothing to do with market timing. As we read in the literature review, high market- to- book ratios are often associated with high growth opportunities (Myers, 1977). For this reason, firms with high market-to-book ratios that decrease their leverage ratio could exhibit such behavior not because of market timing, but because they are reluctant to finance
3
their projects with debt. Hence, this term of the equation can be interpreted as a control variable for growth opportunities. In order to arrive at their final measures, Kayhan and Titman (2007) multiply
both terms at the right hand side of equation 8) byFD. This leaves them with two separate variables, of which the first term is called Yearly Timing measure:
) / , cov(FD M B YT 9)
and the second term they call Long Term Timing measure:
FD B M
LT ( / ) 10)
Following de Bie and Haan (2007), both measures will each time be calculated over the years t to 4
t .Note that the difference between stock return and Yearly Timing (YT) as market timing variable is that our stock return variable measures stock market valuation directly in terms of changes, Yearly Timing (YT) measures stock market valuation in levels, since this measure interacts stock market valuation (measured by the market-to-book ratio) with the amount of capital that a firm actually raises during a particular year (Financial Deficit).
3.2.2.2 Capital structure targeting
As I have explained in detail, an implication of capital structure targeting is that firms move towards their optimal target debt ratio. In this case, firms that have debt ratios higher than their target debt ratio will lower their debt ratio in the coming period, and firms that have debt ratios lower than their target debt ratio will increase their leverage. The proxy variable used in this study to account for capital structure targeting is called ‘deviation from target leverage’. This variable is defined ‘deviation from target leverage’ because it measures the difference between a firms’ actual debt ratio and its target level.
The proxy variable used in this study to account for capital structure targeting is called ‘deviation from target leverage’ (DevTar). This variable is defined ‘deviation from target leverage’
(DevTar), because it measures the difference between a firms’ actual debt ratio and its target level.
borrow from banks. In an imperfect world with asymmetric information about a firms’ ability to repay its debt, firms with little tangible assets will find it difficult to raise much debt. This predicts a positive relation between asset tangibility and the target debt ratio. The second variable, firm size (Size), is also expected to have a positive relation with leverage. The most important reason for this is that large firms are less exposed to asymmetric information. If the public knows more about a firm, this will make it easier for this firm to raise debt. Another reason is that larger firms will be more able to diversify into different investment projects than smaller firms, thereby limiting their risk of temporary bad results from one investment project. This means that large, diversified firms are less likely to face the risk of financial distress. The third variable that Rajan and Zingales (1995) identify as an important determinant of leverage, profitability (Prof), does not have a clear cut effect on leverage. Different authors have expressed different views as to the effect of profitability on leverage. On one hand, the pecking order theory predicts that more profitable firms will issue less debt, as these firms will be able to finance more investment projects with retained earnings. On the other hand, in the models of Ross (1977) and Leland and Pyle (1977), more profitable firms issue more debt because this serves as a signal of good firm performance and confidence in future performance. This line of reasoning predicts a positive relation between profitability and debt. The last variable is a firms’ market to book ratio (M/B). According to Rajan and Zingales (1995) the relation between market to book ratios and leverage may stem from perceived mispricing. If firms issue stock when their price is relatively high compared to their book value, we might expect a negative effect between leverage and a firms’ market-to-book ratio. Another explanation why market-to-book ratio is an important determinant of leverage is because it is considered to be a proxy of firms’ growth opportunities. In this case, market to book ratio has a negative relation with leverage because firms want to keep their financial flexibility (Myers, 1977).
I include industry dummies to control for industry specific effects (appendix 2 provides an overview of the industry specification of firms in our sample). The target capital structure is approximated by regressing leverage (both market leverage and book leverage) on these proxy variables and industry dummies. The residuals from this regression are used as a measure for the ‘deviation from target leverage’ (DevTar) in the further analysis. This procedure is common in capital structure studies (Kayhan and Titman, 2007; de Jong and Veld, 2001; de Haan and Hinloopen 2003). Appendix 4 provides the results of this regression and a further detailed explanation.
3.2.2.3 Pecking order financing
approach’ is also adopted by Baker and Wurgler (2002) and Kayhan and Titman (2007). When the financial deficit variable is positive, the firm needs more funds than it internally generates and when this variable is negative the firm is generating more internal funds than it invests. The pecking order theory predicts that firms with high financial deficits increase their debt ratios, because debt is preferred over equity. Following de Bie and Haan (2007), the proxy variable in this study is the cumulative financial deficit between year t and year t-4.
Kayman and Tittman (2007) point out that a financial deficit or a financial surplus might have different effects on leverage. For example, asymmetric information problems might be larger for a firm that issues shares than for a firm that repurchases shares. Therefore, positive financial deficits can have different effects on capital structure changes than negative financial deficits. To account for this, I introduce a dummy variable (FDdum) that takes the value of one when the financial deficit is positive and a zero when the financial deficit is negative.
The second proxy variable for pecking order financing is profitability. The pecking order theory predicts that profitable firms will be more able to finance projects with internally generated funds, implying a negative relationship between profitability and leverage. Profitability is defined as earnings before taxes, scaled by total assets at the beginning of each year. The proxy variable for profitability in this study is the cumulative earnings before taxes between year t and year t-4. Following De Bie and Haan (2007) and Kayhan and Titman (2007), in the regressions for book leverage I scale profitability by the book value of total assets and in the market leverage regressions I scale profitability by the market value of assets. All definitions of the variables from section 3.2 are given in appendix 1.
3.3 Hypotheses
From the empirical literature we know that a negative relation between stock returns and leverage provides evidence of market timing because this implies that firms are prone to reduce their debt ratio, i.e. issue equity, when their market valuation is relatively high. Our second market timing measure, Yearly Timing, is also expected to have a negative relationship with leverage because this implies that firms that raise external capital when their market valuation is relatively high, are likely to use equity financing. If market timing has a persistent effect on capital structure changes, we expect these relations to hold in both the short and the long run. This leads to the following explicit hypothesis:
1) 0a:
1 0 0 : 11
2) 0b :
0 0 : 1 b
3) 0c:
2 0 0 : 2 1 c
4) 0d :
2 0 0 : 2 1 d
4. Data and Summary Statistics
The sample consists of Dutch listed firms, from the period 1998 till 2007. The data is extracted from Thomson Datastream, which contains financial statements of Dutch listed firms. When data in Datastream is missing, I consult the ‘Jaarboek van Nederlandse Ondernemingen’, which publishes financial statements of Dutch listed firms till 2000. For missing data in the other years I collect the relevant data from annual reports of firms under consideration if it is still available on their websites. Companies are included in the dataset if there are at least eight consecutive years of data available. I exclude financial companies due to their different balance sheet structure. This leaves me with a final sample which consists of 95 firms and covers 1045 firm-year observations (since data from 1997 is also used in order to construct variables for 1998, I have 11*95 firm-year observations). From this data I construct a panel, on which the regressions will be run. In order to compare this study with existing US studies, the variables follow the definitions of US studies (Baker and Wurgler, 2002; Kayhan and Titman, 2007) as closely as possible, as I also did for the methodology. This procedure is also followed by de Bie and Haan (2007) and makes it possible to compare our findings with findings for the Netherlands executed for an earlier time period. Appendix 1 presents definitions of all variables used in this study and includes a detailed explanation of the variable construction. It also provides an exact specification of the moment/date on which the variables are measured. In general, a book year is from January till December. There are a few firms that have book years that start in April or July.
Book leverage is defined as the total book value of debt divided by the sum of total debt and the book value of equity. Market leverage is defined as total debt divided by the sum of total debt and the market value of equity (Kayhan and Titman, 2007; de Bie and Haan, 2007). Net equity issues are defined as the change in the book value of equity minus the change in retained earnings, divided by the book value of total assets. Net debt issues are defined as the change in the book value of total debt divided by total assets (Kayhan and Titman, 2007; de Bie and Haan, 2007). Asset tangibility (Tang) is defined as the ratio of firms’ fixed assets to total assets (Rajan and Zingales, 1995; Kayhan and Tittman, 2007; de Bie and Haan, 2007). Firm size is defined as the logarithm of total sales (Rajan and Zingales, 1995; Kayhan and Titman, 2007; de Bie and Haan, 2007). We take earnings before taxes (EBT), divided by total assets as proxy of profitability. Market to book ratio is defined as the book value of total assets + market value of equity – book value of equity, divided by the book value of total assets (Kayhan and Titman, 2007; de Bie and Haan, 2007). As mentioned, appendix 1 describes the variable construction more extensively and appendix 2 also gives a specification of the different industries in which the firms of our sample are active.
In the results section I take this high correlation into account by estimating alternative model specifications in order to test the robustness of the results.
Table 2 presents the descriptive statistics of this study. Since this study focuses on the determinants of capital structure changes over time, the descriptive statistics are given for each year individually.
Year 98 99 00 01 02 03 04 05 06 07 Net Debt issues
Mean 0.10 0.06 0.04 0.03 -0.07 -0.07 -0.01 0.05 0.05 0.02 Median 0.07 0.07 0.07 0.01 -0.06 -0.06 -0.01 0.07 0.03 0.01 Std. Dev 0.19 0.28 0.76 0.19 0.25 0.19 0.14 0.26 0.18 0.15 Skewness 1.09 -4.97 -7.88 -0.18 0.43 -1.69 -0.68 -5.29 -0.49 -0.27 Kurtosis 6.81 37.17 65.77 5.95 6.62 9.52 5.35 40.34 7.48 5.30 Net equity issues
Mean 0.07 0.04 0.06 0.06 0.01 -0.01 0.03 0.06 0.14 0.11 Median 0.04 0.04 0.04 0.04 0.01 0.01 0.02 0.04 0.04 0.03 Std. Dev 0.24 0.09 0.15 0.23 0.13 0.12 0.12 0.16 0.58 0.63 Skewness 2.19 0.36 2.06 -4.27 -2.25 -2.64 3.60 1.29 7.30 8.03 Kurtosis 9.32 5.32 9.52 33.53 13.74 11.97 21.44 10.80 58.44 67.13 Stock Return Mean -0.01 0.02 -0.05 -0.08 -0.17 0.09 0.08 0.11 0.09 -0.01 Median -0.03 -0.01 0.00 -0.06 -0.11 0.08 0.06 0.10 0.10 0.00 Std. Dev 0.18 0.20 0.19 0.15 0.19 0.14 0.14 0.13 0.12 0.13 Skewness -0.15 0.75 -1.20 -0.64 -0.72 -0.75 0.21 -0.32 -1.68 -1.37 Kurtosis 3.92 4.91 4.59 4.25 3.30 4.35 5.31 4.62 1.10 8.64 Number of Firm-year Obs. 95 95 95 95 95 95 95 95 95 95
The average book leverage is rather stable over time: it moves around 0.62 from 1998 to 2002. From 2003 to 2007 book leverage slightly decreases, moving around an average of 0.56 in this period. Market leverage experiences a peak in the years 2001, 2002, and 2003. This increase in market leverage can be explained by the decrease in stock market prices because of the crash on the stock exchange in 2001 due to the internet bubble. Since market leverage is expressed in market values, it will automatically go up when a firms’ market value decreases.
The above mentioned decrease in stock prices is confirmed by the stock returns of firms in the sample. As reported in table 2, average stock returns indeed sharply decrease in 2001 and 2002 relative to their previous years. The year with the highest average stock return decrease, 2002, is indeed also the year where market leverage is highest. In contrast, the low market leverage of firms from 1998 till 2000 can be explained by the increase of capitalization of firms during these years, where the stock exchange climate was favourable. Market leverage returns to this level again from 2005 till 2007. Average profitability is highest in the years 1998 and 1999, where the average return on assets is 13 % and 11% respectively. The years 2002 and 2003 are the less profitable years in the sample with average return on assets of 3% each year. Profitability increases again from 2004 on, but does not return to the level of 1998 and 1999.
5. Results
This section presents the regression results of this study. In subsection 5.1 I will present the regression results of equation 3), which tests whether market timing provides empirical explanation for capital structure changes of Dutch firms over a short time horizon of 4 years. In subsection 5.2 I will present the regression results for equation 6). This equation tests the ‘persistence’ of the market timing theory, that is, the influence of market timing on capital structure changes is examined over a longer time horizon of 8 years.
5.1 Market timing in the short run
Table 3 presents the regression results of equation 3), which tests whether capital structure changes of Dutch firms can be explained by market timing, while controlling for pecking order financing and capital structure targeting.
As the results indicate I reject hypothesis 1 and 3 that market timing has no significant effect on capital structure changes. The negative significant coefficients for stock return on both market and book leverage indicates that firms use relatively less debt (i.e. more equity) after a stock price
increase. This finding implies that Dutch firms time the equity market and is consistent with empirical evidence for the Netherlands by de Jong and Veld (2001) and de Haan and Hinloopen (2003) and empirical evidence for the US by Kayhan and Titman (2007). The coefficient for the timing measure YT is negative and significant for both market leverage and book leverage, which indicates that firms that raise capital when their market valuation is high, are likely to decrease their debt ratio by issuing relatively more equity. This finding is also consistent with the findings of Kayhan and Titman (2007). Long term timing also has a significant impact on capital structures changes and is negatively related to both market leverage and book leverage. Taking long term timing as a proxy for growth
opportunities, this negative coefficient indicates that firms with higher growth opportunities are more likely to issue equity than debt. This is in line with the prediction that firms with high growth opportunities issue less debt to maintain financial flexibility (Myers, 1977) and consistent with the findings of Kayhan and Titman (2007). Our control variable for capital structure targeting also has a significant effect on capital structure changes. The significant negative coefficient found for target leverage is in line with the trade-off theory, since this implies that firms with higher debt ratios than their optimal target ratio bring back their debt ratio towards their optimal target debt ratio. This finding confirms the findings of Kayhan and Titman (2007) for the US, and also supports the findings of de Jong and Veld (2001) for the Netherlands, who find that Dutch firms are more likely to issue shares when their debt ratio is above their optimal target debt ratio. The coefficient for cumulative
This table reports the regression results of the equation: it it it it it it it it it it it it it it it
it L Stockreturn YT LT DevTar FD dumFD PROF IndustryDummy
L 41 , 4 2 , 43 , 44 45 , 46 , 4 7 , 4 8 4
it
it L
L is the dependent variable and is the change in leverage (both market leverage and book leverage) between year t and year t-4, Stockreturnit,it4is the 4 year cumulative stock log return between year t and year t-4, YTit,it4is the covariance between the market to book ratio and the financial deficit from year t to year t-4, LTit,it4is the product of the average market to book ratio and the average financial deficit (FD) between year t and year t-4, DevTarit4 is the difference between the actual debt ratio and the target debt ratio at year t-4, where the target debt ratio is approximated by the predicted value of the debt ratio (see appendix 4), FDit,it4is the total external financing from year t till year t-4, dumFDit,it4 is a dummy variable that interacts with the financial deficit and takes the value of one when FD is positive and zero when FD is negative, PROFit,it4 is the 4 year cumulative profitability between year t and year t-4. The regression is estimated with generalized least squares with random effects. Variable definitions are given in section 4 and in appendix 1. Standard errors are White standard errors. * denotes significance at the 1% level, ** denotes significance at the 5% level. Industry dummy statistics are suppressed.
Book Leverage Market Leverage Coefficient Std. Error Coefficient Std. Error
firms prefer to finance projects with internally generated funds and will prefer debt financing over equity financing when there is a need for acquiring external capital. Evidence for the pecking order theory in the Netherlands already has been found by Cools (1993) and de Haan and Hinloopen (2003). The coefficients for FD and dumFD are not significant, which means that I find only limited evidence of pecking order behavior in the short run. The positive coefficient for FD in the regressions for both market leverage and book leverage would indicate that firms that face financial deficits are more likely to increase their debt ratio. This is in line with the pecking order theory, as this theory predicts that firms with financial deficits will prefer debt financing over equity financing when there is a need for acquiring external capital. The positive coefficient of dumFD on leverage would imply that financial deficits have a stronger impact on capital structure changes when it is positive, that is, when firms are raising capital instead of retiring (paying out) capital. However, both variables are not significant in either book or market leverage regressions. Because these variables are not significant, I test an alternative model specification where I leave out FD and dumFD. We do this to test the robustness of our regression equation, i.e. whether our results as to the influence of market timing behavior on Dutch capital structures are conclusive. An additional reason for testing these alternative model specifications is the relatively high correlation between FD and LT (see appendix 3). Therefore, I re-estimate the model while leaving out either FD (and dumFD) or LT.
Appendix 5a shows the regression results where I re-estimate the model while leaving out FD and dumFD. The results show a clear picture: re-estimation of the model does not alter the outcomes compared to the full model ones. Thus, it appears that our results about market timing behaviour of Dutch firms in the short run are conclusive. In the light of the high correlation between LT and FD, I re-estimate the model once again leaving out LT and including FD and dumFD, and look whether the results change compared to the full model estimation. The results are clear cut: the market timing coefficients are still significant, and the coefficients for dumFD and FD remain insignificant.
It is worthwile noticing that the R-squared and adjusted R-squared coefficients for the market leverage regression are higher than for the book leverage regression. This implies that the market leverage model fits the data better than the book leverage model, and that the variability of capital structure changes in terms of market values are better explained by our proxy variables than the variability of capital structure changes in terms of book values. This is also found by de bie and Haan (2007), and is in line with the idea that market leverage changes corresponds better with market timing behavior than book leverage changes (Welch, 2004).
examined by de Bie and Haan (2007). We also find significant evidence of capital structure targeting and limited evidence for pecking order behavior in the short run, since only profitability has a significant effect on leverage changes. These results correspond more or less with the findings of Kayhan and Titman (2007), who also find that in the short run capital structure changes are driven by market timing behavior, but also by pecking order financing and capital structure targeting.
5.2 Market timing in the long run
This section examines whether market timing has a persistent effect on capital structure changes. We do this by estimating equation 6), where I examine how the explanatory variables over a 4 year period from t-8 till t-4 influence changes in leverage over an eight year period t-8 till t.
Table 4 present the regression results of equation 6). We observe that stock returns do not longer have a significant effect on book leverage. In fact, the positive coefficient would indicate that Dutch firms are prone to issue debt after stock price increases. On market leverage stock return still has a significant effect, but this can be due to the direct mechanical effect of stock prices on market leverage. As previously discussed, market leverage can fall because of stock price increases even when no equity is issued. The yearly timing measure has lost its significant on both book leverage and market leverage. We must therefore conclude that neither of our market timing measures have a persistent effect on capital structure changes. This evidence is consistent with earlier evidence for the Netherlands from de Bie and Haan (2007) and evidence for the US from Kayhan and Titman (2007), who also find that market timing does not have a persistent effect on capital structure changes.
From table 4 I observe that the deviation from target leverage (DevTar) and cumulative profitability (Prof) do have a persistent effect on capital structure changes. The coefficient of DevTar is more negative in the long run regression estimation (table 4) than in the short run regression (table 3). But because the standard error is higher too, I can not conclude that this variable has a larger negative effect on capital structure changes in the long run than in the short run. The results do neverthless provide significant evidence that, in the long run, companies that are overleveraged bring their leverage back towards their target leverage. Cumulative profitability seems to have a more pronounced relation with capital structure changes in the long run than in the short run: in the long run we observe a significant negative effect of profitability on both book and market leverage, whereas in the short run cumulative profitability only had a significant effect on book leverage. The profitability coefficient is also larger in the long run than in the short run for both book and market leverage, but because the standard error is higher too I can again not conclude that profitability has a larger effect on capital structure changes in the long run than in the short run.
TABLE 4: DOES MARKET TIMING HAS A PERSISTENT EFFECT ON DUTCH CAPITAL STRUCTURES?
This table reports the regression results of the equation:
it it it it it it it it it it it it it it it
it L Stockreturn YT LT DevTar FD dumFD PROF IndustryDummy
L 81 4, 82 4, 83 4, 84 85 4, 86 4, 87 4, 88
8
it
it L
L is the dependent variable and is the change in leverage (both market leverage and book leverage) between year t and year t-8. Stockreturnit it4, 8is the 4 year cumulative stock log return between year t-4 and year t-8. YTit it4, 8 is the covariance between the market to book ratio and the financial deficit from year 4 to year t-8.LTit it4, 8is the product of the average market to book ratio and the average financial deficit (FD) between year t-4 and year t-8. DevTarit8 is the difference between the actual debt ratio and the target debt ratio at year t-8,where the target debt ratio is approximated by the predicted value of the debt ratio (see appendix 4), FDit it4, 8is the total financial deficit between year t-4 and year t-8, dumFDit it4, 8 is a dummy variable that interacts with the financial deficit and takes the value of one when FD is positive and zero when FD is negative, PROFit it4, 8 is the 4 year cumulative profitability between year t-4 and year t-8.
Book Leverage Market Leverage
Coefficient Std. Error Coefficient Std. Error
now also do not have a significant effect on market leverage. This confirms our idea that market timing is not longer present in the long run. The results do not lead to other conclusions compared to the full model estimation: the static trade-off theory seems to explain capital structure changes best in the long run. Alltogether, I cannot reject hypotheses 3 and 4 that market timing has no influence on capital structure changes. Like in section 5.1 we also see that in all models the R-squared and adjusted R-squared coefficients of the market leverage regressions exceed the book leverage regressions: it seems that in the long run our proxy variables explain variablity in market leverage better than variablity in book leverage.
