The Determinants of Capital Structure
A search for the factors which may affect the composition of the
capital structures of companies
The Determinants of Capital Structure
A search for the factors which may affect the composition of the
capital structures of companies
Henryk Jansen April 24, 2006 Groningen Rijksuniversiteit Groningen Student # 1198157 Dr. Ir. N. Brunia
To all the people who have been waiting for this, especially J & J.
Cover: “Early seesaw” by Ritchie. Used with kind permission from www.worth1000.com. The image is meant to refer to the balance between the pro’s and cons of debt and equity.
Henceforth his might we know, and know our own So as not either to provoke, or dread
New war, provok't; our better part remains To work in close design, by fraud or guile
What force effected not: that he no less At length from us may find, who overcomes
By force, hath overcome but half his foe. John Milton, Paradise Lost
Preface
For several years, I’ve been taught a lot about economic principles in ideal situations. However, the fact that real life might be different wasn’t always highlighted as is maybe should have been. Theoretical knowledge is indispensable of course as a base for further analysis, but so is the knowledge of when it is applicable and when the boundaries of theory are being crossed by practice.
It was only in the final stage of my study that a substantial part of the program was addressed to the implication of financial theories into practice. As this was in the area of corporate valuation and I was pleasantly surprised by the challenging character and broadness of the problematic, I decided I wanted to go further into the subject and write my thesis about it. Now, corporate valuation is one of the most debated issues in finance, probably because it involves the most money and because it influences and is influenced by the core of corporations, namely their operations. Searching for a specific subject that is both relevant and sufficiently marked out to enable a thesis like this is like walking in a jumble-shop seeing all kinds of interesting items but being allowed to take home only one. Therefore, I had to make a choice and decided to analyze what exactly determines the differences in capital structure of companies which can be observed in practice.
Based on the theories studied in the past, certain relationships can be assumed but the question remains to what extend these are reflected in practice. As it is such an extensively analyzed and debated subject, there is a lot of literature about it. However, as every author has its own way of researching and its own methodology, the exact way to model the analysis was a true challenge and, as in every thesis, changed till the last moment. The result of this search is of course the thesis as you see it before you.
Once the subject was clear to me, throughout the process of writing I have never doubted this thesis to be successfully completed. As most students however, I also have experienced the difficulties of analyzing and writing with which such a project is of course surrounded. The hardest part to me was not the collection and analysis of relevant literature or my own data, but to work on it on a regular base. To settle myself down to the necessary tasks seems often more difficult than assumed. This should not distract however from the time it took to get everything lined up as I wanted it to be; especially some data took long to be made consistent and suitable for use.
All in all, it has been a very good experience for me, which has increased both my knowledge about financing issues as my practical skills. The subject and results may not be pathbreaking, I do think this thesis to be a relevant addition to the current literature on capital structures. In any case, it was made with the utmost care and intended to be in accordance with other, much more experienced, economists.
To conclude, some general remarks: First, I have written this thesis in English, though both Dutch and English were allowed. I’ve chosen for English because I wanted to improve my writing skills and due to the subject many terms would be English anyway. Moreover, I thought the subject to deserve an international approach. I did not experience any difficulties with the language, though sometimes expressing myself rightly was a little harder and the chance of repeating certain expressions increased a little. All in all, I think it worked out very well. Second, for the references, I followed the style of Grinblatt and Titman (2002). In many articles, even in very respectable journals, I found inaccurate references. As I experienced this to be irritating and frustrating occurrences, I have tried to avoid making the same mistakes (and others as well of course) to the best of my abilities.
Finally, I would like to thank all the people who helped me finishing my study: my parents, brother, family, friends, fellow students and professors, especially Mr. Brunia, who, sometimes even without them knowing it, encouraged me to overcome the hurdles and continue the struggle towards the greater goal.
Contents
Preface... 4
Contents... 6
1. Introduction ... 7
2. Theories behind the determination of capital structures ... 10
2.1. Introduction ... 10
2.2. The Modigliani-Miller theorem ... 10
2.2.1. Introduction... 10
2.2.2. The introduction of debt... 11
2.2.3. Loosening basic assumptions and drawing new conclusions... 12
2.3. The trade-off theory... 14
2.3.1. Background ... 14
2.3.2. The benefits and costs of debt ... 15
2.3.3. The interaction between stakeholders and the level of debt ... 16
2.3.4. Concluding remarks... 18
2.4. The pecking order theory ... 19
2.5. Criticism of Welch and his “inertia” hypothesis... 21
3. Empirical analysis ... 24
3.1. Introduction ... 24
3.2. Methodology ... 25
3.2.1. The model ... 25
3.2.2. Measuring the debt-ratio... 25
3.2.3. Variables ... 27
3.2.3.1. Evidence in financing literature on the determinants of capital structure ... 27
3.2.3.2. The variables included in the regression analysis ... 28
3.3. Data description... 31 4. Results ... 32 4.1. Balances... 32 4.2. Debt-ratios... 35 4.3. Industry differences ... 37 4.4. Variables ... 39
4.5. Concluding remarks on results regression analysis... 43
5. Summary and Conclusions... 44
1. Introduction
The composition of the capital structure of companies is a much debated subject in finance, even the relevance of the subject itself has been questioned. In their pioneering and pathbreaking article, Modigliani and Miller (1958, MM) developed a theorem in which, under certain conditions, the composition of capital structure is irrelevant for the value of companies. However, more sophisticated theories have been developed since, which has led to a better understanding and explanation of the practice of financing, and which are being supported by empirical research. The trade-off theory provides a broadly accepted framework for a causal relationship between capital structure and firm value.
Given this causal relationship, the question remains of course what determines the composition of these capital structures. It is this question that is the subject of this thesis: “What determines the composition of the capital structures of companies?”
It may be assumed that if all firms would be equal, all would have the same capital structure (which is optimal for those firms). However, in practice nearly no firms have identical capital structures, so there must be reasons why some firms have more or less debt than others. These reasons must be sought in the differences between firms. The question therefore is in which respect companies differ in practice and which consequences this has for the composition of capital structure. In other words: with which variables is capital structure related and what is the nature of these relationships?
There are three important theories that explain the financing behavior of firms that lead to the particular capital structures: the trade-off theory, pecking order theory and principal-agent theory.
According to the trade-off theory, firms distinct and weigh arguments to increase or decrease their level of debt, with the shareholder’s value of the firm as the decisive variable to be maximized.1 Some factors make firms want to increase their debt-ratio, others should lead to a lower debt-ratio. Corporate taxes for example make it attractive to lever a firm’s assets as interest payments are tax deductible. By subtracting a firm’s interest payments from its earnings, a so-called tax shield is created, lowering total cash outflows and therefore increasing market value. However, there is also a downside to debt: as the level of debt increases, so does the risk of default (i.e. the risk that a firm can no longer meet its interest and principle payment obligations) and the related costs of financial distress. The trade-off theory not only explains why market value is affected by the composition of capital structure, it also suggests possibly explanatory variables for the composition itself.
The presence of costs due to having either too little (less tax shield) or too much debt (default risk) raises the question whether or not an optimum could be determined in which a balance is found between debt increasing and debt decreasing considerations. Such an optimal debt-ratio could be applicable to all firms equally, or differ per industry as various industries have different characteristics which could affect optimal debt-ratios. The travel and leisure industry for example is found to have very low debt-ratios, while firms in the personal and household goods industry in general appear to have high levels of debt.2
1 As shareholder’s value should always be the standard on which all company decisions are ultimately based. See
also Jensen (2001).
As the composition of capital structure affects the market value of firms, managers and analysts have a great interest in knowledge about the determinants of capital structure. When these determinants are known, optimal capital structures could be determined. As investors have knowledge of these optimal capital structures and markets are efficient, it is to be expected that deviations from it will be negatively reflected in the value of a firm. Therefore, firms with other than optimal capital structures will have lower market values and, more importantly, are able to increase their value by changing their capital structure. This knowledge could give investors better insight in the correctness of current share prices and also information about the expected direction for future share prices (information most highly valued by virtually all human inhabitants of this planet). As in most cases where large amounts of money are involved, a lot of theoretical and empirical research has been (and is still being) done in this respect.
As each theory has its alternative, not surprisingly this is also the case for the trade-off theory. The pecking order theory offers one alternative explanation for the financing behavior of companies. It is not completely contrary to the trade-off theory, but more of an addition. Where the trade-off theory explains the total capital structure of a company as a whole at a certain moment, the pecking order theory focuses on the financing of new investments projects.3 The pecking order theory states that firms want to finance new investments in such a way that the total costs to the firm are as low as possible in order to maximize shareholder value. Therefore, firms favor retained earnings for financing investments. If these are insufficient, the firm issues new debt; equity issues are the least favorable source of financing. This pecking order is largely based on information asymmetries between managers and either shareholders or debt-holders, also subject of the principal-agent theory.
Of course there are always people who question all (rational) explaining theories. Ivo Welch (2002) for example has developed an “inertia” theory which basically states that managers fail to react to external changes in capital structures (changes in the share price) and therefore there would be no explanation for current capital structures except for their past composition. In addition, Miller (1977) has stated that managers have no objective economic rationales at all for their financing decisions. This would not matter much though, because it does not affect the value of companies, as stated in his earlier work with Modigliani.
In this thesis, the significance of various possible explanatory variables put forward by the trade-off, pecking order and principal-agent theories is being tested in a linear regression model using information about public Dutch firms during the 1995-2004 period. This method is in line with i.a. Booth et al. (2001) who state that “recent empirical research has focused on explaining capital structure choice by using cross-sectional tests and a variety of variables that can be justified using any or all of the three models” [i.e. the pecking order, principal-agent and trade-off theories].4
The results will be compared with those reported in various similar studies, e.g. Rajan and Zingales (1995) and Titman and Wessels (1988). It is found that the composition of capital structures in the Netherlands does not substantially deviate from those in other countries (not only industrial but developing countries as well). On average, Dutch firms were found to have a debt-ratio of 35% for book values and 24% using market values. These numbers however were found, as expected, to differ substantially between industries. Furthermore, it is found that the included variables have a significant influence on the debt-ratios of companies, all of them in the way predicted by the current finance theory (though not all were statistically
3 So, where the trade-off theory assumes managers to be constantly assessing the complete capital structure, the
pecking order theory only assumes managers to seek the least expensive source of financing when new investments are necessary.
significant). Tangibility of assets and for example is positively related to the debt-ratio, whereas profitability has a strong decreasing effect. When on average a firm with certain characteristics may be expected, based on these relationships, to have a certain capital structure, and the actual capital structure deviates from the expected, there must be either another satisfying explanation which justifies this deviation or the capital structure should be adjusted when this increases shareholder’s value.
The rest of this thesis is being organized as followed: First, the theories behind the determination of capital structure will be described. Second, the model, included variables and data will be described. Third, the results will be presented. Finally, a conclusion will be drawn.
2. Theories behind the determination of capital structures 2.1. Introduction
When trying to analyze what determines capital structure, three theories which give insight in financing decisions are the trade-off theory, the principal-agent theory and the pecking order theory. These will be described below and it will be shown how the variables that possibly affect capital structure proceed from these theories.5 As the development of the relevant
theories is to some extent connected with the work of Modigliani and Miller (MM, 1958, 1963), their theorem will be described shortly first. This will give a better understanding of why the theories have been developed and insight in the background of the direction of their development.
2.2. The Modigliani-Miller theorem 2.2.1. Introduction
In all sciences, there can be appointed some “decisive moments”: discoveries, events, insights or publications which mark their origin or which have fundamentally changed people’s opinions about a certain subject, and thereby have shaped the current state of knowledge. For most sciences, these breakthroughs have occurred at least decades, but in most cases rather centuries (if not millennia) ago. Sometimes they were recognized immediately as decisive, in which case they were often controversial at the time; in other cases their importance was noticed only in retrospect. Naturally, decisive moments in more specific areas have occurred more and more recent.
For modern financing theory, Franco Modigliani and Merton Miller’s 1958 article on capital structures, i.e. the way in which a company is being financed6, can be considered one of those “decisive moments”. Their propositions about the irrelevance of debt for the value of companies are among the very first things current finance students are being taught. However, as an enormous amount of research has been done since, extending and deepening our knowledge, it becomes clear that the real world is much more complicated than assumed by Modigliani and Miller (MM). Some people say that for economists, real life is a special case; though this is true to a certain degree, even economists themselves feel sometimes a little awkward about the distance between certain assumptions that are being made and practice. As students become more and more inclined to interpret MM’s propositions as being unrealistic, as I have personally experienced, chances are their theories are being ridiculed. Miller himself (1977) states about this: “The arbitrage proof of this proposition can now be found in virtually every textbook in finance, followed almost invariably, however, by a warning to the student against taking it seriously.”7 However, after a more careful consideration of their theorem, it becomes clear that there is more to it than some students may think.
MM assess the effects of capital structure on the cost of capital and market value of firms. This has had major consequences for the financing theory developed so far, and gave an enormous impulse to the further development of financing theory and the practice of financial research. Before MM, it was assumed that the return on investments in physical assets are certain and known to the investors of the firm. Under this kind of certainty, the rate of return
5 The relationship between the trade-off, pecking order and principal-agent theory is complex; elements from
these theories partly overlap and supplement each other. The principle-agent theory is sometimes considered to form a framework for the other two theories, while it can also be argued that it is only a part of either theory. Here, the trade-off and pecking order theory are discussed separately, supplemented with the relevant elements from the principal-agent theory.
6 See Appendix 1 for a list of abbreviations and an explanatory vocabulary. 7 Miller (1977): p. 262.
to the shareholders should equal the market rate of interest on bonds, as both have the same financial risk characteristics.8 Because of this, the market value of a firm and the average cost of capital under certainty are independent of the capital structure of the firm. In order to adjust the model for uncertainty, the “traditional” view was to increase the market rate of interest with a risk premium, or decrease the rate of return on investments with a risk discount to a “certainty equivalent” yield. However, the determinants of the size of this discount and the relationship with other variables had not been clearly and fully explained and its estimation therefore was highly subjective. This has led MM to develop a new model.
According to MM, the market value of a company is independent of its financing structure (also under uncertainty). MM therefore deny the possibility of an optimal capital structure as put forward by the traditional view. For their theory, MM assume firms can be divided into different “equivalent return classes.” When equity is the only source of financing, each firm’s expected return on shares is proportional to any other firm in that class, and all expected returns share the same probability distribution. Therefore, the return on investment of firms in the same class can only vary in level (as the investments of each firm can differ in level), but not in rate. Hereby it is assumed that the distinction between income to the firm (retained earnings) and income to the shareholders (dividends) can be neglected. This is true when management acts in the best interest of the shareholders and in the absence of taxes and transaction costs, as in that case shareholders are indifferent between the firm making investments or extracting the profits from the firm in the form of dividends and investing these in another firm of the same class. Both would give the same rate of return and have the same financial risk, and therefore shares of firms in the same class can be assumed to be perfect substitutes. For this reason, given perfect markets, all shares in the same class will have to sell at the same price per dollar’s worth of expected return.9
2.2.2. The introduction of debt
With the introduction of debt, two sources of financing become available to the firm. This has major complications for both the capital structure of the firm and the market for shares. Assuming bonds provide a constant, certain income per period, and are traded in perfect markets and the absence of taxes, all bonds can be considered to be perfect substitutes for each other. Therefore, all bonds will sell at the same price per dollar’s worth of return and have the same rate of return (the interest rate r), just like shares in the situation described above. In this case, investors could duplicate the characteristics of a levered firm at no costs by making a portfolio of shares in an unleveraged firm and privately borrowing in the bonds market. Therefore, they are not willing to pay more or less for levered firms than for unleveraged firms, and both will have to provide the same rate return. Any deviations will lead to arbitrage opportunities that will automatically correct the arisen inequalities. This leads to MM’s famous Proposition1:
“The market value of any firm is independent of its capital structure and is given by capitalizing its expected return at the rate appropriate to its class.”10
Equally, it can be stated that “the average cost of capital to any firm is completely independent of its capital structure and is equal to the capitalization rate of a pure equity stream of its class.”11
The compensation for equity and debt investors, however, fundamentally differs in character. Bondholders, in contrast to shareholders, are entitled to a fixed and sure stream of (interest)
8 Thus, the average cost of capital equals both the rate of return to shareholders and the market interest rate, and
is independent of the amount of leverage.
9 MM (1958): p. 267. 10 MM (1958): p. 268. 11 MM (1958): p. 269.
payments, thereby changing the expected net returns to the firm in comparison with unleveraged firms. The introduction of debt therefore lowers the expected returns to shareholders and increases the financial risk of the firm. As firms in the same class can now have different debt-ratios, they no longer share the same probability distribution of returns and therefore do no longer have the same financial risk characteristics. Their shares are no longer “homogeneous” or perfect substitutes, and therefore can trade at different prices. As a consequence, shares of firms in a world with or without debt differ by the effect of debt on the net returns to the firm. According to MM’s Proposition 2:
“The expected yield of a share of stock is equal to the appropriate capitalization rate for a pure equity stream in the class, plus a premium related to financial risk equal to the debt-to-equity ratio times the spread between [the average cost of capital and the interest rate].”12 In other words, the change in the rate of return on stocks as a consequence of the introduction of debt is exactly the amount at which the effects of the introduction of debt are neutralized, so that the average cost of capital stays constant. This differs fundamentally from the traditional view in which each firm has an optimal capital structure where the average cost of capital is the lowest. According to MM, no such optimal capital structure could exist unless there would be “large and systematic imperfections in the market which permanently bias the outcome.”13
Where MM state that the yield of common shares will decrease for higher debt-ratios (because the average cost of capital has to remain constant), the traditional theory states that this yield will initially remain constant and even sharply rise for high debt-ratios. One reason advocates of the traditional view give for this, is that as long as the debt-ratio lies within a certain “normal” range, increases in the level of debt do not increase the riskiness of the shares, as the firm is sufficiently healthy to meet its interest payment obligations. Another reason is that certain large institutional investors are restricted to investments in bonds, through which firms will be able to borrow at more favorable terms than in a perfect market. Only at high levels of debt would the financial risk of the firm increase to such a degree that the shareholders would be enforced to demand a higher rate of return on their shares as compensation for the increased risk. These differences are explained by the determination of dependent and independent variables. In MM’s line of reasoning, the average cost of capital (ACC) is determined independently of its components and then each of the components (the rate of return on shares i and the interest rate r) are determined dependent on the ACC (r as an exogenous variable and i as the closing variable): i = ƒ(ACC, r). This is contrary to the (traditional) method in which a variable is constructed as a function of its independently determined components: ACC = ƒ(r, i).
2.2.3. Loosening basic assumptions and drawing new conclusions
Hitherto, MM’s propositions have been based on three basic assumptions: the absence of taxes, the existence of only one kind of bond and one interest rate, and perfect capital markets. While the latter will remain unaffected, loosening the two other assumptions makes the MM theorem more realistic and has important consequences for the capital structure of a company. In their original paper, these consequences were underestimated, as MM later admitted. Therefore, they published another paper, in which some of their initial statements were corrected (MM, 1963).
Based on propositions 1 and 2, MM had developed a “theory of investment”, summarized by Proposition 3:
12 MM (1958): p. 271.
“The cut-off point for investment in the firm will in all cases be [the capitalization rate of a pure equity stream of its class14] and will be completely unaffected by the type of security used to finance the investment.”15
However, it has become clear that both claims do not hold when taxes and multiple sorts of debt are being introduced. This is because the average cost of capital was proven to be affected by the tax rate and the amount of debt in a firm.
The initial assumption of the existence of only one kind of bond and one interest rate being applicable to all firms is not very realistic. In the real world, a large variety of bonds is available to a firm. Practically, a bond’s interest rate is dependent of its maturity, the provided collateral, and the risk (financial, business, country) of the borrowing firm. According to MM, without identifying the underlying logic, the level of the interest rate varies, ceteris paribus, proportionally with the debt-equity ratio (D/S) of the borrowing firm.16 MM state this does not affect Proposition 1, as investors can still undo the effects of the additional debt by changing their personal investment portfolio. However, they explicitly ignore the tax effects of such an increase in the level of debt. But even without this omission their statement does not seem to hold. According to MM, the average cost of capital should still be the same for all firms in the same class. This would mean that if interest rates rise because of higher debt levels, the yield of equity should be reduced. MM state that “if r increases with leverage, the yield i will still tend to rise as D/S increases, but at a decreasing rather than a constant rate.”17 At high levels of debt, when r becomes very large as well, the yield will even start to decrease in order to keep the average cost of capital constant. The only reason MM give for this being accepted by current shareholders, is because they could offset the lower returns on their shares by adding the issued high-interest bonds to their personal investment portfolio. However, if the demanded return on shares is assumed to be determined independently of the average cost of capital by autonomous investors, it follows that the average cost of capital will necessarily change with changes in the debt-ratio of a firm, for both the level and the relative importance of the interest rate will change.
The most important problem with their initial propositions, however, arises because of the existence of taxes and the resulting deductibility of interest payments, changing a firm’s taxable profits and thus the expected return to its shareholders. When a firm can deduct its interest payments, net interest rates will decrease and higher levels of debt will increase the tax shield, thereby decreasing taxable net profits and thus tax payments. MM admit: “With a corporate income tax under which interest is a deductible expense, gains can accrue to stockholders from having debt in the capital structure, even when capital markets are perfect.”18 This is a crucial statement, as it shows that it had become clear in their first paper
already that capital structures do affect the (market) value of a firm. In retrospect, it seems that, as a result, Proposition 1 is no longer valid or relevant. However, this conclusion is (of course) not drawn in their first article. Although they do state that “certain interpretations [of Proposition 1 and 2] must be changed”19, they note that the gains of tax deductions are relatively small.
In their second article MM had to admit that the tax advantages of debt are considerable, and they show that the average cost of capital is also affected by the tax rate, the interest rate and the debt-ratio. An explanation for this new view is the recognition that the tax advantages of debt are not solely due to the deductibility of interest payments from taxable income, but that
14 Which equals the average cost of capital, according to MM. 15 MM (1958): p. 288.
16 r = r * (D / S). MM (1958): p. 273.
17 MM (1958): P. 274-275. See also Appendix 2. 18 MM (1958): p. 294.
the nature of the various cash flows play an important role as well.20 This thought has led to one of the basic equations in current financing theory, namely that the value of a levered company equals the value of the company in case it was unlevered plus the value of the tax shield. The value of the firm in the case it was unlevered can be calculated by discounting the expected earnings after taxes at the market capitalization rate for unlevered companies with otherwise the same characteristics. The value of the tax shield can, under certain assumptions,21 be calculated by multiplying the interest payments with the marginal tax rate and discounting this at the cost of debt. The equation should hold because otherwise arbitrage opportunities would be created.
As the average cost of capital is affected by both the tax rate and the debt level, Proposition 3 remains true only in the sense that each investment should have a rate of return higher than the cut-off point in order to act in the best interest of the current shareholders and increase the market value of the firm. This point however is not “completely unaffected by the type of security used to finance the investment”22, but is known as the Weighted Average Cost of
Capital (WACC), and depends on the return on equity, the interest rate, the tax rate and the debt-ratio.23 This is of major importance because, as a consequence, when the rate of return on shares exceeds the market interest rate (which is a reasonable assumption), the value of the company will increase when the amount of debt increases.24 This means that a firm could increase its market value by adding more debt and would lead to the conclusion that firms should strive for an optimal capital structure with a debt-ratio where the average cost of capital is the lowest.
2.3. The trade-off theory 2.3.1. Background
The MM theorem has had a major effect on the development of financing theory; their propositions about the irrelevance of capital structures have been subject of much debate ever since. Especially the adherents of the “traditional” view had difficulty with the complete independence between capital structure and market value. They started to improve the traditional view by building a better theoretical foundation and put forward arguments that supported their own view instead of MM’s (initial) theorem.25 This has resulted partly in what is now known as the “trade-off theory” of capital structure.
The trade-off theory states there are benefits and costs to the use of debt in a firm. These benefits and costs consist in essence of changes in the market value of a company proceeding from both the direct financial consequences of the use of debt (like changes in principle and interest payment obligations) and the consequences for the firm of the behavior of stakeholders in the company as a result of financing decisions. Because of these latter consequences, changes in the capital structure affect not only the financial position of a firm, but its operational performance as well. According to the trade-off theory, companies, considering all consequences of (changes in) capital structure and maximizing shareholder value, make a trade-off between the benefits and costs of debt, such that the benefits of debt
20 Certain and uncertain cash flows have to be discounted at different rates, and debt is in general less risky to
investors than equity.
21 See Grinblatt and Titman (2002): p. 465. 22 MM (1958): p. 288..
23 WACC = r
e * D / (D + E) + (1 – t) * rd * D / (D + E).
24 I.e. the average cost of capital, with which the cash flows are discounted, decreases when more of the less
expensive debt is added, because of which the level of discounted cash flows will increase. However, at the critical debt-ratio the costs of equity will start to rise sharply, thereby increasing the average cost of capital.
are utilized and the costs limited as much as possible, and (are inclined to) compose their capital structure accordingly. This is illustrated in Appendix 3.
Because of the assumed importance of the costs and benefits of debt, it is necessary to know what these benefits and costs of debt are and how they are related to the behavior of the stakeholders of a firm. This will be discussed in the two following paragraphs. Resulting from this, it becomes relevant to know which variables affect the level of debt, because this is related to a firm’s ability to increase its debt-ratio in order to utilize the benefits of debt, and to the limits of a firm’s ability to borrow because of the negative impact on the market value of a firm. These variables and their relationship with the debt-ratio will be analyzed and empirically tested in chapter 3.
2.3.2. The benefits and costs of debt
The benefits of debt consist foremost of the tax shields created because of the deductibility of interest payments. When interest payments are deductible, companies are allowed to offset them against taxable income in the same year. Therefore, the taxable income, on which taxes payable is based, becomes lower, decreasing the tax obligations of the company. When there is no taxable income of which the interest payments could be subtracted or if this would result in negative taxes, companies are (under specific conditions)26 allowed to carry them backward or forward against taxable income in other years. This tax advantage of debt is being referred to as the “tax shield”. Because of this, the total cost of debt to the firm becomes lower and (future) cash-flows are being increased. Shareholders will incorporate this in the share price of the firm, thereby increasing its market value. Because of the importance of the tax shield for the trade-off theory, variables that affect the tax-shield are likely to be important in the determination of capital structures as well.27 Another benefit of debt is what Jensen and
Meckling (1976) call “the incentive effects associated with highly leveraged firms.”28 This will be shown in the next paragraph.
The suggestion, based on the benefits of debt, that companies would have to be entirely financed with debt, in order to take full advantage of the deductibility of interest payments, neglects the negative aspects of debt.29 First, there are of course the interest payments themselves, which decrease the net profits of the firm. Second, the costs of debt consist of the costs of financial distress and bankruptcy costs. These are not the same:30 bankruptcy costs occur only at the exact moment of bankruptcy, are relatively well observable in practice, relatively fixed, and borne by both debt- and equity-holders. These costs could consist of i.a. the legal and administrative costs of default and the costs of reorganization. The costs of financial distress, however, usually appear long before factual bankruptcy, increase as the company approaches the bankruptcy, are less objectively observable, are affected by many variables specific to the company and are mostly borne by the shareholders. These costs generally consist of behavioral consequences of the threat of bankruptcy, and contracting and monitoring costs.
Principal and interest payments have to be paid out of the revenues of the firm, which fluctuate over time. The remainder is for the current shareholders of the company. But where
26 These conditions vary per country and over time.
27 John Graham has published some important articles on various aspects and consequences of corporate tax
rates, like international differences between corporate tax rates and the complexities of the measurement of marginal corporate tax rates. Companies (mostly multinationals) vary in their effective (marginal) corporate tax rate, e.g. because of individual tax credits or possibilities to shift funds between subsidiaries (especially abroad).
28 Jensen and Meckling (1976): p. 334.
29 See e.g. Scott (1976): p. 34, Ross (1977): p. 24 and Bradley, Jarrell and Kim (1984): p. 857. 30 See e.g. Babenko (2003): p. 2, 7.
the level of the latter is uncertain and has an optional character31, the compensation for debt holders is obligatory.32 Therefore, when these fixed payments are being increased more and more, not only the expected returns available for the shareholders become smaller, but the risk that the debt obligations will exceed the revenues (default risk) increases. Of course, the more revenues fluctuate and the higher the fixed payments, the higher this risk becomes. If the debt-ratio increases too much, this could make bondholders worry about the principal and interest payments. Because of the seniority of certain kinds of debt over others, each debtor, in order to secure full repayment, will try to be the first to withdraw his loans, ultimately leading to bankruptcy. This financial distress will lead to lower market values as shareholders fear for lower returns or even losing their investments. Not surprisingly therefore, earnings and cash flow volatility was reported by Graham and Harvey (2001. p. 210) to be an important decision making criterion for CFOs. Because of the importance of financial distress, the relationship of variables that affect financial distress with capital structure has to be analyzed. Financial distress is not only related to debt- and equity-holders however, other stakeholders are being affected by it as well and will react accordingly, as will be shown in the following subparagraph.
2.3.3. The interaction between stakeholders and the level of debt
In addition to the direct financial consequences described above, the trade-off theory covers the interaction between various stakeholders and (changes in) the capital structure. This relationship works both ways: on the one hand are stakeholders being affected by (changes in) the capital structure of companies. This particularly concerns the certainty about the performance of concluded contracts, one of the basic requirements for business to take place. Their behavior in reaction to this has important consequences for companies. On the other hand do firms reckon with the behavior of their stakeholders when making financing decisions. Capital structures are in turn being adjusted in anticipation of, or in reaction to, the (expected) behavior of various stakeholders. Relevant stakeholders in this respect are i.a. suppliers, employees, customers, competitors and management.
Because suppliers and employees are for their own financial well-being more or less dependent on the ability of firms in which they have a stake to meet their financial obligations, suppliers and employees have a natural interest in the financial situation of these firms. As the compensation of suppliers and employees for supplied services is (in most cases) monetary, they react to changes in the financial position of firms in which they have a stake. As shown above, debt increases the level of fixed payments, and thus decreases the financial flexibility of companies. Because the possibility of difficulties with the payment of bills and salaries and the risk of bankruptcy therefore increases, this may cause suppliers and employees of a firm to sever their relationship with the company. Firms will then no longer be able to attract the best suppliers, credit facilities and employees. In addition, employees with performance-based contracts based on financial objectives could get a lower incentive to work because their goals become unattainable. This could lead to a decrease in productivity and product quality, which in turn will decrease sales and cash flows and therefore the market value of firms. Companies want to avoid this of course, and therefore take the consequences of financing decisions for employees and suppliers into account. Relevant thereby is the dependence of employees and suppliers on the company. Because of the large economic influence firms can have, not all suppliers and employees will be able to terminate their relationship with a company. In some regions for example there is only one large employer. Employees of those companies then are unable to find another job. This increases the consequences of a possible bankruptcy and thus the costs of financial distress. The costs of
31 Net earnings to the shareholders could be negative, zero or positive; in case of the latter, they can be either
reinvested in the company or paid out as dividends.
32 I.e. companies could be enforced by the debt holders to meet their principal and interest payment obligations,
financial distress will therefore increase with the dependence of stakeholders on the company. Therefore, variables that affect the dependability of stakeholders on firms should be taken into account when analyzed the determinants of capital structure. For completeness, it should be noted that financial distress may make employees work harder and suppliers accept more favorable terms in order to avoid bankruptcy. The gains of this however are likely to be very small.
Customers are an important factor in the costs of financial distress, as in the end they are the ones who generate the sales. Though they are not financially dependent on firms (it is the other way round), they behave to a certain extend based on characteristics of the firms which provide them with products and services. (Potential) customers consider i.a. the price, quality, maintenance and warranties of their (planned) purchases. When these factors are being affected by changes in the capital structure, the behavior of (potential) customers will be affected as well. When financial distress increases, prices may be increased to increase sales, quality may get worse because of a cut down on expenses on material or because of diminished dedication of employees (as described above), and maintenance and warranties may be cancelled because of bankruptcy. Therefore, customers might not buy products of financially distressed firms, or only at more favorable terms like lower prices33; this of course
is fatal for the continuity of the company. From Graham and Harvey (2001, p. 210) it can be learned that customer and supplier comfort is definitely a factor taken into consideration by CFOs when determining the composition of capital structures. Variables that affect the uncertainty under customers about price, quality, maintenance and warranties are therefore likely to affect financing decisions as well.
Competitors may affect the determination of capital structures because, as Welch (2002) suggests, firms may adjust their debt-ratio towards the industry average. One reason for this is that firms use the industry average debt-ratio as a reference. Another possibility is that firms independently assess their own optimal debt-ratio, but that, because firms in the same industry make similar investments, their optimal debt-ratios are naturally similar as well. According to Graham and Harvey (2001, p. 210) many CFOs base their debt policy at least partly on comparable firm debt levels.
The role of management within firms is much debated in economic literature. It is part of the trade-off theory, as capital structures are assumed to affect management behavior, but it is subject of the principal-agent theory as well. This theory covers the nature and consequences of the relationship between two parties who have concluded a contract under which one (the agent) is to perform on behalf of the other (the principal) some service which involves delegating some decision making authority to the agent.34 Relevant here is that the objectives
of managers and investors may not always be compatible. Because each maximizes its own utility, the actions of management may limit the maximization of shareholder value or harm the interests of debt holders. Information asymmetries make it more costly for investors to monitor management behavior. Capital structure can then be used to direct management behavior into a more favorable direction from an investor’s point of view (the incentive effects).
Management weights the risks and estimated returns of investments. Assuming their only objective is to maximize shareholder value, problems arise between debt-holders and management. When investments are largely financed with debt, this may stimulate managers to invest in projects with high returns if successful, even if the probability of success is very small. If successful, shareholders capture most of the gains, while debt-holders bear most of
33 Babenko (2003): p. 8.
the costs if the investment turns out unsuccessful.35 To reduce this problem, debt-holders limit the total level of debt in companies so that the losses in case of unsuccessful investments will be limited, and they require managers to take part in the company so that these will suffer personal financial losses if investments turn out unsuccessful. In addition, because creditors have a certain degree of influence on a firm, managers themselves can be reluctant to have high levels of debt in the firm. Debt-holders are often able to oppose certain limitations on financing or even operations that restrict management in its opportunities.36 Examples of this are solvability and liquidity rates requirements, discouraging certain (risky) investment projects and a seat in the board of directors.
Leaving the assumption of purely maximizing shareholder value for that of managers maximizing their own utility creates another agency problem: that between management and shareholders. Of course, managers have the incentive to make decisions that will increase expected future cash flows as they want to guarantee the continuity of the firm and improve their own reputation.37 However, in order to increase their own utility, they also tend to enrich
themselves and to generate excessive perquisites at the expense of shareholders.38 This affects capital structure in two ways.39 First, shareholders will make management to take part in the company to bring their interests in line with each other. Second, shareholders will increase the level of debt because of the monitoring by debt-holders and the restrictive effects of financial distress (managers will limit their perquisites when the financial space for it becomes smaller).
2.3.4. Concluding remarks
From the preceding, it becomes clear that financing decisions reach further than just the composition of capital structure. Because of their consequences for all the stakeholders of a company, financing decisions directly affect the operating performance, and thus the market value, of firms. Note that this relationship was entirely overlooked by MM, who only consider the effects of debt on the discounting rate of (operating) cash flows, not on the level of the operating cash flows themselves, an important omission in their theorem. Jensen and Meckling state: “It is now recognized that the existence of positive costs associated with bankruptcy and the presence of tax subsidies on corporate interest payments will invalidate this [MM’s] irrelevance theorem precisely because the probability distribution of future [operating] cash flows changes as the probability of the incurrence of the bankruptcy costs changes, i.e., as the ratio of debt to equity rises.”40
The benefits and costs of debt should be weighted to determine a target debt-ratio that will maximize the market value of the firm (given the firm’s assets and future investment plans). This is the point where the marginal costs of financial distress equal the marginal revenues of the tax-shield. Overall, “firms select capital structures depending on attributes that determine the various costs and benefits associated with debt and equity financing”41 to maximize shareholder value. The process of assessing this target debt-ratio and striving towards reaching it through changes in the capital structure of a firm is not a unique, static event, but a continuous, dynamic process which is subject to change all the time as the firm is liable to both external shocks (e.g. in the share price) and internal processes (e.g. the development of growth opportunities).
35 Jensen and Meckling (1976): p. 334. 36 Grinblatt and Titman (2002): p. 581-583. 37 See e.g. Myers (1977): p. 148.
38 Jensen and Meckling (1976): p. 312-313.
39 Jensen and Meckling (1976): pp. 334-337, Fama and Jensen (1983): pp. 312-315. 40 Jensen and Meckling (1976): p. 333.
Miller (1977) claims that the trade-off theory “goes wrong” when it assumes there is a one-on-one trade-off between bankruptcy costs and the gains of the creation of tax-shields. According to him, the costs of bankruptcy “seem disproportionately small relative to the tax savings they are supposedly balancing.”42 This seems to be somewhat in line with the traditional view that shareholders do not demand a higher yield on their shares when a company increases its debt level as long as the debt-ratio stays below the critical level (see figure 2 in Appendix 2). However, the costs of debt consist of direct financing expenses as well as bankruptcy costs and financial distress which, together, can form a considerable large amount. In addition, Miller pays attention to problems at the tax-shield side of the trade-off theory. First, he claims that capital structures have shown little change over time, even when tax-rates have changed enormously. Second, he states that increases in the value of shares due to increased tax-shields are cancelled by larger personal income tax obligations. This would not count for large institutional investors however who form a large part of the shareholders. Third, Miller states that the same corporate tax rate is being equally applied to all companies in the same country, so that it can not explain differences in capital structures of firms in the same country at any given time. Firms however do differ in their marginal tax rate, e.g. as a result from operations abroad; besides, the corporate tax rate could be explanatory for differences between firms in different countries and for different periods in time.43
2.4. The pecking order theory
A different approach is being taken by the pecking order theory.44 The central issue here is the way companies finance new investments most favorable. The pecking order theory states that, in general, companies prefer to finance new investments with internal funds (retained earnings and investment reserves) rather than external capital (shareholder’s equity and debt). If the internal funds are insufficient (which is mostly the case), firms prefer debt to equity issues. Hybrid financing structures like convertible bonds are preferred to equity but not to straight debt. The difference with the trade-off theory is that the trade-off theory considers all equity equal (as opposed to debt), whereas the pecking order theory distinguishes internal and external equity, of which internal is the most and external the least preferred by companies for financing. Where the trade-off theory leads to the assumption of a capital structure where the benefits and costs of debt are perfectly balanced, according to the pecking order theory target debt-ratios are irrelevant; equity is not homogeneous and the amount of debt depends on the availability of internal funds. Overall, variables that affect the level of internal funds are then likely to affect capital structure as well.
There are multiple reasons for the pecking order of financing decisions, some being cost related and others related to information asymmetries, which are to some extent connected to the principal-agent theory:
First, historical data show that, in general, bondholders demand lower returns than shareholders, mainly because of lower risk. Variables that reduce this risk will probably lower the demanded returns as well.
Second, transaction costs can form a considerable large expense, though their exact level varies per transaction.45 Retained earnings and investment reserves are readily available to the firm and no services by third parties are required for using them. Therefore, no legal fees or administrative or underwriting costs have to be paid. Debt
42 Miller (1977): p. 262.
43 See e.g. Graham (1996a,b) and Marsh (1982): p. 132.
44 See e.g. Myers (1984), Myers and Majluf (1984) and Grinblatt and Titman (2002): pp. 612-614.
45 See Gilson (1997) for an extensive study of the relationship between transaction costs and debt, especially for
or equity issues however involve extensive service by expensive lawyers, consultants and bankers. In general, the costs of issuing debt are lower than for equity: on average 2.24% for straight bonds compared to 3.79% for convertible bonds, 7.11% for seasoned equity issues and 11% for initial public offerings.46 As initial public offerings are more expensive than seasoned offerings, private firms are more inclined to issue debt than public firms.47 In addition, large issues are relatively less expensive than small issues because of economies of scale (fixed costs are being spread over a larger amount).
Third, companies want to keep some flexibility in their capital structure, also referred to as “financial slack”.48 Companies keep cash and marketable securities at hand and don’t use the maximum possible amount of debt because they want a buffer to remain to be able to finance new investments quickly when they occur. Graham and Harvey (2001, p. 210) report this to be the single most important factor on which CFOs base debt issuing decisions. Therefore, if the debt-ratio has reached critical limits (as e.g. set by banks) a firm could issue new equity to counterbalance, even if there is no direct need for the funds.49 This is connected of course with the space within the cash-flows for increased fixed payments which go with higher levels of debt. Some argue that because of the benefits the availability of cash produces (e.g. the rapidity with which new investments can be made), it should be valued at more than its book value. However, cash also has drawbacks, e.g. that it increases managers’ ability to enrich themselves or to make value-destroying investments. Therefore, cash is assumed to have the same value inside and outside companies.
Fourth, outside investors can have considerable influence on companies. Equity investors have a legally regulated right to appoint the supervisory board and have to approve management decisions and the annual report. Bondholders can affect both financial and operational decisions because of the terms of debt. By using internal capital, this kind of exertion by third parties is being prevented altogether, and managers avoid being “subject to the discipline of the capital market.”50
Fifth, there are the consequences of information asymmetries, the basis of the both the principal-agent and the pecking order theory. Managers are assumed to know more about the future performance of the firm than outside investors. Because of this knowledge, they consider shares to be underpriced, overpriced or rightly priced. In general, they are reluctant to issue underpriced shares, but are eager to issue overpriced shares.
The possibility of underpriced shares was already mentioned by MM. They stated that investors and analysts “place very heavy weight on current and recent past earnings in forming expectations as to future returns.”51 When a firm then creates an investment opportunity with higher than expected returns the share price “may fail to capitalize the new venture.”52 Therefore, the share price would be undervalued and managers do not want to issue new shares at this low price to finance the investment project. Managers would therefore have to communicate the higher expected returns of the new investment to the shareholders. If the information asymmetry continues to exist, a
46 Total direct costs as a percentage of total proceeds, average for 1990-1994 period. Lee et al. (1996): p. 62. 47 See also Grinblatt and Titman (2002): p. 15.
48 Brealey, Myers and Marcus (2001): p. 446. 49 Taggart (1977): p. 1484.
50 Myers (1984): p. 582. 51 MM (1958): pp. 292. 52 Idem.
better strategy would be to issue debt initially, and later redeem this with equity issued when the true returns of the project become known. Another (more elegant and probably less expensive) solution would be to issue convertible debt. A third option would be to place the investment project in a separate entity.53
The opposite is the case when managers consider the shares of the firm to be overpriced. The value added by the new project is then of lesser importance; the management simply thinks the shares should be worth less than they actually are, given the firm’s prospects. An equity issue would then raise more funds than would be the case if investors were aware of the real worth of the shares. Investors are aware of this signaling function54 and therefore the share price is likely to drop after equity issue announcements, something managers want to avoid. When the share price is relatively high, managers are more likely to issue new equity. Therefore, after an increase in share price, managers are more likely to issue equity then debt (when financing new investments).55 This contradicts the trade-off theory, which predicts
companies to issue debt in such a case, as, when share prices have increased, the debt-ratio has fallen to below the target level.
Of course, it depends on the individual circumstances of companies how these reasons affect the decision about the issuing of debt and equity. It should also be noted, like Brealey et al. (2001) state, that “the pecking order theory does not deny that taxes and financial distress can be important factors in the choice of capital structure [like the trade-off theory states]. However, the theory states that these factors are less important than manager’s preference for internal over external funds and for debt financing over new issues of common stock.”56
2.5. Criticism of Welch and his “inertia” hypothesis
Ivo Welch (2002) has argued that neither the trade-off nor the pecking order theory adequately describes the business processes that, in general, lead to the composition of capital structures. In his view, managers fail to react appropriately to changes in equity values at all, so that capital structures are not being adjusted to shocks in the share price of a company (at least not in the short run) as generally predicted in the literature.57 According to Welch, there is no rationale for current debt-ratios, except for the fact that the level of debt has not (yet) been adjusted to the current equity value of the firm (“inertia”). This is fundamentally at odds with both the trade-off and the pecking order theory, though for different reasons. According to the trade-off theory, the current capital structure is the result of a careful consideration of all the currently relevant variables. According to the pecking order theory, current capital structures can be explained by past investment financing decisions and past profitability. (In addition, Welch explicitly assumes firms should return to their original debt-ratio after a change in their share price. The pecking order theory however does not attach much importance to past debt-ratios). In Welch’s view though, it is past (exogenously determined) stock returns that determines equity value and thus debt-equity ratios. In his own words: “The variables prominently featured in some other studies as explanators of capital structures (…) seem to function only through their mechanistic correlation with (past) returns and equity capitalization. Once we include our mechanistic inertia debt ratio, these variables lose their power.”58
53 Idem.
54 Myers (1984): pp. 583-585. 55 Myers (1984): p. 586.
56 Brealey, Myers and Marcus (2001): p. 446. 57 See e.g. Taggart (1977) p. 1469 and 1475. 58 Welch (2002): p. 5.
One of Welch’s explanations for the inertia is that it may be expensive for firms to issue new equity when the debt-ratio has increased due to decreases in the share price. However, this would only apply for the short run and for small shocks in the share price. His other explanations are that managers are reluctant to repurchase equity to issue more debt in case of an increase in share price (either because the shares are overvalued or because they do not see the positive features of debt) or, in case of a decrease in share prices, managers don’t want to issue what they think are underpriced shares.59
Welch finds that debt-ratios are better explained by the “inert debt-ratio” (debt-ratio of the previous year adjusted for the difference in the external stock return) rather than the previous debt-ratio (which he assumes to be optimal). In his view, this makes other variables superfluous. However, he ignores the possibility that the external stock returns are being caused by announcements about changes in these variables, which would make them the underlying fundamental causes of current debt-ratios after all. It could perfectly be that a previous year’s debt–ratio wasn’t optimal in terms of the trade-off theory, or that announcements have been made about future growth opportunities that led to changes in the share price, or that in the meantime adjustments to the debt-ratio have been made towards a more favorable capital structure. This would not be reflected in Welch’s results and leaves his reasoning plausible. It would however contradict his conclusions as it would after all be the variables described by Titman and Wessels (1988) and others that best explain current debt-ratios. Nevertheless, previous year’s debt-ratio could give some relevant information about current capital structures and should not be ignored altogether.
According to Welch’s findings, not even short-term debt is being adjusted to external share price shocks, which would be easiest to do60. It would take at least five years for a firm to
return to its optimal debt-ratio. His conclusions are hard to accept when capital markets are believed to behave reasonable efficient, and investors and analysts have about the same information about firms as managers. Managers, who ignore optimal debt-ratios and act with great inertia or not at all to changes in share prices, surely would not stay long with a firm when investors believe there to be at least some truth in the trade-off theory.
There is, however, in my opinion, also a sort of middle way and that is when there are reasons for the observed inertia which simultaneously can be considered to be value maximizing for current shareholders. One such, in my view justifiable, reason to postpone adjustments to the current debt-ratio after a change in the share price is that long-term interest rates are being expected to decrease in the near future. In that case, it is in the best interest of the shareholders to wait with the issue of new long-term debt till interest rates have decreased.61 Another reason could be that issues of debt and equity are timed such that they coincide with cash outflows for investments in order to “reduce the transaction costs of switching funds in and out of liquid assets until they are needed. (…) The pattern of expenditures could then influence the timing of security issues, reflecting firms’ attempts to coordinate their inflows and outflows.”62
Welch’s concept is somewhat related to the comment of Miller (1977) on “neutral mutations”.63 According to Miller, “actual decision procedures are inevitably heuristic,
judgmental, imitative and groping even where (…) they wear the superficial trappings of
59 Welch (2002): p. 4. These and other arguments are being more extensively discussed in §2.4.
60 Taggart (1977) also finds “relatively slow adjustment speeds to the long-term debt and permanent capital
targets” (p. 1475). However, he does find short-term debt to be adjusted to external share price shocks.
61 Based on Taggart (1977): p. 1476. 62 Taggart (1977): p. 1481).
hard-nosed maximization.”64 Managers simply use rules-of-thumb and their intuition rather than rational financing objectives to decide about the financing of a firm. As the capital structure of a firm has no influence on the value of that firm, as Miller advocates, this heuristic behavior of managers is rather harmless to the shareholders of the firm. Therefore, where harmful heuristics will eventually “die out, neutral mutations that serve no function, but do no harm can persist indefinitely.”65 Like Welch’s theory, this “neutral mutations” theory is hard to believe when the costs of various capital structures differ substantially and the market puts weight to differences in capital structures. That this is indeed the case can be seen by the fact that share prices change when announcements are being made about changes in the capital structure of a company.66 In addition, both Welch’s and Miller’s conclusions are contrary to much empirical evidence presented by numerous authors, e.g. Rajan and Zingales (1995): “All this suggests that the levels of leverage that we see in different countries do not rise randomly, but are a consequence of conscious financing choices made by firms.”67
64 Miller (1977): p. 272.
65 Miller (1977): p. 273. 66 Myers (1984): p. 576.
