Testing Theories of Capital Structure: Evidence from Germany, France, and the UK

Testing Theories of Capital Structure: Evidence from

Germany, France, and the UK

Author

Niels Spieker

Student Number: 1389165

University of Groningen

Faculty of Business and Economics

Master's Thesis, MSc Business Administration

Specialization: Finance

Supervisor

Dr. Ing. N. Brunia

ABSTRACT

The majority of the existing empirical literature on capital structure focuses on one theory and on firms from only one country. Conclusions of these studies often contradict each other and therefore results are hard to compare. This study evaluates a combination of the three main theories of capital structure: the static trade-off theory, the pecking order theory, and the market timing theory. I examine whether a combination of these theories can explain the capital structure of firms from the three largest economies of the European Union - Germany, France, and the UK - for the period 1989 till 2008. Using this combination of theories provides the opportunity to test each theory while controlling for the other two. Findings imply that pecking order behavior exists in France and in the UK, but its effect on leverage is not persistent. Evidence for market timing behavior is found only during the years around 2000, and the effect was also temporary. Overall, from these three theories the static trade-off theory explains capital structure the best because the most of the corresponding factors are found to be significant for all (sub)samples and periods.

JEL Classification: C23, G32

Table of contents

1. Introduction ...4  

2. Literature review ...7  

2.1 Static trade-off theory...7  

2.2 Pecking order theory...8  

2.3 Market timing theory ...10  

2.4 Empirical literature ...11   3. Methodology...16   3.1 Model specification ...16   3.2 Hypotheses ...17   3.3 Variables...17   3.3.1 Dependent variables ...18   3.3.2. Independent variables ...18   4. Data...21  

4.1 Company accounts data ...21  

4.2 Macroeconomic variables...24  

5. Results ...26  

6. Conclusion ...31  

References ...34  

Appendix A: Definitions of Firm Characteristic Variables...36  

Appendix B: Equity Risk Premium ...37  

Appendix C: Summary statistics and correlation tables...38  

1. Introduction

Modigliani and Miller (1958) proposed that the capital structure of a firm is irrelevant, assuming perfect and frictionless capital markets. According to their proposition, the choice of financing method does not affect the value of a firm. In reality, however, firms use various combinations of equity and debt to finance their activities. Therefore one of the most interesting challenges of corporate finance is to provide an explanation for the fact that some firms finance their incremental investments with debt, some with equity, and others with a combination of both.

established by cumulative historical financing decisions made based on the relative cost of equity.

Empirical studies published during the last decade have found evidence for all three theories. Especially studies that used data of US firms found results that confirm pecking order and market timing behavior of firms. However, the only theory that constantly proved to be a reliable predictor of capital structure is the static trade-off theory. Although confirming results for the pecking order theory are found in the 1970s and 1980s (Shyam-Sunder and Myers (1999)), the pecking order theory seems to have become less credible as explanation of capital structure. Using more recent data from the 1990s and the beginning of the current century, net equity issues are found to track the external financing needs of firms better than net debt issues (Bessler, Drobetz and Grüninger (2010); Seifert and Gonenc (2010)). These results contradict the predictions of the pecking order theory. Although it has not been around for a long time, the credibility of the market timing theory is also diminishing. Many studies, including Kayhan and Titman (2007) find that it only has a short-term impact on the capital structure, after which firms return to their target debt ratio. An omission of existing empirical literature is that the effect of a different institutional environment, other than that of the US, is usually not considered. Since most research is conducted in the US, still little is known about capital structure in many other areas in the world. Although there are studies that research the static trade-off and pecking order theory in various countries, they usually focus on one country only which makes results between countries hard to compare. Europe provides an excellent sample to compare the determinants of capital structure since the largest economies of the European Union - Germany, France, and the UK - are fairly homogenous in their economic development and all provide samples of firms that are large enough to be able to do a meaningful analysis (Rajan and Zingales (1995); Wanzenried (2006)).

Do the pecking order and market timing theory provide incremental value, relative to the static trade-off theory, in explaining the capital structure of German, French, and UK firms?

If the pecking order and market timing theory do add value in explaining the capital structure, it might be useful to take their predictions into account when making assumptions about the future capital structure of firms. Since the factors that influence pecking order and market timing behavior are hardly predictable, it will be very difficult to find and implement a solution that improves the prediction of capital structures. Therefore I focuses on the question if these theories help explain capital structures and leave the implementation problem to future research.

The static trade-off theory is tested by including firm characteristics like size, type of assets, and profitability as well as factors that are expected to influence the time-varying costs of debt like the interest rate. The pecking order theory is tested by including the external financing needs of the firm, when the pecking order theory indeed affects the capital structure these external financing needs would increase the leverage of the firm, since debt is preferred over equity. To test the market timing theory, this financing deficit is interacted with a direct measure of the relative cost of equity, the market equity risk premium. To check for robustness and persistence, the theories are tested using subsamples from different periods and by lagging variables for different lengths of time. All tests are run separately for all three countries, but are performed in exactly the same way to yield comparable results.

statistics, section 5 will present the results, and section 6 provides the conclusion of this research as well as suggestions for further research.

2. Literature review

In this section I will discuss all three theories and the existing empirical evidence for these theories.

2.1 Static trade-off theory

In their correction on their own capital structure irrelevance proposition (Modigliani and Miller (1958), Modigliani and Miller (1963) discussed one of the most important properties of debt: the tax shield. Although Miller (1977) argues that personal taxes have an effect that cancels out the benefits of corporate taxes, empirical evidence shows that taxes do affect debt policy (Mackie-Mason (1990)) and that the use of debt to shield taxes increases firm value (Graham (2000)). Kraus and Litzenberger (1973) recognized the most important cost of debt: the cost of financial distress. The costs of financial distress refer to the costs of liquidating or restructuring the firm, the direct costs of financial distress, and to indirect costs of financial distress like reputational damage when the firm's creditworthiness is in doubt. Besides these obvious costs and benefits there are other costs and benefits of debt that have to be taken into account. Jensen and Meckling (1976), and Jensen (1986) developed the concept of agency costs and the role debt can play to reduce them. They stress that managers will act in their own interest and therefore will seek higher-than-market salaries or more job security. These personal interests can, for instance, be satisfied by investing the free cash flows in projects that increase the size of the firm or reduce its risk, but possibly decrease the value of the firm. A higher debt level reduces the free cash flow since it increase interest payments, and therefore reduces the agency costs that are associated with the free cash flow. Conflicts of interest between bond holders and shareholders also affect the trade-off between equity and debt in the form of the asset substitution problem, risk-shifting, and the underinvestment problem (Jensen and Meckling (1976); Myers (1977)).

adjustment costs cannot be neglected in the real world, so usually a weak form of the static trade-off theory is used to test the theory. This weak-form static trade-off theory allows firms to temporarily deviate from their optimal capital structure, but posits that within a reasonable time (e.g. two to three years) a firm adjusts its capital structure to its structure. Unfortunately, the optimal capital structure cannot be determined because the factors that are thought to influence it are not always directly observable. That is why firms are assumed to have a target capital structure that equal or close to this optimal capital structure. Again, this target is not directly observable, but there are roughly three approaches to estimate the target leverage. The first approach calculates the average debt ratio of the firm over a longer period, the second approach calculates the average debt ratio across its industry. The first approach assumes that the leverage ratio is mean reverting and that the mean is equal to the target leverage. The second approach assumes that firm characteristics that are thought to influence the trade-off, like firm risk and the type of assets that can serve as collateral, are similar across firms of the same industry. The third approach uses the firm characteristics, and sometimes macroeconomic factors, that are expected to influence the trade-off as factors to determine the target capital structure. Firms characteristics usually involve factors like the type of assets, tangible assets are more suitable as collateral than intangible assets, and firm size, since the probability of default for larger firms is considered lower than that of smaller firms. Macroeconomic factors can include corporate tax rates, a higher tax rate can make debt more attractive but on the other hand less debt is needed to provide the same tax shield, and the default spread, the return investors require to bear the risk of default of the firm. Written as a formula this factor model yields that the target leverage of firm i at time t (Lit,target) is a

function of firm characteristics (xit) and macroeconomic variables (zt):

(1) Lit,target = f ( xit , zt )

2.2 Pecking order theory

place and a valuable real investment opportunity, but with insufficient cash and equivalents available to finance the investment. Because of the information asymmetry, investors cannot determine the exact value of the assets, nor the value of the investment opportunity. If the firm then issues equity, investors can interpret this news as good news and as bad news. Good news means investors interpret the issue as financing for an opportunity with a positive net present value, bad news means that shares are overvalued and the manager tries to transfer wealth from new shareholders to the existing shareholders by issuing overvalued shares. Because rational investors know that managers act in the interest of existing shareholders, the share prices will fall to compensate for possible overvaluation. Since it is not in the interest of existing shareholders to issue undervalued shares, the manager is likely to pass up on the investment opportunity. Since debt is a senior claim on the assets and earnings of the firm relative to equity, which is the residual claim, debt is less exposed to errors in valuation. Therefore, the downward effect on the share price of announcing a debt issue should be smaller than the effect of announcing an equity issue, a prediction that is confirmed by Smith (1986) and Shyam-Sunder (1991)

The above leads to the pecking order theory of capital structure (Myers (1984)):

1. Firms prefer internal financing over external financing (information asymmetry is assumed to be relevant only for external financing).

2. Dividends are sticky (Lintner (1956)). This means that dividend cuts are not used to finance capital expenditures, and that changes in cash requirements are not compensated by short-term changes in dividends pay-outs.

3. If external funding is needed, the safest security available is issued first. This means that debt is issued before equity. If internally generated cash flows exceed the financing needs, the excess cash is used to pay down debt rather than buy back shares.

4. The capital structure of a firm therefore is the result of the cumulative requirement for external financing.

The strong-form pecking order predicts that equity issues are very rare because they only appear if firms are not able to issue debt at reasonable terms anymore, usually because of a large probability of financial distress. Shyam-Sunder and Myers (1999) translate this strong-form pecking order theory in a simple testable model. They define a financing deficit, the part of the investments that cannot be financed with internal funding (DEFit), and state that this

should on average be equal to the net amount of debt issued, since equity issues are very rare and therefore can be neglected. In formula form, their pecking order model is:

in which ΔDit is the net debt issued. The strong-from pecking order predicts that in this model

α is equal to zero and that β is equal to one because all of the deficit is funded by issuing debt or retiring debt if the deficit is negative (hence a financing surplus). Because in reality, equity issues are not very rare, the weak-form of the pecking order allows the beta to be a little less than one an alpha to be non-zero.

This weak-form pecking order theory leaves room for interpretation: at what beta (and what alpha) should the pecking order theory be rejected? Let's define the debt level at which a firm's preference for debt and equity financing is equal as the debt capacity of the firm. Then if the financing deficit is less than or equal to the debt capacity minus the current debt level, the financing deficit should be completely financed with debt and if the financing deficit is larger than this 'excess' debt capacity, it should be (partly) financed with equity. Both the weak and strong form of the pecking order theory predict that this debt capacity is equal to a debt level at which the firm is very close to financial distress. Since financial distress is not in the interest of the shareholders and certainly not in the interest of the manager because the manager loses his job in a case of bankruptcy, it seems unlikely that in reality firms will borrow up to that point. Instead, it can be expected that they will switch to equity earlier, that point is when the debt capacity is reached. Unfortunately, just like that target debt level for the static trade-off, the debt capacity is not directly observable and impossible to determine exactly. However, this debt capacity is likely to be dependent on largely the same factors that are expected to determine the target leverage of the static trade-off theory. Factors that reduce the (in)direct cost of financial distress like tangibility and size can be expected to affect the debt capacity of a firm. The most important factor that is expected to be different for the debt capacity of the pecking order theory compared to the target leverage of the static trade-off theory is profitability. Although the debt capacity is expected to be higher for more profitable firms (enough income to cover large interest payments), the pecking order theory predicts a lower debt level for these firms because they need less external financing than firms that are less profitable.

2.3 Market timing theory

current capital structure is a cumulative outcome of historical attempts to time the market when issuing or repurchasing equity. The market timing theory can be regarded as a special kind of pecking order theory since it also predicts a pecking order of financing methods. The difference between the pecking order theory and the market timing theory is that the pecking order of financing methods within the market timing theory can change due to market conditions, while the pecking order theory predicts that the order is fixed. If stock prices are high, relative to their fundamental value, equity is issued before debt and if stock prices are relatively low, debt is issued before equity.

In the subsection above I discussed the pecking order theory from an information asymmetry perspective. Another approach would be to look at the costs of the financing method. I define the cost of a financing method as the return investors require for the investment in the security related to that financing method. According the capital asset pricing model the expected return of an investment consists of a riskless and a risky part in which the riskless part is equal to the risk-free rate, and the risky part depends on the level of risk and on the compensation investors require to bear that risk. For equity, this compensation is called the market equity risk premium which is equal to the expected return on the market portfolio minus the risk-free rate. The market equity risk premium is implied in the stock prices and can be estimated by calculating the internal rate of return that equates the present value of all future cash flows to common shareholders to the current stock price (Huang and Ritter (2009)). When stock prices are high, relative to their fundamental value, the market equity risk premium is low, and when stock prices are relatively low, the market equity risk premium is high. The market equity risk premium is also called the relative cost of equity, because it measures the compensation shareholders require to bear the market risk of equity. If the market equity risk premium is low, equity is a relatively cheap financing method, and if it is high, equity is relatively expensive as a financing method. The majority of the studies that test market timing theory use the market-to-book ratio of individual firms to measure relative under or overvaluation (Baker and Wurgler (2002); Kayhan and Titman (2007)). The advantages of using the market-to-book ratio compared to the market equity risk premium are that it is easy to measure and that provides the possibility to track mispricing of individual firms. An important disadvantage is that a high market-to-book ratio not necessarily means that a firm is overvalued, it can also be caused by large future growth opportunities.

2.4 Empirical literature

trade-off theory and the pecking order theory were developed earlier, the largest part of the studies focuses on one or both of these theories.

Shyam-Sunder and Myers (1999) use their simple model of the pecking order (equation (2)) and a simple target adjustment model to test both the pecking order theory and the static trade-off theory separately. By using a partial adjustment model, they tested the weak form of the static trade-off theory that allows for positive adjustment costs. Although Shyam-Sunder and Myers (1999) find evidence for both theories, they find that the pecking order theory provides more explanatory power than the static trade-off theory. Shyam-Sunder and Myers' (1999) sample, however, consists of only 157 US firms from 1971-1989 and they required the firms to survive during the whole sample period. Frank and Goyal (2003) used Shyam-Sunder and Myers' simple pecking order model to test the pecking order theory for a broad sample of US firms from 1971-1998. Although their results show that the financing deficit explains a significant part of the net debt issued, they reject both the strong and the weak form of the pecking order hypothesis. Frank and Goyal also used the conventional set of variables as distilled by Rajan and Zingales (Rajan and Zingales (1995` to explain leverage and find signs on the coefficients that are consistent with the pecking order theory. When combining these conventional variables with the financing deficit, they found that the financing deficit adds explanatory value and has a significant positive sign, especially for large firms.

long run they do not find a persistent effect of market timing on capital structures like Baker and Wurgler (2002) did find for US firms.

Table I: Overview of the results of previous empirical studies

The presented signs are based on the author's results from the main results table. A + means that a positive coefficient is found, a - means that a negative coefficient is found and a 0 means that no significant relation is found. Numbers in the parentheses are the relevant coefficients for that variable found by that study.

Independent variables

Authors Key question Data Dependent

variable Tangi-bility Profit-ability Size R&D expenses Financing Deficit Market-to-book/Q Stock returns Market ERP Rajan and Zingales (1995) Evaluating firm characteristics as determinants of leverage.

Panel of firms from G-7 countries from 1987-1991 Book leverage/market leverage + - +* - Shyam-Sunder and Myers (1999)

Testing the pecking order theory against the static trade-off theory.

Panel of 157 surviving US listed firms from 1971-1989

Gross debt issues + (0.75)

Baker and Wurgler (2002)

Testing the pecking order, static trade-off, and market timing theory.

Panel of US IPO firms from 1968-1999

Book leverage and

market leverage + - + - -

Frank and Goyal (2003)

Testing the pecking order theory.

Panel of US listed firms from 1971-1998

Gross debt issues + - + + (0.27) -

Welch (2004) Testing determinants of firm

leverage, especially stock returns.

Panel of US listed firms from 1962-2000

Market leverage 0 0 -

Kayhan and Titman (2007)

Testing market timing theory while controlling for pecking order and static trade-off theory.

Panel of US listed firms from 1960-2003 Book leverage/market leverage - + (0.11/0.20) - - De Bie and De Haan (2007)

Same as Kayhan and Titman (2007).

Panel of Dutch listed firms from 1987-2003

Book

leverage/market leverage

- - +

Huang and Ritter (2009)

Testing market timing theory while controlling for pecking order and static trade-off theory.

Panel of US listed firms from 1964-2001

Book

leverage/market leverage

0 - + - - +

3. Methodology

3.1 Model specification

In theory I expect that observed capital structures are affected by the trade-off between the costs and benefits of debt that determine a target leverage ratio, as well as cash flows, investments, and the market conditions of equity and debt. To test these expectations I follow Huang and Ritter (2009) and regress both book and market leverage on proxy variables for the static trade-off, pecking order, and market timing theories:

(3) Lit = α + β1xit + β2zt + β3PDEFit-k + β4 (ERPt-k-1 PDEFit-k) + εit

In which Lit stands for the leverage ratio of firm i at the end of year t, α is a constant, xit-1 is a

vector of firm characteristics that are expected to influence the trade-off between equity and debt, and zt-1 is a vector of macroeconomic variables that are expected to influence the

trade-off between equity and debt. PDEFit-k is equal to the financing deficit of firm i during year t-k

if it is positive and zero otherwise, and ERPt-k-1 is the implied market equity risk premium at

time t-k-1. Since about two-thirds of the sample firms does not report research and development (R&D) expenses I follow Huang and Ritter (2009) and include a dummy to capture the effect of not reporting R&D expenses. This R&D dummy variable is equal to one if R&D expenses are reported and zero otherwise. Because a dummy variable is used, the use of fixed effects is not possible so equation (4) is estimated using a pooled OLS regression, the results remain robust when estimated without R&D expenses and the dummy. To be able to test if the effects of pecking order behavior and market timing behavior are lasting, I follow Huang and Ritter (2009) and estimate the regression for values of k ranging from 0 to 8. With a larger k, possible effects are expected to reduce, but if the effects are lasting, the coefficients on their proxies should stay significantly positive. The complete definitions of all variables are found later on in this section.

The model used enables us to evaluate the pecking order and market timing theory while controlling for static trade-off variables. Kayhan and Titman (2007) use a two-stage estimation method to estimate a 'leverage deficit' that they include in their regression, instead of including the static trade-off proxies directly. A advantage of that method is that it gives the possibility to estimate the deviation of a target leverage ratio. The advantages of directly including the proxies into the equation are that it gives the opportunity to evaluate the

!

opposite expected sign for the pecking order theory compared to the static trade-off theory. Also, using the target leverage proxies directly in the regression equation potentially reduces the estimation error due to imputed regressors (Hovakimian (2004)).

3.2 Hypotheses

The hypotheses for the variables in the vectors of proxies for target leverage are discussed in the next subsection with their definition. The literature review developes the hypothesis that both the pecking order theory and the market timing theory have an effect on the leverage of firms, at least temporarily. The pecking order theory posits that the positive financing deficit increases leverage because debt is preferred before equity. The market timing theory posits that a low market equity risk premium leads to lower leverage, because equity is then preferred over debt and vice versa. Hence, a positive relation is expected between the market timing proxy and leverage. This leads to the following explicit hypotheses:

1. H0a : β3 > 0

H1a : β3 = 0

2. H0b : β4 > 0

H1b : β4 = 0

If the pecking order theory significantly affects the capital structure of firms, H0a should not

be rejected. If market timing significantly affects the capital structure of firms, H0b should not

be rejected. The vector for firm characteristics, xit in equation (3), also contains variables that

could explain pecking order and/or market timing behavior, namely profitability and Tobin's Q. Therefore, the analysis of the hypotheses above should be looked at in the context of the results found on these firm characteristic variables.

3.3 Variables

3.3.1 Dependent variables

The only dependent variable that is used, is leverage. Leverage is measured in two ways: book leverage and market leverage. Seminal work on capital structure theory is not specific on which measure should be used, market values are the most logical choice from a valuation perspective. Following the majority of the studies on this subject I will estimate and report the results of both measures. I define book leverage as the book value of total debt over the sum of book value of common equity and book value of total debt, and market leverage as book value of total debt over the sum of the market value of common equity and the book value of total debt.

3.3.2. Independent variables

I will define the independent variables in the order they appear in the regression equations. The first regressor of the equation is the vector of firm characteristics which is constituted by tangibility, size, profitability, Tobin's Q, R&D expenses, and capital expenditures. Tangibility is defined as the proportion of net property, plant and equipment of the firm's assets (total assets adjusted for operating liabilities). Tangibility is expected to have a positive effect on leverage for three reasons: first, it is easier for providers of debt to assess the value of tangible assets like land and buildings compared to intangible assets like goodwill. Second, tangible assets are usually easier to sell in a case of liquidation of the firm which leads to a higher expected recovery rate. Third, after an innovation by the competition, intangible assets can decrease in value very quickly while tangible assets are rather stable in their value. The proxy for size is the natural log of net sales or revenues. Larger firms are expected to have a number of properties that reduce that probability of financial distress and therefore it is expected to have a positive impact on leverage. One of these properties is that larger firms are often more diversified, both geographically as industrially. Also, the direct fixed costs of bankruptcy are a smaller portion of the firm value for larger firms (Titman and Wessels (1988)). Profitability is defined as the earnings before interest, taxes, depreciation, and amortization (EBITDA) as proportion of the firm's assets. According to the static trade-off theory, profitability has a positive effect on leverage because more profitable firms have a higher tax liability they can shield using debt, face more agency costs of the free cash flow, and have enough cash flows to cover large interest payments. The pecking order theory, however, posits that more profitable firms need less debt because they can fund their projects internally and therefore it predicts a negative impact of profitability on leverage.

ratio. Since there is no direct measure of future growth opportunities, I follow Huang and Ritter (2009) and use three measures for growth firms: Tobin's Q, R&D expenses, and capital expenditures. Tobin's Q is the ratio of the market value of assets to their replacement value, R&D expenses and capital expenditures are scaled by the firm's assets. The expected sign of the coefficient of Tobin's Q is negative. It has to be noted that Tobin's Q potentially also captures the market timing effect and that it has a mechanical relation with market leverage since the market value of equity is in the numerator of Q and in the denominator of the dependent variable, market leverage. The expected sign of R&D expenses is negative too, since high current R&D expenses are expected to generate future opportunities. It is unclear what should be expected of capital expenditures because it mainly consists of additions to tangible assets like property, plants, machinery, and equipment. On the one hand, these expenditures potentially generate future growth opportunities. On the other hand, investments in tangible assets can increase the tangibility of the firm which is expected to have a positive effect on the debt capacity of the firm. Because it is not clear what part of the capital expenditures consists of tangible assets, I do not predict a specific sign on this coefficient. Following Huang and Ritter (2009) I include the following debt-related macroeconomic variables in the regression: the real interest rate, the default spread, the term spread, the corporate tax rate, and the real growth rate of the gross domestic product (GDP). The real interest rate is defined as the one-year risk-free rate minus the inflation rate. It is expected to have a positive effect on leverage, because a higher interest rate increases the tax shield for every dollar of debt. The default spread (measured as the difference between Moody's Baa- and Aaa-rated corporate bond yields (Huang and Ritter (2009)) is expected to have negative impact on the leverage ratio because it measures the relative cost of debt: if the default spread is high, the borrower pays the lender more for bearing the same risk. However, the tax effect of a higher nominal interest rate caused by a higher default spread might cancel out this effect. Term spread, measured as the difference in yield between the 10 year and one year government bond yield, is expected to have a positive effect on leverage. When the term spread is low, long-term debt is relatively cheap so firms might increase the use of long-term debt. According to the static trade-off theory, the tax shielding property of debt is one of the main reasons to issue debt, so an increase in the corporate tax rate is expected to positively influence leverage.

is raised. The definition of the financial deficit used in this study is equal to the one employed in the studies mentioned above. The financing deficit (DEF) is defined as the sum of investments (capital expenditures, I), dividends (D), and changes in net working capital (ΔWC), net of net cash flow (CF, net income plus depreciation and amortization). This sum is identical to net debt issues (Δd) plus net equity issues (Δe):

(4) DEF = I + D + ΔWC - CF ≡ Δd + Δe.

When this variable is positive, the firm's investments (including dividend distributions) are greater than the internally generated funds. When the financing deficit is negative, it is actually a surplus that can be used for future investments, reduction of debt, share repurchases, or a combination of those. According to the pecking order theory managers are reluctant to issue or repurchase shares, due to information asymmetry. Therefore, the financing deficit should lead to a debt issue if it is positive and a debt reduction if it is negative. It is, however, likely a financial deficit and a financial surplus affect leverage differently (Kayhan and Titman (2007)). For instance, it is possible that issues regarding information asymmetry are different for share issues than for share repurchases and therefore have different effects on pecking order or market timing behavior. Therefore, I follow Huang and Ritter (2009) and use only the positive financing deficit (PDEF). Since the pecking order theory predicts that firms that need external financing (hence, have a positive financing deficit), prefer debt financing over equity financing. Therefore, the positive financing deficit is expected to have a positive effect on leverage.

4. Data

I will describe the data that is used in this study in two subsections: company accounts data and data on macroeconomic variables.

4.1 Company accounts data

One of the biggest problems of performing research on capital structure outside the US is the lack of consistent accounting and market information (Rajan and Zingales (1995)). Therefore, this study is limited to large European countries where there are sufficient firms available to make meaningful comparisons. In particular, this study focuses on financial and non-utility firms from Germany, France, and the UK, the three largest economies of the European Union. Firm-level data is extracted from Thomson Financial's Worldscope database using DataStream for the years 1988 up to and including 2008. Financial institutions are excluded because of their different balance sheet structure, utilities are removed from the sample because a large part of these companies were regulated during the sample period. Observations that have a blank or zero value for total assets are removed as well as observations with a negative value for book equity. Observations with missing values on other variables are also excluded from the final sample, except for research and development expenses. Since more than half of the firms does not report R&D expenses I set the missing values for this variable to zero and rely on a dummy variable to capture the effect of missing values in the analysis (Huang and Ritter (2009)). Because Worldscope occasionally makes recording errors, the top and bottom one percent of certain variables are eliminated, which is in line with comparable studies.

This leaves us with a final sample of 8,813 German observations, 8,401 French observations, and 19,152 UK observations. To be able to compare the results of this study with the results of Huang and Ritter (2009) I will follow their definitions as closely as possible. However, in Germany severe problems arise with accounting definitions for liabilities1, a record Huang and Ritter (2009) use to measure leverage. To overcome these problems, I do not use the record for total liabilities but use total debt instead, according to Rajan and Zingales (1995) this is an appropriate way to measure leverage.

In this study, book leverage is defined as the book value of debt divided by the sum of the book values of equity and debt. Market leverage is defined as the book value of debt divided

1_{"An indication of the importance of these differences is that 29 percent of the liabilities of a German}

by the sum of the book value of debt and the market value of equity. The positive financing deficit (PDEF) is equal to the financing deficit when it is positive and zero otherwise. The positive financing deficit, tangibility (net PP&E), capital expenditures, profitability (EBITDA), and R&D expenses are scaled by the firm's total assets net of operating liabilities. Table II shows the summary statistics for book and market leverage, and figure 1 displays the trends of the median market leverage for all three countries. Book leverage in Germany increases on average during the 1990s to decrease again from around 2000. When we look at the median, the pattern is clearer and shows a dip after 1998. German market leverage shows an even larger dip in 2000, which might be caused by the dot-com bubble and the high stock prices during that period. In France, book leverage starts fairly high at 0.47 in 1988 to decrease and stay steady at around 0.36. Again, when looking at the median, market leverage shows a dip in 2000, albeit relatively small. Average book leverage in the UK does not show a clear trend although it is a little lower, on average, during the second half of the sample period compared to the first half. Like for Germany and France, the median market leverage for the UK shows a dip in 2000, but the largest dip in the UK is in 2006, the year in which the UK housing market bubble reached its peak (Barrell and Davis (2008)). Summary statistics of all company accounts variables are presented in Appendix C.

Table II

Summary statistics firm leverage

Year 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008

Book leverage Germany

Mean 0.286 0.312 0.342 0.336 0.349 0.350 0.311 0.320 0.297 0.288 0.318

Median 0.237 0.285 0.298 0.292 0.318 0.331 0.263 0.287 0.255 0.259 0.307

Std. Dev. 0.219 0.242 0.268 0.269 0.286 0.284 0.281 0.279 0.268 0.256 0.266

Skewness 0.558 0.380 0.358 0.446 0.323 0.323 0.496 0.424 0.531 0.572 0.379

Kurtosis -0.514 -0.897 -1.073 -0.924 -1.151 -1.091 -1.004 -1.066 -0.834 -0.657 -0.972

Market leverage Germany

Mean 0.240 0.182 0.216 0.212 0.246 0.242 0.221 0.257 0.251 0.229 0.230 Median 0.185 0.132 0.146 0.153 0.183 0.194 0.128 0.191 0.183 0.173 0.170 Std. Dev. 0.217 0.176 0.206 0.207 0.233 0.226 0.245 0.252 0.250 0.229 0.223 Skewness 0.810 1.045 0.852 0.932 0.656 0.680 0.934 0.743 0.768 0.889 0.763 Kurtosis -0.414 0.454 -0.224 -0.066 -0.686 -0.561 -0.331 -0.570 -0.500 -0.129 -0.441 N 133 258 301 343 395 426 589 555 483 513 571

Book leverage France

Mean 0.472 0.412 0.427 0.375 0.361 0.361 0.380 0.370 0.351 0.340 0.349

Median 0.505 0.427 0.433 0.361 0.350 0.347 0.382 0.384 0.376 0.327 0.334

Std.Dev. 0.202 0.233 0.252 0.243 0.235 0.229 0.234 0.239 0.242 0.237 0.248

Skewness -0.211 -0.073 0.067 0.212 0.277 0.206 0.086 0.075 0.229 0.375 0.317

Kurtosis -0.660 -0.846 -0.904 -0.865 -0.788 -0.861 -0.946 -0.979 -0.728 -0.628 -0.795

Market leverage France

Table II (continued)

Summary statistics firm leverage

Year 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Book leverage UK Mean 0.256 0.312 0.305 0.269 0.285 0.293 0.286 0.283 0.260 0.242 0.261 Median 0.248 0.299 0.288 0.256 0.282 0.293 0.271 0.258 0.237 0.196 0.222 Std. Dev. 0.180 0.212 0.208 0.192 0.203 0.218 0.231 0.234 0.231 0.235 0.244 Skewness 0.545 0.511 0.541 0.573 0.496 0.499 0.523 0.535 0.687 0.757 0.638 Kurtosis -0.026 -0.155 -0.045 0.042 -0.033 -0.233 -0.439 -0.503 -0.222 -0.293 -0.558 Market leverage UK Mean 0.190 0.235 0.251 0.186 0.201 0.192 0.208 0.219 0.192 0.164 0.194 Median 0.162 0.211 0.201 0.140 0.161 0.155 0.151 0.170 0.145 0.104 0.145 Std. Dev. 0.161 0.183 0.207 0.171 0.182 0.175 0.203 0.207 0.190 0.179 0.201 Skewness 1.104 0.833 0.972 1.218 1.103 1.027 0.918 0.888 0.983 1.213 1.049 Kurtosis 1.184 0.326 0.462 1.217 0.896 0.782 0.074 0.121 0.395 1.187 0.520 N 728 839 752 863 994 1035 951 888 865 1089 1050 4.2 Macroeconomic variables

To estimate the market equity risk premium, the values of five different variables need to be known: share prices, book value of common equity (per share), earnings forecasts, and the risk-free rate. Share prices, book values of common equity, and the number of shares are obtained from DataStream. The sample that is used for the estimation consists of all non-financial and non-utility constituents of the Dow Jones EUROSTOXX 50 Index during the period 1989-2008 for which earnings estimates are available. The estimate used for the market equity risk premium is the non-weighted average of the individual implied equity risk premiums of the sample in every year, the sample varies from 21 firms in 1989 to 30 firms in 2008. The use of only 'well-established', large firms might cause the market equity risk premium estimates to be relatively conservative since large firms are expected to be less riskier (on average) than small and medium-sized enterprises that are included in the regression sample. As a consequence, the results are likely to understate the importance of valuation in the choice between equity and debt because differences in equity risk premium can be more extreme in reality (Huang and Ritter (2009)).

from the online database of the Deutsche Bundesbank that provides EURIBOR rates as well as FIBOR rates which are used before 1999, the first year EURIBOR is reported. Before 1991, the Frankfurt Bank Reported interest rates are used since FIBOR was first reported in 1991. For the UK, LIBOR rates are used for the risk-free interest rate which are obtained from the online statistics database of the Bank of England. Corporate bond yields on Moody's Aaa and Baa ratings to estimate the default spread are retrieved from the US Federal Reserve database using the WRDS. The interest rates used to calculate the term spread are downloaded from the databases of the Deutsche Bundesbank and the Bank of England. I use the difference between the 1-year and 10-year government bond yields to estimate the term spread. The corporate tax rates that are reported in KPMG's "Corporate and Indirect Tax Rate Survey 2009" are used as corporate tax rates and for the years that are not covered by KPMG's report, the data from Djurovic-Todorovic (2002) is used. Growth percentages of the consumer price indexes (inflation rates) are drawn from the Financial Statistics database of the International Monetary Fund (IMF).

Figure 2 is a graph of the market equity risk premium (ERP) for Germany and France, and for the UK. This graph shows two very pronounced 'dips' in the Germany and France ERP that indicate a temporary overvaluation of equity by the market. Especially the dip around 1999-2001 is notable since it coincides with the dips observed in market leverage for all three countries during that period (see Table I). The differences between the UK, and Germany and France are explained by the difference in risk-free rate. The EURIBOR (FIBOR) seems to follow the LIBOR interest rates with a small lag which causes the ERP of Germany and France to follow the ERP of the UK with a small lag.

5. Results

This section presents the results of this study. The results of regression equation (3) for both book and market leverage will be discussed in the light of all three theories for all three countries. After discussing the results of the complete sample of every country I discuss the results of subsamples from different periods.

The regression is estimated with and without the macroeconomic variable (zt in equation (3)).

Since the macroeconomic variables were often found to be insignificant and did not alter my conclusions, the results presented in Table III are from the estimation without the macroeconomic variables. This is in line with the common econometric practice that it is better to leave irrelevant variables out of the equation (Brooks (2008)). The estimation results including the macroeconomic variables are presented in Appendix D .

The pecking order theory, however, does predict that more profitable firms have lower debt ratios, because they have more internal funding available. The pecking order theory also predicts that firms that need external financing prefer to finance their deficit with debt instead of equity, which would increase their leverage. Therefore the pecking order theory predicts a positive sign for the coefficient of the positive financing deficit. A significant positive sign on the financing deficit (row (i)) is indeed found for book leverage in France and for both book and market leverage in the UK. Combined with the negative sign on profitability (row (d)), these results provide evidence for the pecking order theory in the UK and to some extent in France. No evidence for the pecking order theory is found in Germany. When comparing this result to the results of Bessler et al. (2010) and Jalal (2007), these results are not consistent since Bessler et al. (2010) and Jalal (2007) find less evidence for the pecking order theory in common law countries, like the UK, compared to civil law countries like Germany and France while my results indicate the opposite. Jalal (2007) explains differences in pecking order behavior by differences in shareholder protection by law which they argue is better in common law countries which reduces information asymmetry costs and therefore causes less pecking order behavior. Bessler et al. (2010) also stresses that the pecking order theory performs better with large firms, but both the average and the median size of UK firms are smaller that those of German and French firms. Therefore there is no obvious explanation for the differences I find between the countries in this study. The positive coefficient (2.219) on the market timing proxy for market leverage (row (j), column (2)) in Germany indicates market timing behavior. However, the coefficient is only marginally significant. For France, the market timing proxy yielded no significant results, while the coefficient for market timing in the UK is even negative for both book and market leverage.

Following Huang and Ritter (2009) I also estimated equation (3) with lagged pecking order and market timing variables for different periods (2, 3, and 4 years). These lagged regressions did not show consistent significant effects of market timing, the results of the 3-year lagged regression is presented in Appendix D. Jungqvist and Wilhelm (2003) provide another explanation for this decrease in leverage during the dot-com bubble. They find that IPO underpricing was very high during the dot-com bubble and that this might have provided incentives for managers to issue a lot of equity during that period.

Table III: Effects of firm characteristics, financing activities, and market conditions on leverage The following equation is estimated using German, French, and UK data from 1990-2008

Lit = α + β1xit + β2PDEFit-k + β3(ERPt-k-1

€

× PDEFit-k)

In which Lit, the dependent variable, is the book or market leverage of firm i at the end of year t. The vector xit represents firm characteristics of firm i at the end of year t. PDEFit-k is the

positive financing deficit of firm i at the end of year (t-k) and ERPt-k-1 is the market equity risk premium at the end of year t-k-1. In this table, k = 0. A dummy is included for firms that do not

Table IV: Effects of firm characteristics, financing activities, and market conditions on leverage (1995-2004) The following equation is estimated using German, French, and UK data from 1995-2004

Lit = α + β1xit + β2PDEFit-k + β3(ERPt-k-1

€

× PDEFit-k)

In which Lit, the dependent variable, is the book or market leverage of firm i at the end of year t. The vector xit represents firm characteristics of firm i at the end of year t. PDEFit-k is the

positive financing deficit of firm i at the end of year (t-k) and ERPt-k-1 is the market equity risk premium at the end of year t-k-1. In this table, k = 0. A dummy is included for firms that do not

6. Conclusion

This study evaluates the three most studied theories of capital structure: the static trade-off, the pecking order, and the market timing theory. Since existing literature finds that none of these theories completely explains the capital structure of firms, it would be interesting to see if a combination of these theories would lead to better results than one of the theories on their own. The model used in this thesis consists of book leverage and market leverage as dependent variables, and proxies for the three theories of capital structure as independent variables. I use proxies for the type of assets, size, profitability, and growth opportunities to account for the target leverage that is predicted by the static trade-off theory. The pecking order predicts that almost all external financing of a firm consists of debt issues, so the positive financing deficit is included to test that prediction. Where most other studies use a market-to-book ratio as a proxy for market timing, I follow Huang and Ritter (2009) in using the implied market equity risk premium as a direct measure of the relative cost of equity as financing method. The sample used consists of listed firms from the three largest economies of the European Union: Germany, France, and the United Kingdom (UK). Where most studies do not include data after 2001, the dataset used in this study covers data from 1989 - 2008.

timing proxy and firm leverage. For the period 1995-2004, however, I have found this positive relation which indicates that during that period firms exhibited market timing behavior. The fact that for the complete period this relation is not found, and that lagging the market timing variable did not yield positive results, leads to the conclusion that this market timing effect is only temporary. Kayhan and Titman (2007) and Huang and Ritter (2009) cover a longer period from the 1960s until 2001 and do find market timing behavior for the complete period. Kayhan and Titman (2007) do find market timing behavior for their complete sample, but also find that this effect is only temporary. Huang and Ritter (2009), who also use the market equity risk premium as proxy for market timing, do find a persisting market timing effect on leverage. Based on these results I conclude that the pecking order theory and market timing theory add relatively little information to the static trade-off theory in explaining capital structures of firms. Therefore the current practices for predicting future capital structures that are based on the static trade-off theory do not need to change.

A returning problem in research on capital structure is finding a good proxy. Because the predicted behavior of theories is often not directly measurable, one or more proxies need to be used to test the prediction of a theory. Take, for instance, the financing deficit that is used as a proxy for the pecking order in this study. This proxy is developed by Shyam-Sunder and Myers (1999) and it is used in a lot of studies, but it is also criticized by Chirinko and Singha (2000). The model used in this study does not require that equity issues are very rare, which is the case for Shyam-Sunder and Myers' (1999) model the way they use it in their study. However, the fact that it is still unclear what the expected effect should be of a negative financing deficit (hence a surplus) limits this study to only examining the positive financing deficit. A better understanding of the behavior of firms that have a surplus might lead to additional insights into for instance share repurchase programs.

Most studies on market timing use the market-to-book ratio as a proxy for the relative cost of equity. This study uses the market equity risk premium (ERP) for that matter, which is a more direct measure of the relative cost of equity. Using the ERP instead of the market-to-book ratio overcomes the problem that a high market-to-book ratio not necessarily means that a firm is overvalued. A high market-to-book ratio, after all, could also indicate expected future growth opportunities. On the other side, the market ERP is an aggregate measure and is therefore equal for every firm, while it is likely that there are differences between firms and industries in overvaluation. During the dot-com bubble of the late 1990s for instance, the real ERP for technological ('dot-com') firms might have been much lower than the average of the EUROSTOXX 50 firms. The real ERP for more 'old-fashioned' industries might have been much higher. Estimating the ERP for every firm separately might not be feasible for several reasons. First, for smaller firms earnings forecasts are generally not reported in the I/B/E/S database. Second, there is a large potential measurement error since the exact date of the forecast is unknown which makes it difficult to estimate the correct share price to use in the estimation. By taking an average of a sample of firms, these measurement errors are reduced and firm risk (beta) does not have to be estimated. Estimating the ERP for various industries separately, might prove useful. During the dot-com bubble, for instance, it was a specific industry that was particularly overvalued.

References

Baker, Malcolm, and Jeffrey Wurgler, 2002, Market timing and capital structure, Journal of

Finance 57, 1-32.

Barrell, Ray, and E. Philip Davis, 2008, The evolution of the financial crisis of 2007-8,

National Institute Economic Review 5-14.

Bessler, Wolfgang, Wolfgang Drobetz, and Matthias C. Grüninger, 2010, International tests of the pecking order theory, Working Paper (University of Hamburg).

Brooks, Chris, 2008. Introductory econometrics for finance (Cambridge University Press, Cambridge).

Brounen, Dirk, Abe de Jong, and Kees Koedijk, 2006, Capital structure policies in europe: Survey evidence, Journal of Banking & Finance 30, 1409-1442.

Chirinko, Robert S., and Anuja R. Singha, 2000, Testing static tradeoff against pecking order models of capital structure: A critical comment, Journal of Financial Economics 58, 417-425.

De Bie, Tijs, and Leo De Haan, 2007, Market timing and capital structure: Evidence for Dutch firms, De Economist 155, 183-206.

Djurović-Todorović, Jadranka, 2002, Corporate income tax in EU countries comparative analysis, Economics and Organization (Faculty of Economics Niš, University of Niš, Niš).

Donaldson, Gordon, 1961, Corporate debt capacity: A study of corporate debt policy and the determination of corporate debt capacity, (Division of Research, Harvard Graduate School of Business Administration, Boston).

Drobetz, Wolfgang, and Pascal Pensa, 2007, Capital structure and stock returns: The european evidence, SSRN eLibrary.

Feltham, Gerald A., and James A. Ohlson, 1995, Valuation and clean surplus accounting for operating and financial activities, Contemporary Accounting Research 11, 689-731. Frank, Murray Z., and Vidhan K. Goyal, 2003, Testing the pecking order theory of capital

structure, Journal of Financial Economics 67, 217.

Graham, John R., 2000, How big are the tax benefits of debt?, Journal of Finance 55, 1901-1941.

Graham, John R., and Campbell R. Harvey, 2001, The theory and practice of corporate finance: Evidence from the field, Journal of Financial Economics 60, 187-243. Hovakimian, Armen, 2004, The role of target leverage in security issues and repurchases,

Journal of Business 77, 1041-1071.

Huang, Rongbing, and Jay R. Ritter, 2009, Testing theories of capital structure and estimating the speed of adjustment, Journal of Financial & Quantitative Analysis 44, 237-271. Jalal, Abu M., 2007, The pecking order, information asymmetry, and financial market

efficiency, SSRN eLibrary.

Jensen, Michael C., 1986, Agency costs of free cash flow, corporate finance, and takeovers,

The American Economic Review 76, 323-329.

Jensen, Michael C., and William H. Meckling, 1976, Theory of the firm: Managerial behavior, agency costs and ownership structure, Journal of Financial Economics 3, 305-360.

Jungqvist, Alexander L., and Jr Williamj Wilhelm, 2003, Ipo pricing in the dot-com bubble,

Journal of Finance 58, 723-752.

Kayhan, Ayla, and Sheridan Titman, 2007, Firms' histories and their capital structures,

Journal of Financial Economics 83, 1-32.

Kraus, Alan, and Robert H. Litzenberger, 1973, A state-preferene model of optimal financial leverage, Journal of Finance 28, 911-922.

Lee, Charles M. C., James Myers, and Bhaskaran Swaminathan, 1999, What is the intrinsic value of the dow?, Journal of Finance 54, 1693-1741.

Mackie-Mason, Jeffrey K., 1990, Do taxes affect corporate financing decisions?, Journal of

Finance 45, 1471-1493.

Miller, Merton H., 1977, Debt and taxes, Journal of Finance 32, 261-276.

Modigliani, Franco, and Merton H. Miller, 1958, The cost of capital, corporation finance and the theory of investment, The American Economic Review 48, 261.

Modigliani, Franco, and Merton H. Miller, 1963, Corporate income taxes and the cost of capital: A correction, The American Economic Review 53, 433-443.

Myers, Stewart C., 1977, Determinants of corporate borrowing, Journal of Financial

Economics 5, 147-175.

Myers, Stewart C., 1984, The capital structure puzzle, Journal of Finance 39, 575-592. Myers, Stewart C., and Nicholas S. Majluf, 1984, Corporate financing and investment

decisions when firms have information that investors do not have, Journal of

Financial Economics 13, 187-221.

Rajan, Raghuram G., and Luigi Zingales, 1995, What do we know about capital structure? Some evidence from international data, Journal of Finance 50, 1421-1460.

Ritter, Jay R., and Richard S. Warr, 2002, The decline of inflation and the bull market of 1982-1999, Journal of Financial & Quantitative Analysis 37, 29-61.

Seifert, Bruce, and Halit Gonenc, 2010, Pecking order behavior in emerging markets, Journal

of International Financial Management and Accounting 21, 1-30.

Shyam-Sunder, Lakshmi, 1991, The stock price effect of risky versus safe debt, Journal of

Financial & Quantitative Analysis 26, 549-558.

Shyam-Sunder, Lakshmi, and Stewart C. Myers, 1999, Testing static tradeoff against pecking order models of capital structure, Journal of Financial Economics 51, 219-244. Smith, Clifford W. Jr, 1986, Investment banking and the capital acquisition process, Journal

of Financial Economics 15, 3-29.

Titman, Sheridan, and Roberto Wessels, 1988, The determinants of capital structure choice,

Journal of Finance 43, 1-19.

Wanzenried, Gabrielle, 2006, Capital structure dynamics in the UK and continental europe,

The European Journal of Finance 12, 693-716.

Appendix A: Definitions of Firm Characteristic Variables

Table A1: Definitions of Firm Characteristic Variables

Variable Symbol Definition DataStream/Worldscope code(s)

1 Book leverage Lbook D/(D+E) See items 3 and 4

2 Market leverage Lmarket D/(D+MV) See items 3, 4, and 5

3 Book value of debt D Total debt + preferred stock WC03255 + WC03451

4 Book value of equity E Book value of common equity WC03501

5 Market value of equity MV Market value of equity MV

6 Financing deficit* DEF (PDEF) ΔWC + CAPEX + DIV - CF See items: 7, 8, 9, and 10 7 Working capital* WC Current assets - current liabilities WC02201 - WC03101

8 Dividends paid* DIV Total cash dividends paid WC04551

9 Capital expenditures* CAPEX Capital Expenditures WC04601

10 Cash flows from operations* CF Cash flows from operations: net income + depreciation and amortization

WC01551 + WC01151 11 Tangibility* TANG Net property, plant, and equipment WC02501

12 Size SALES Natural log of net sales or revenues WC01001

13 Profitability* OIBD Operating Income Before Depreciation WC018198

14 Tobin's Q Q (MV+D)/TA See items 3, 5, and 16

15 R&D expenses* R&D Research and development expenses WC01201 16 Book value of assets TA Book value of total assets adjusted for operating liabilities D + E

Appendix B: Equity Risk Premium

The implied equity risk premium (ERP) is estimated using forecasted earnings from the International Brokers' Estimates System (IBES). Following Lee, Myers and Swaminathan (1999) I use the residual income valuation model developed by Feltham and Ohlson (1995). The model consists of the book value of a firm (Bt) at time t plus a finite sum of the present

values of the economic values added (EVA) and a continuing value term. By setting the cost of equity at a value that makes that the value of the stock calculated by the model equals the market value of the stock the implied cost of equity is determined. The equity risk premium is calculated as the difference between this cost of equity and the risk-free rate (risk-free rate in this case is, following Huang and Ritter (2009) the annualized one month EURIBOR, before 1999 it is the one month Frankfurt Interbank Offered Rate (FIBOR) and before 1991 it is the one month rate reported by the Frankfurt Banks, all retrieved from the BKK Statistics database of the Deutsche Bundesbank.. For the UK, LIBOR rates are used. For estimating the aggregate equity risk premium we take the unweighted average of the (financial and non-utility, ICB Industry codes 8000 and 7000) constituents of the Dow Jones EURO STOXX 50 Index risk premiums estimated as described above.

(B1)

€

V

= B

+

FEPS

− R × B

(1+ R)

+

FEPS

− R × B

(1+ R)

+

FEPS

− R × B

(1+ R)

_R

Stock prices, book value of equity and number of shares are retrieved from Datastream and the Worldscope database (codes: P, NOSH, and WC03501).

Where:

€

V

€

FEPS

€

B

€

Appendix C: Summary statistics and correlation tables

Summary statistics company accounts data Germany

Year 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Book leverage Mean 0.286 0.312 0.342 0.336 0.349 0.350 0.311 0.320 0.297 0.288 0.318 Median 0.237 0.285 0.298 0.292 0.318 0.331 0.263 0.287 0.255 0.259 0.307 Std. Dev. 0.219 0.242 0.268 0.269 0.286 0.284 0.281 0.279 0.268 0.256 0.266 Skewness 0.558 0.380 0.358 0.446 0.323 0.323 0.496 0.424 0.531 0.572 0.379 Kurtosis -0.514 -0.897 -1.073 -0.924 -1.151 -1.091 -1.004 -1.066 -0.834 -0.657 -0.972 Market leverage Mean 0.240 0.182 0.216 0.212 0.246 0.242 0.221 0.257 0.251 0.229 0.230 Median 0.185 0.132 0.146 0.153 0.183 0.194 0.128 0.191 0.183 0.173 0.170 Std. Dev. 0.217 0.176 0.206 0.207 0.233 0.226 0.245 0.252 0.250 0.229 0.223 Skewness 0.810 1.045 0.852 0.932 0.656 0.680 0.934 0.743 0.768 0.889 0.763 Kurtosis -0.414 0.454 -0.224 -0.066 -0.686 -0.561 -0.331 -0.570 -0.500 -0.129 -0.441

Positive financing deficit (PDEF)

Table C1 (continued)

Summary statistics company accounts data Germany

Year 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Tangibility Mean 0.716 0.771 0.838 0.804 0.728 0.660 0.518 0.499 0.461 0.397 0.397 Median 0.638 0.626 0.654 0.631 0.579 0.544 0.368 0.363 0.361 0.282 0.275 Std. Dev. 0.473 0.600 0.725 0.687 0.654 0.595 0.637 0.920 0.450 0.428 0.563 Skewness 1.817 2.097 2.738 2.783 3.209 2.988 5.661 15.394 2.893 2.816 7.589 Kurtosis 5.661 6.341 9.301 11.130 18.628 14.174 57.743 304.222 17.782 14.292 92.202 Size Mean 13.066 12.296 12.428 12.310 11.940 12.007 11.862 11.753 11.871 11.804 11.822 Median 13.121 12.352 12.320 12.261 12.049 12.135 11.812 11.682 11.727 11.661 11.609 Std. Dev. 1.886 2.059 1.960 2.051 2.378 2.487 2.200 2.252 2.229 2.232 2.223 Skewness -0.513 -0.240 -0.070 -0.377 -0.513 -0.613 0.089 0.025 0.167 0.258 0.213 Kurtosis 1.723 0.485 0.294 1.277 0.623 1.119 0.464 1.382 0.816 1.050 0.759 Q Mean 2.120 3.278 3.096 4.426 4.278 9.814 3.477 2.156 1.998 2.090 2.381 Median 1.275 2.079 1.828 1.876 1.618 1.601 1.767 1.374 1.289 1.422 1.549 Std.Dev. 3.344 4.662 8.941 19.888 18.627 137.032 4.446 3.195 2.671 2.437 3.050 Skewness 6.683 6.032 14.986 12.998 11.070 20.276 3.890 7.418 7.062 4.632 6.872 Kurtosis 51.150 44.590 243.747 180.343 127.768 415.042 21.394 73.766 72.588 26.166 72.863

Research and development expenses

Table C2

Summary statistics company accounts data France

Year 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Book leverage Mean 0.472 0.412 0.427 0.375 0.361 0.361 0.380 0.370 0.351 0.340 0.349 Median 0.505 0.427 0.433 0.361 0.350 0.347 0.382 0.384 0.376 0.327 0.334 Std. Dev. 0.202 0.233 0.252 0.243 0.235 0.229 0.234 0.239 0.242 0.237 0.248 Skewness -0.211 -0.073 0.067 0.212 0.277 0.206 0.086 0.075 0.229 0.375 0.317 Kurtosis -0.660 -0.846 -0.904 -0.865 -0.788 -0.861 -0.946 -0.979 -0.728 -0.628 -0.795 Market leverage Mean 0.355 0.248 0.335 0.288 0.316 0.273 0.267 0.277 0.282 0.256 0.248 Median 0.332 0.207 0.316 0.248 0.286 0.224 0.209 0.250 0.251 0.210 0.201 Std. Dev. 0.219 0.200 0.250 0.237 0.244 0.222 0.223 0.225 0.227 0.222 0.219 Skewness 0.291 0.602 0.324 0.568 0.475 0.857 0.720 0.577 0.530 0.905 0.878 Kurtosis -0.759 -0.658 -1.047 -0.722 -0.808 0.116 -0.383 -0.628 -0.626 0.235 0.060

Positive financing deficit (PDEF)