Loading...

Target leverage and capital structure adjustment speed across German industries

Seminar Paper 2010 37 Pages

Economics - Finance

Excerpt

Table of Contents:

List of Abbreviations

List of Symbols

List of Appendices

1. Introduction

2. Theories on capital structure
2.1. Trade-off theory
2.2. Other theories on capital structure

3. The dynamic framework
3.1. The two-stage dynamic partial-adjustment capital structure model
3.2. Definitions of leverage
3.3. Determinants of the target leverage
3.4. Determinants of the speed of adjustment

4. Empirical analysis
4.1. Data description
4.2. Descriptive statistics of observed leverage ratios
4.3. Determinants of the target leverage
4.4. Determinants of the speed of adjustment

5. Conclusion

Appendix

List of Cited Literature

List of Abbreviations

illustration not visible in this excerpt

List of Symbols

illustration not visible in this excerpt

List of Appendices

A.1. Calculations for market leverage as dependent variable

A.2. Calculations for book leverage as dependent variable

A.3. Data and variable definitions

List of Tables

Table 1: Case 1 - Results of the first-stage regression

Table 2: Case 2 - Results of the first-stage regression

Table 3: Case 1 and 2 - Speeds of adjustment across industries for the two-stage partial-adjustment model

Table 4: Case 1 and 2 - Correlation coefficients for the two-stage partial-adjustment model

Table 5: Descriptive statistics of market leverage from 1991-

Table 6: Correlation matrix of independent variables

Table 7: Major industry groups

Table 8: Three main factors affecting the speed of adjustment

Table 9: Descriptive statistics of market leverage by industry

Table 10: Descriptive statistics of book leverage from 1991-

Table 11: Correlation matrix of independent variables

Table 12: Major industry groups

Table 13: Three main factors affecting the speed of adjustment

Table 14: Descriptive statistics of book leverage by industry

1. Introduction

Since Modigliani/Miller ’ s famous theorem (1958) that capital structure is irrelevant for firm valuation,1 firms’ capital structure choice has been one of the most signifi- cant subjects in the modern finance theory. The subsequent theoretical literature has found evidence to negate the irrelevance theorem.2 Most empirical studies applied a static framework and are capable to explain differences in the optimal leverage ratios across firms, using observed leverage ratios as proxies for the optimal target leve- rage, but do not explain observed differences in firms’ leverage ratios itself.3 One broadly accepted reason for a firm’s deviation from their target leverage ratio is the existence of adjustment costs. In the presence of adjustment costs, firms may deviate from their target leverage and find it not cost effective to adjust their leverage ratio frequently or fully within one period, even if they recognize that their existing capital structure is not optimal. This shows the need for developing and using a dynamic approach in order to examine firms’ capital structure.4 Besides examining the deter- minants of firms’ target leverage, the recent literature focuses on firms’ speed of ad- justment towards their target ratios.5 It is interesting to investigate how fast firms adjust towards their optimal leverage and especially which factors determine the speed of adjustment. This paper investigates the dynamics of capital structure of German industries for the period from 1991 to 2009 based on the two-stage dynamic partial-adjustment model that is commonly used in literature.6

The paper is organized as follows. Section 2 provides a brief overview of the three main theories of capital structure. Section 3 specifies the dynamic partial-adjustment model and describes the variables that may affect the target capital structure as well as the adjustment speed. Section 4 reports the empirical results and Section 5 con- cludes the paper.

2. Theories on capital structure

The financial literature discusses three main theories that try to explain firms’ capital structure choice. The most influential and classical theories of capital structure choice are the trade-off and the pecking order theory. Welch (2006) points out that these two theories are not mutually exclusive.7 A more recent approach is the market timing hypothesis. Several empirical studies show that none of these theories are capable to explain observed capital structures completely.8 About 50 % of the ob- served changes in capital structure can be explained by the trade-off theory, whereas only 10 % can be attributed to each market timing and pecking order strategy.9 For this reason, the next paragraphs provide a short outline of all three theories.

2.1. Trade-off theory

One popular model is the static trade-off theory suggested by Myers (1984).10 This theory focuses on only one time period and assumes the existence of a constant op- timal leverage that maximizes a firm’s value. Optimal leverage is determined by ba- lancing the benefits against the costs of debt financing within one period, holding the firm’s assets and investment plans constant. Benefits of debt include the tax deducti- bility of corporate interest payments. By substituting equity for debt, firms can bene- fit from higher debt tax shields and, consequently, these lead to an increasing firm value.11 However, the magnitude of this effect depends on the extent of the higher personal tax rate on debt, relative to equity, that may offset or even outweigh the benefits from tax shields.12 By contrast, higher leverage also causes higher direct and indirect costs of financial distress. Direct costs of financial distress refer to legal and administrative costs of liquidation or reorganization, whereas indirect costs are asso- ciated with agency problems such as moral hazard or adverse selection.13 The reason why firms do not adjust back towards their targets instantaneously is explained by adjustments costs and economic changes as well as firm-specific characteristics, such as stock price realizations.14 If there were no or only small adjustments costs, a firm would always be at its target due to instant adjustments. In practice a wide variation in actual leverage ratios can be observed. This could be an indicator for the existence of significant adjustment costs and an explanation for different capital structures among otherwise similar firms.15 Thus, the type and magnitude of adjustment costs determine the frequency and the optimal degree of adjustment.16 One can see that the static trade off theory only works out to some extent. For this reason, it has been ex- tended to the dynamic version.

The dynamic trade-off theory is based on the assumptions of the static trade-off theory, but instead of setting a constant optimal leverage ratio an optimal range of leverage is determined. This optimal range varies over time with firm-specific cha- racteristics such as size or profitability.17 If a firm exceeds the upper or lower bound, it will undertake capital structure adjustments and drift back to the opposed boundary step by step.18

2.2. Other theories on capital structure

The other main theory is known as pecking order theory, previously suggested by Donaldson (1961) and further developed by Myers and Majluf (1984). Instead of having leverage targets, it is assumed that the management has superior information about the firm’s value than potential investors.19 This results in the existence of asymmetric information, such as adverse selection, and leads to a hierarchy of prefe- rences in terms of funding sources. In general, firms tend to rely on internal sources of funds, and prefer debt to equity if external financing is required.20

However, both the trade-off and the pecking order theory do not have the capability to explain the statistical and economical impact of historical market-to-book ratios on capital structures. Baker and Wurgler (2002) have found a way to explain this effect, called market timing hypothesis. They argue that firms time their equity issues to the equity market conditions. Consequently, this leads to changes in capital structure. Due to the persistence of these changes, the authors conclude that capital structure is the result of the cumulative outcome of past attempts to time the equity market. Generally, companies issue new equity when shares are considered to be overvalued and repurchase when they are perceived to be undervalued.21

3. The dynamic framework

In this section I introduce the framework for measuring the speed of adjustment to- wards the desired target leverage. First, the model and the estimation methodology are presented. Second, the determinants of the target capital structure are described and finally the variables that influence the speed of adjustment to the target leverage are discussed.

3.1. The two-stage dynamic partial-adjustment capital structure model

An appropriate and commonly used model in the finance literature to measure the speed of adjustment is the two-stage dynamic partial-adjustment model.22 The firststage can be written as follows:

Stage 1 D

illustration not visible in this excerpt

(1)

where[Abbildung in dieser Leseprobe nicht enthalten]denotes a firm’s i desired target leverage ratio,[Abbildung in dieser Leseprobe nicht enthalten]is a vector of firm characteristics related to costs and benefits that determine a firm’s desired target le- verage, and represents a coefficient vector. The data of all firms are pooled and the parameters of equation (1) are estimated by applying simple ordinary least squares (OLS). Using simple OLS with panel data is called a pooled OLS regression. It is subject to a large number of error types23, but it is a simple and quick benchmark to which more sophisticated regressions can be compared. Practically, I put all data together and do not make any distinction between cross-section and time series. The firms’ historical leverage ratios are employed as dependent variable. In a further step, I use the estimated equation to determine the target leverage ratios[Abbildung in dieser Leseprobe nicht enthalten]for each firm.

Before running the second-stage regression, I classify each firm of my sample into its major industry group.24

The second stage uses the estimated target leverage ratios from the first-stage regression in order to measure the speed of adjustment towards the target leverage for each major industry group. This regression can be written as follows:

Stage 2 D

illustration not visible in this excerpt

(2)

where is the adjustment parameter representing the speed of adjustment to the desired target leverage for a major industry group, starting from the previous year’s leverage ratio. [Abbildung in dieser Leseprobe nicht enthalten]and[Abbildung in dieser Leseprobe nicht enthalten] represent firm’s i actual leverage in period[Abbildung in dieser Leseprobe nicht enthalten]and[Abbildung in dieser Leseprobe nicht enthalten], respectively, and [Abbildung in dieser Leseprobe nicht enthalten] is a statistical error term assumed to have mean 0 and constant variance. This equation is estimated by applying pooled OLS again. I interpret as the average speed for a “typical” firm within its major industry group. If 1, a firm respectively a major industry group fully adjust to its target instantaneously and the actual leverage ratio never deviates from the target leverage. This will be the case in a frictionless world. However, in a real world with adjustment costs firms may not find it optimal to adjust fully from period t 1 to t, but partially. Thus, is ex- pected to be less than one.25 Finally, if 1, a firm adjusts more than necessary and is not at its target level again.26

3.2. Definitions of leverage

Before estimating equations (1) and (2), the leverage ratio needs to be defined. Sev- eral definitions of leverage have been used in the literature. In addition, it is not clear a priori whether leverage based on book or market values and total debt or only long- term debt should be preferred in capital structure studies.27 Earlier studies pay more attention to book-valued leverage ratios, whereas previous research focuses on ana- lyzing market-valued leverage ratios. When authors have analyzed market and book leverage, the results have been very similar.28 Nevertheless, I use both market (Case

1) and book leverage ratios (Case 2) as dependent variables in my analysis. I define

market leverage ([Abbildung in dieser Leseprobe nicht enthalten]) as follows:29

illustration not visible in this excerpt

(3)

where[Abbildung in dieser Leseprobe nicht enthalten]and[Abbildung in dieser Leseprobe nicht enthalten] denote firm’s i interest-bearing short-term and long-term debt in period t, respectively, and[Abbildung in dieser Leseprobe nicht enthalten] represents firm’s i market capitalization in period t. The market capitalization is calculated by outstanding shares times market price in the particular period. Book leverage ([Abbildung in dieser Leseprobe nicht enthalten]) can be written as follows:30

illustration not visible in this excerpt

(4)

where[Abbildung in dieser Leseprobe nicht enthalten] i, t denotes firm’s i total assets in period t.

3.3. Determinants of the target leverage

Researchers have found that target leverage depends on firm-characteristic determi- nants.31 Due to the fact that many of them are unobservable and difficult to quantify, e.g. bankruptcy costs, proxies for the benefits and costs of leverage are used.32 I can base the selection for my sample only to a limited extent on previous studies, because the determinants of firms’ target leverage ratios vary considerably across countries.33 However, Rajan et al. (1995) and Antoniou et al (2008) identify a core set of deter- minants34 appearing to influence the capital structure of U.S. firms that is fairly simi-lar in the major industrialized countries including Germany.35 They suggest employ- ing a four-factor model that has often been used as benchmark model in the litera- ture.36 The four factors that are used as independent variables and proxies for the benefits and costs of leverage are growth, profitability, tangibility, and size.37 How- ever, Frank and Gayol (2009) illustrate that using only these four variables and omit- ting others can materially change inferences on other factors that are included in the model. They found a model with a set of six determinants that are reliably important and robustly to explain more than 27 % of the variation in leverage, whereas the 19 remaining factors only explain additional 2 % of the variation.38 For this reason, I include expected inflation and the median industry debt ratio. The following para- graphs provide a brief overview of these determinants included in my model and the expected effect on leverage39 according to the classical capital structure theories mentioned in Section 2:40

A typical measure of growth (-) is the market-to-book asset ratio.41 I expect an in- verse relation between leverage and growth opportunities for two reasons. First, ac- cording to Myers (1977), it is more likely for high-levered firms to pass up profitable investment opportunities than for low-levered ones. For this reason, firms expecting high future growth should prefer equity financing.42 Second, the trade-off theory pre- dicts that growing firms face higher costs of financial distress. Thus, managers of those companies tend to replace debt for equity resulting in lower leverage.43

The impact of profitability (-) is ambiguous. According to the trade-off theory, more profitable firms face lower costs of financial distress. Therefore, one would expect that profitability is positive correlated with leverage.44 However, the pecking order theory predicts that firms prefer to finance new investments with internal funds ra- ther than debt. Therefore, more profitable firms will become less levered over time,holding investments and dividends constant. Following previous empirical studies supporting the pecking order theory, I expect to obtain a negative relationship.45

Both theoretical and empirical studies argue for the relevance of a firm’s size (+) as a determinant of target leverage. The impact of size is not clear as well. One side, the pecking order theory predicts an inverse relation. Larger firms are better known in public and provide more information. This reduces informational asymmetries and should increase the preference for equity relative to debt. On the other side, a positive relationship between the size of a firm and its leverage is expected because generally larger firms tend to be more diversified, and thus, are less likely to fail than smaller ones. This is consistent with the trade-off theory. Previous empirical studies provide evidence for the significance of the trade-off theory.46

Tangibility (+) is the ratio of net property, plant and equipment to total assets. The greater the proportion of tangible assets, the higher should be the leverage for two reasons. First, tangible assets can serve as collateral for debt financing, thereby lowering bankruptcy costs. Second, in case of bankruptcy tangible assets tend to retain more value than intangible ones. Therefore, lenders are more willing to supply loans to firms with a higher proportion of tangible assets.47

I expect the industry median leverage (+) to be positively correlated with leverage because managers usually use the industry median leverage as benchmark to which they adjust.48 Thus, a higher industry median leverage should lead to a higher leverage.49 This is in line with the trade-off theory.

Moreover, one macroeconomic factor that should be included in my model is ex- pected inflation (-). It is assumed that firms tend to have high leverage when the in- flation is expected to be high. This indicates a positive relationship between leverage and expected inflation.50 However, I recognized after running the first-stage regres- sion that this factor is not significant at the 10 % level.

[...]


1 Cf. Miller, M. et al. (1958), p. 267-271.

2 See for example Alonso, P. de A. et al. (2005), p. 391-407.

3 See for example Rajan, R. et al. (1995), p. 1421-1460, Titman, S. et al. (1988), p. 1-19.

4 Cf. Drobetz, W. et al. (2006), p. 941-942, Heshmati, A. (2001), p. 1, Strebulaev, I. (2007), p. 1748.

5 Cf. Elsas, R. et al. (2008), p. 26-30, Huang, R. et al. (2009), p. 239.

6 See for example Cook, D. et al. (2009), p. 6-8, Flannery, M. et al. (2004), p. 472, Hovakimian, A. et al. (2001), p. 5-6.

7 Cf. Welch, I. (2006), p. 18.

8 Cf. Elsas, R. et al. (2008), p. 6-7, Huang, R. et al. (2009), p. 237-238.

9 Cf. Flannery, M. et al. (2006), p. 471.

10 Cf. Myers, S. (1984), p. 577-581.

11 Cf. De Haas, R. et al. (2006), p. 135, Elsas, R. et al. (2008), p. 4, Fama, E. et al. (2002), p. 1, Myers, S. (1984), p. 577.

12 Cf. Fama, E. et al. (2002), p. 6.

13 Cf. De Haas, R. et al. (2006), p. 135.

14 Cf. Antoniou, A. et al. (2008), p. 72, Kayhan, A. et al. (2004), p. 27, Titman, S. et al. (2007), p. 404, Welch, I. (2006), p. 19.

15 Cf. Myers, S. (1984), p. 577-578.

16 Cf. Barclay, M. et al. (2005), p. 8-17, Dudley, E. (2008), p. 3, 68. Leary and Roberts (2005) distin- guishes between three types of adjustments costs. Adjustments costs may be fix, proportional, or fix plus proportional. Cf. Leary, M. et al. (2005), p. 2578-2580.

17 These factors are discussed in Section 3.3.

18 Cf. Elsas, R. et al. (2008), p. 6, 26, Fischer, E. et al. (1989), p. 20.

19 Cf. Myers, S. et. al (1984), p. 187.

20 Cf. Frank, M. et al. (2003), p. 220, Myers, S. (1984), p. 581, Myers, S. et. al (1984), p. 187.

21 Cf. Baker, M. et al. (2002), p. 25-27, Frank, M. et al. (2009), p. 6.

22 The following description refers to Flannery, M. et al. (2006), p. 472.

23 For example, Papke and Wooldridge show that there are methodological problems using linear models for fractional data. They use the quasi-maximum likelihood method to estimate the firms’ desired target leverage. Cf. Papke, L. et al. (1996), p. 619-632.

24 I use Datastream’s four digit numeric code assigned to each company to represent its industry group. A major industry group is represented by the first two digits.

25 Cf. Cook, D. et al. (2009), p. 7, Drobetz, W. et al. (2006), p. 944.

26 L öö f (2004) points out that unanticipated changes in the economic environment may be an explana- tion for overadjustment. Cf. Lööf, H. (2004), p. 458.

27 Cf. Cook, D. (2009), p. 8, Frank, M. et al. (2009), p. 11-12, Lööf, H. (2004), p. 455.

28 Cf. Drobetz, W. .et al. (2006), p. 950, Flannery, M. et al. (2006), p. 471-472.

29 Cf. Cook, D. (2009), p. 8, Flannery, M. et al. (2006), p. 471.

30 Cf. Cook, D. (2009), p. 8.

31 Cf. Flannery, M. et al. (2006), p. 499.

32 Cf. Roberts, M. (2000), p. 3.

33 Cf. De Haas, R. et al. (2004), p. 162. This could be explained by different institutional characteris- tics. Cf. Wald, J. (1999) p. 184.

34 Including growth, profitability, size, and tangibility.

35 Cf. Antoniou, A. et al. (2008), p. 59, Rajan, R. et al. (1995), p. 1421.

36 Cf. Frank, M. et al. (2003), p. 219.

37 Cf. Rajan, R. et al. (1995), p. 1451, Kayhan, A. et al. (2004), p. 14.

38 Cf. Frank, M. et al. (2009), p. 3.

39 The expected sign is written in brackets.

40 Appendix 3 provides the exact formula for each measure.

41 Adam and Goyal (2008) show that the market-to-book asset ratio is the most reliable proxy for growth opportunities. Cf. Adam, T. et al. (2008), p. 55.

42 Cf. Myers, S. (1977), p. 149-155, Titman, S. et al. (1988), p. 4.

43 Cf. Cf. Antoniou, A. et al. (2008), p. 62.

44 Cf. Elsas, R. et al. (2008), p. 10.

45 Cf. Cf. Antoniou, A. et al. (2008), p. 62, Cook, D. (2009), p. 10, Frank, M. et al. (2009), p. 7, Myers, S. et al. (1984), p. 187.

46 Cf. Antoniou, A. et al. (2008), p. 64, Frank, M. et al. (2009), p. 7-8, Hovakimian, A. et al. (2001), p. 7, Rajan, R. et al. (1995), p. 1451, Titman, S. et al. (1988), p. 5-6.

47 Cf. Antoniou, A. et al. (2008), p. 63-64, Elsas, R. et al. (2008), p. 9, Frank, M. (2009), p. 9, Hova- kimian, A. et al. (2001), p. 8, Rajan, R. et al. (1995), p. 1451, Titman, S. et al. (1988), p. 3.

48 Cf. Elsas, R. et al. (2008), p. 10, Hovakimian A. et al. (2001), p. 14.

49 Cf. Elsas, R. et al. (2008), p. 10, Frank, M. (2009), p. 9.

50 Cf. Frank, M. et al. (2009), p. 3-4.

Details

Pages
37
Year
2010
ISBN (eBook)
9783656082347
ISBN (Book)
9783656082514
File size
898 KB
Language
English
Catalog Number
v183708
Institution / College
University of Regensburg
Grade
1.3
Tags
target german

Author

Share

Previous

Title: Target leverage and capital structure adjustment speed across German industries