How disagreement affects corporate decisions : an analysis of security issues

(1)

How Disagreement Affects Corporate Decisions:

An Analysis of Security Issues

Master Thesis

Amsterdam, 1 July 2016

Abstract

Using a sample of nonconvertible debt issues and seasoned equity offerings for US listed firms between 1993 and 2014, this paper tries to empirically analyze the effect of manager-investor disagreement on security issuance decisions. It helps explain the regular occurrence of equity issues by firms by finding evidence that firms are more likely to issue equity when manager-investor disagreement is low. Furthermore, this paper tests an extension of the disagreement theory by Boot & Thakor (2011) and finds that firms with high valued existing assets will always issue debt, regardless of the level of disagreement. It serves to show the conditionality of security issuance theories. Additional robustness tests present several characteristics of firms and industries where this theory does not hold. Specifically, this paper finds that the disagreement theory especially holds for smaller, younger or low levered firms in the manufacturing or service industry.

Keywords: Manager-investor disagreement, Issuance behaviour, Security issue theories

Student Taco Trimbach

UvA ID 10001142

Study MSc Business Economics

Track Finance

(2)

2

Statement of Originality

This document is written by student Taco Trimbach who declares to take full responsibility for the content of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business of the University of Amsterdam is responsible solely for the supervision of completion of the work, not for the contents.

Amsterdam, July 2016

Signature

Acknowledgements

I would like to express my deepest gratitude to my thesis supervisor professor Arnoud Boot for his guidance and feedback during my thesis writing process. His insightful comments have motived me whenever I got stuck and helped me to continuously improve my thesis.

Furthermore, I would like to thank my fellow classmates and friends that have shared and endured this process with me on the third floor of the Roeterseiland. I want to give a special mentioning to Pieter and Else for making my master finance that much more workable and fun. Lastly, but certainly not least, I would like to thank my parents who always supported me and Liset for being awesome.

(3)

3

1. Introduction

How do firms make decisions when management and shareholders have fundamental different views of the future and no certainty whose view is correct? In today’s ever changing economic environment this is a relevant question. Turmoil in financial markets has added to the uncertainty and increased divergence of opinion between market participants. For public firms this poses an extra challenge. They have to justify their decisions to their shareholders who might not share the same view on what is important in the future. Firms might find that operating in such an environment is extra challenging as they are judged by the value of their stock. Large conflicts of opinion between managers and shareholder can impair decision making and damage the continuity of a firm. It shows how firms are governed by their shareholders and how shareholder sentiment can affect corporate decisions. Yet, firms desire a certain level autonomy to be able to do what they think is right. Many corporate decisions are therefore a trade-off between the benefits of autonomy and the effect of a decision on the stock price. This trade-off has to be taken into account when firms decide whether to go public or not, what investments to do and, in the case of this paper, what kind of securities to issue.

One of the key interests of researchers in finance is to explain security issuance behaviour of firms. Many theories have been developed to address this issue although few are able to explain the many seen equity issues in today’s markets. These issues seem to coincide with rising stock prices and are in stark contract with the predictions of two main theories of security issuance: the trade-off theory and the pecking order theory. The trade-off theory (Jensen & Meckling, 1976; DeAngelo & Masulis, 1980) states that the optimal debt level should be a trade-off between the cost of additional debt, such a bankruptcy costs, and the tax benefit of this additional debt. Consequently, debt will always be issued as long as the optimal debt level is not achieved and rising stock prices should induce debt issues as this decreases the effective leverage ratio. The pecking order theory (Myers & Maljuf, 1984) states that asymmetric information problems would cause managers to abstain from issuing equity as shareholders interpreted this as a negative signal of firm value. This could adversely affect the stock price. Thus, debt will be issued until the maximum debt level is reached and equity will only be issued under extreme circumstances. As both theories fail to explain equity issuance motives the need became clear for a theory that could provide a possible explanation. The market-timing theory (Baker & Wurgler, 2002) is able to explain firm’s frequent equity issues when stock prices are high. They argue that this is a result from managers trying to take advantage of the perceived overvaluation of their firm to obtain cheap equity. Thus, managers try to “time the market” when issuing equity.

While the market-timing theory is able to explain security issuance behaviour, this paper seeks to investigate an alternative theory of security issuance. This theory is centred on disagreement between managers and investors and was first suggested by Dittmar & Thakor (2007). Its goal is to explain how this disagreement affects the decision to issue either debt or equity. Dittmar & Thakor

(5)

5

(2007) argue that disagreement has a negative relationship with the probability of an equity issue. In other words, firms with higher levels of agreement are more likely to issue equity. They argue that a manager values autonomy in project choice and equity holders are less likely to disagree with management decisions when disagreement is low. Low disagreement also causes the stock price to rise, but even when controlling for the stock price they find that disagreement is an important determinant of issuance decisions. Throughout this paper disagreement between managers and investor will be referred to simply as disagreement unless mentioned otherwise. Boot & Thakor (2011) elaborate on this theory. They argue that debt can still be issued when disagreement is low if the value of the assets in place of a firm is sufficiently high. If this is the case a manager can achieve greater project autonomy by issuing debt as payoffs to debtholders do not depend on the outcome of the investment opportunity. The model by Boot & Thakor (2011) shows that this theory is also conditional upon certain situations. Therefore, the aim of this paper is to study how disagreement between managers and investors affects security issuance decisions under different conditions. The main research question can be formulated as follows: What is the effect of disagreement between managers and investors on firms security issuance decisions?

This paper hopes to add to the existing theories of security issuance by providing an in depth analysis of the effect of disagreement of security issuance decisions. By doing this it hopes to shed light on how shareholders and other investors affect corporate decisions and activities. In a world where views of the future state of the world can lie far apart from each other, understanding this is essential in understanding how firms behave and what effect it has on financial markets. To test the disagreement theory a sample of non-financial firms that issued nonconvertible debt or seasoned equity between 1993 and 2014 is used. Two proxies for disagreement will be employed: the actual minus forecasted earnings per share and the standard deviation of analysts’ forecasts. To exclude market-timing motives in security issues decisions several stock price variables will be controlled for. The final sample consists of 3,960 unique firms with a total of 9,819 total security issues of which 3,416 are debt issues and 6,403 are equity issues.

Using logit regression analysis, this paper finds a significant relationship between disagreement and security issuance decisions. Specifically, the results provide evidence that firms who have less disagreement between managers and investors are more likely to issue equity. It shows that managers value autonomy and issue equity to achieve this when disagreement is low. The results are robust to using both disagreement parameters and multiple market-timing theory controls. This paper also finds that stronger evidence emerges for the earlier years of the sample and that the relationship disappears in later years, possibly due to the recent financial crisis. Additional robustness analysis shows that the disagreement theory is more likely to hold for younger, smaller, and low levered firms without rated debt. Considering the industries the firms in the sample operate in, this paper finds that equity issues dominate the manufacturing and service industry while debt issues dominate the mining, construction, transportation, and retail industry. As the results imply, one might

(6)

6

expect the disagreement is a more important determinant of issuance decisions in the industries where equity issues dominate. Lastly, the extension of the disagreement theory by Boot & Thakor (2011) is confirmed. This paper finds evidence, although limited in some cases, that firms with higher valued assets in place are more likely to issue debt regardless of the level of disagreement. It shows that the value of the assets in place has a significant influence on issuance decisions. As with the previous findings, the results appear much stronger when an adjusted sample period is used. Additional descriptive statistics show that firms with higher valued assets in place are more likely to be in the mining, constructing and retail industry. As such, the theory by Boot & Thakor (2011) is more likely to hold in these industries.

On a more general note, these findings suggest that a universal theory of security issuance has yet to be discovered. Although manager-investor disagreement does appear to have significant explanatory power over issuance decisions many more factors play a role in this decision. As with the other theories of security issuance there are situations or circumstances where this theory does not hold. It illustrates the fragility of security issuance theories by emphasizing the conditionality of these theories. One has to take this into account when making assumptions based on these theories. The remainder of this paper is organized as follows. Section 2 will discuss related literature and elaborate on the theories of security issuance. The methodology and formulation of testable hypotheses will be discussed in section 3. Section 4 clarifies the data sources and presents the descriptive statistics for the sample. Results will be presented in section 5. Finally, the conclusion and discussion will be provided in section 6.

2. Literature Review

This section will discuss the main theories developed to describe the security issuance behaviour of firms. The main empirical papers and their findings will be presented and the economic relevance of each of them will be discussed. Firstly, both the trade-off theory and the pecking order theory will be discussed as both of these theories predict that equity will only be issued as a last resort. Secondly, the market-timing theory will be discussed, which provides us with a possible explanation for the regular occurrence of equity issues. Lastly, the newly developed theory of security issuance will be discussed. This theory incorporates divergence of opinions between managers and investors and the effect of this divergence on security issuance decisions. An extension of this theory, which adds a managerial autonomy component to the equation, is also presented in this section. The goal of this literature review is to provide an extensive outlay of existing research regarding security issuance theories and to emphasize the importance and contribution of this paper in the academic field of finance.

(7)

7 A. Trade-off Theory & Pecking Order Theory

The static trade-off theory is essentially comprised out of the agency costs theory of Jensen & Meckling (1976) and the debt tax shield theory of DeAngelo & Masulis (1980). The theory states that a firm has an optimal debt level and that it should strive to maintain its optimal capital structure throughout the corporate lifecycle. The issuance of new securities is thus a trade-off between the agency- and bankruptcy costs and the tax-shield benefit of additional debt. According to the trade-off theory a firm should issue debt when its stock prices are rising. This results from the fact that rising stock prices decrease the effective leverage ratio of a firm. In order to maintain its optimal capital structure a firm should therefore issue debt. The theory also predicts that debt will be bought back when the optimal debt level is exceeded.

In response to the high leverage ratios some firms exhibit a new theory of security issuance was developed by Myers & Majluf (1984). This “pecking order” theory revolves around the problem asymmetric information, which arises from different levels of information between the insiders (i.e. management) and outsiders (i.e. shareholders) of a firm. The pecking order theory predicts a certain order of financing preferences that influence a firm’s issuance decisions. It states that a firm prefers to finance a new investment using internal funds first. Only when internal funds are depleted and additional funds are still required will a firm move on to external financing. The theory then predicts that a firm will first issue debt to satisfy its demand for external financing and that it will continue to issue debt until its debt capacity is reached. Equity will only be issued once a firm has reached its debt capacity. The rationale behind this is that insiders of a firm are usually better informed about the quality of an investment opportunity than outsiders. This level of information asymmetry generates adverse selection costs in the sense that the price at which a firm is able to sell its securities will be less than the value given to the security by a firm’s insiders. Investors interpret the issue as a signal that management believes that the stock is currently overvalued. Thus, an equity issue by a firm before it has reached its debt capacity sends a negative signal to investors which is consequently reflected in the stock price. The theory also predicts that firms might even pass up on positive net present value (NPV) projects when the debt capacity is reached and internal funds are exhausted to avoid the negative effect of an equity issue on the stock price. The main difference between the pecking order theory and the trade-off theory is that the pecking-order theory reflects a firm’s cumulative need for external financing whereas the trade-off theory mainly explains more normal capital structures (Myers, 1984).

Although both theories provide rational explanations for security issuance behaviour they are unable to explain the fact that firms tend to issue equity when stock prices are high. Additionally, several papers, among others Fama & French (2005), Leary & Roberts (2010), Brounen et al. (2006) and Graham & Harvey (2001), provide compelling evidence against these theories. Using a more direct approach than the usual cross-section regressions and a sample of firms in the period 1973-2002, Fama & French (2005) find that the pecking order theory is violated more than 50% of the

(8)

8

times in their sample. They employ a broad range of equity issue types including not only SEOs but also private placements, convertible debt, warrants and more. They find that on average 67% of the firms in the sample issued equity in the period 1973-1982 and that this number increases to 86% for the period 1993-2002. More strikingly, they find that equity issuers are typically not under pressure to issue equity, as is predicted by the pecking order theory. Firms that issue equity generally have moderate leverage levels and financing surpluses. Fama & French (2005) conclude by stating equity financing is not a last resort, that asymmetric information is not the only determinant of capital structures and that the pecking order theory cannot provide a general and universal explanation for the issuance behaviour observed by firms. Leary & Roberts (2010) use a novel empirical model to find similar results when applying a strict interpretation of the pecking order theory. In this case they find that the predictive power of the pecking order theory is only 50% in their sample. Only when they allow certain components to vary according to factors described by alternative theories, do they find that this predictive power increases. In addition, they ascribe the limited evidence they find in favour of the pecking order theory more to incentive conflicts and less to information asymmetry. Graham & Harvey (2001) and Brounen et al. (2006) provide some striking survey evidence regarding firm’s capital structure decisions. They questioned 392 US CFOs and 313 European CFOs, respectively. Both papers show that security issuance decisions are only moderately driven by trade-off- and pecking order behaviour and that these decisions are dominated by the degree of financial flexibility and credit ratings. Furthermore, they find limited evidence that asymmetric information influences security issuance motives.

Boot & Thakor (1993) directly contradict the predictions of both theories by proposing an alternative theory. This theory states that, in the presence of asymmetric information, issuing an “informationally insensitive” security, such as debt, does not always dominate issuing an “informationally sensitive” security, such as equity. They developed a model whereby an issuing firm benefits from separating its assets cash flows into both an informationally insensitive and an informational sensitive security. The assumption is that the assets cash flows are ex ante unknown to outside investors but can be discovered by some at a cost. By offering securities of different sensitivity they allow informed investors to transfer their capital to more informational sensitive securities, where they can earn a higher return. This way, higher-valued firms can stimulate informed trading and encourage information production which, in turn, moves the price of the security closer to its fundamental value. Thus, Boot & Thakor (1993) state that higher valued firms will always choose to issue equity, in contrast to the trade-off- and the pecking order theory. In response to this theory, Fulghieri & Lukin (2001) discuss an alternative. They argue that the pecking order theory might still apply in some situations depending on several factors. One factor in this case is the cost of becoming informed. In line with Boot & Thakor (1993), they argue that if the cost of becoming informed is sufficiently low, a firm can benefit from issuing equity. However, when the cost of becoming informed increases to such a point where they outweigh investors’ potential benefit from informed

(9)

9

trading, a firm rather chooses to issue debt. Thus, the security issuance decision does not only depend on the level of information asymmetry but also on the cost and precision of information production (Fulghieri & Lukin, 2001).

The two theories described above provide a good example of the inability of security issuance theories to fully predict the corporate decisions that firms make. These decisions depend on many more factors than just those incorporated in the theories. Myers (2001) argues that these theories are not meant to be universally accepted and generally applicable and that they are only conditionally useful. One has to bear this in mind while discussing the empirical relevance of each theory.

B. The Market-Timing Theory

To address the failing predictability of both the trade-off theory and the pecking order in explaining the stylized fact that firms tend to issue equity when stock prices are high, Baker & Wurgler (2002) developed a new theory of security issuance; the market-timing theory. This theory provides a clear rationale for the many equity issues by firms. It involves the behaviour of firms to “time the market”, that is, a firm’s tendency to issue equity when they perceive their stock to be overvalued measured by book value and past market values. Overvaluation occurs when irrational investors fail to update their valuations to incorporate the information contained in the equity issue or other news-carrying corporate actions. By issuing equity firms take advantage of the irrational high stock price and are able to raise a larger amount of funds. The theory also implies that firms tend to buy back equity when their market values are perceived to be low and that market-timing has large and long-lasting effects on a firm’s capital structure. Using a sample of listed firms Baker & Wurgler (2002) find strong empirical evidence in favour of the market-timing hypothesis. The explanation they offer for their results states: “A firm’s capital structure is the cumulative outcome of attempts to time the equity market” (Baker & Wurgler, 2002, p.3).

More compelling evidence is provided by Graham & Harvey (2001) and Brounen et al. (2006). Both surveys find that the degree to which a stock is over- or undervalued is considered of high importance among a CFOs decision to issue equity. Furthermore, the surveys show that recent stock price rises have a positive effect on the decision to issue equity. Other papers offer empirical evidence in favour of the market-timing theory. Dong et al. (2012a) find that the market-timing theory is stronger among firms with potential growth opportunities and high share turnover. Using a forward-looking approach of misvaluation they test how the sensitivity of security issuance to misvaluation varies across several subsamples. Elliot et al. (2008) find that even when controlling for growth options and adverse selection effects, the choice of issuing equity is still significantly affected by the level of overvaluation. Jenter (2005) reports similar results by concluding that managers’ views of the fundamental value of their firm differs consequently from market valuations and that this alleged mispricing is a fundamental driver of security issuance decisions. Using a sample of Canadian firms,

(10)

10

Dong et al. (2012b) test the market-timing theory against the pecking order theory. They find that firms will only issue equity when they are overvalued if they are not financially constrained as financially constrained firms might not have to flexibility to issue and repurchase equity according to market-timing behaviour. Correspondingly, they find that when firms are not overvalued they prefer debt financing over equity financing. This is in line with evidence from Dong et al. (2012a), who find that the probability of equity issuance is increasing in the amount of overvaluation but that this is only among overvalued stock. Lastly, Dong et al. (2012b) show that among overvalued firms the post issue long-term return is significantly lower. Interestingly, this issue is also addressed by Schultz (2003) and Loughran & Ritter (1995). Both papers study the observed long-run underperformance of initial price offerings (IPOs) and seasoned equity offerings (SEOs) and argue that this is due to mispricing and managements’ exploitation of this window of opportunity. Schultz (2003) addresses a phenomenon he refers to as pseudo market-timing, which states that the probability of an equity issuance is increasing in the amount a firm can receive for its stock. Thus, equity issues that occur at peak prices result in lower returns. On a contrary note, DeAngelo et al. (2010) find that it is a firm’s need for cash that is the main driver behind equity issues and that market-timing opportunities and lifecycle stage are only secondary drivers. In their paper, they evaluate the explanatory power of the market-timing theory in comparison with a simple lifecycle theory that states that more mature firms (low market-to-book) will fund investment through internal funds and debt while younger firms (high market-to-book) tend to issue equity to fund investments. Using a sample of firms in the period 1973-2001 they find that 62.6% of firms in their sample would have run out of cash without the proceeds of the equity issue and that 81.1% of firms would have cash balances below average. Lastly, in his well-known empirical paper Welch (2004) finds that neither tax considerations, bankruptcy costs nor market-timing proxies can explain issuance behaviour when stock returns are controlled for. He concludes that in the long-run the stock price is the best proxy to explain issuance decisions and that debt-equity ratio changes made as a result of changes in the stock price have long lasting effects on the capital structure.

C. Manager-Investor Disagreement Theory

Although the market-timing theory of Baker & Wurgler (2002) provides a possible answer to the many observed equity issues, this paper seeks to address an alternative theory about security issuance. This theory was first suggested in a paper by Dittmar & Thakor (2007). It is based on the premise that firms are more likely to issue equity when disagreement between managers and investors is low. They argue that when disagreement is low, stock prices will be higher and this, in turn, will encourage managers to issue equity. This means that the choice of issuance revolves around the presumption how this will affect investment choice and post-investment stock price. A manager will choose to finance a new investment if he beliefs that this will benefit the stock price. Since asymmetric

(11)

11

information problems play a significant role, whether this investment benefits the stock depends on whether investors belief the investment is a good project. An investor that is more likely to agree with a manager is therefore more likely to share a manager’s view about the profitability of the investment opportunity. This clearly highlights the importance of disagreement on the choice of security issuance. The choice of issuing equity, in this case, is affected by the autonomy it gives to the manager. Thus, by issuing equity when disagreement is low a manager ensures autonomy in project choice. Similarly, debt can be more restrictive. If bondholders do not share the same views as the manager about the profitability of an investment opportunity, they have the option to include restrictive covenants in their debt contracts. These protect bondholders but limit the freedom the manager has in choosing their desired project. A manager might therefore end up investing in a project that might not maximize shareholder interests. Thus, Dittmar & Thakor (2007) predict that equity will be issued when disagreement is low and stock prices are high, and that debt will be issued when disagreement is high and stock prices are low. Furthermore, the model predicts that when there are no investment opportunities available a firm will not issue equity but it might issue debt.

Dittmar & Thakor (2007) test their predictions using a sample of firms over the period 1993-2002 that issued seasoned equity or nonconvertible debt. The sample eventually consists of 4,496 equity issues and 3,321 debt issues. They employ several proxies for manager-investor disagreement, such as the difference between a firm’s actual earnings per share (EPS) and the mean of the analysts forecast of the EPS. Taking into account the effect of stock prices and the level of asymmetry on the security issuance decision they add several control variables to their regression. Even when controlling for these factors they find strong empirical evidence in favour of their model. They also find that firms that issue equity experience a strong increase in their capital expenditures which is they do not find for firms that issue debt.

Boot & Thakor (2011) present a model that elaborates on the theory and evidence presented by Dittmar & Thakor (2007). They extent the notion that equity will be issued when disagreement is low by emphasizing “managerial autonomy” in security issuance decisions. Their theory allows for debt to be issued when disagreement is low if this ensures manager autonomy in project choice. The model presents a situation where debt will be issued, regardless of disagreement, if the value of the assets in place is high. When the value of assets in place is sufficiently high, payoffs to debtholders do not depend on the cash flows from new investment opportunities, which make them indifferent about the project choice. A manager can thus ensure greater project autonomy by issuing debt if the current value of the firm is sufficiently high. In line with Dittmar & Thakor (2007), the model by Boot & Thakor (2007) does predict equity to be issued if both disagreement as well as the value of the assets in place is low. However, in the case where disagreement is high and the value of the assets in place is low a manager might still prefer to issue debt. Note that in this situation a manager’s choice of security issuance could depend on other factors as issuing both equity as well as debt will reduce managerial autonomy. Boot & Thakor (2011) conclude with several testable empirical predictions,

(12)

12

some of which this paper will address. The importance value of managerial autonomy is not only highlighted by these two papers. In their survey paper Graham & Harvey (2001) find that the degree of financial flexibility has substantial influence over security issuance decisions.

In a related paper, Gomes & Philips (2004) present similar results. They study why public firm’s issue public and private securities and whether the types of securities differ depending on each market. Specifically, they examine how two major determinants; asymmetric information and firm risk & investment opportunities, affect security issuance motives. Using a dataset of over 13,000 security issues they find that firms with higher disagreement parameters are more likely to issue securities in a private market and in particular, issue private equity and convertibles1. This result is also addressed by Boot & Thakor (1993) who argue that firms with high asymmetric information might choose to issue more information sensitive securities, such as equity, to promote information production and make insider trading more profitable. Additionally, Gomes & Philips (2004) find that, conditional upon issuing in a public market, firms with higher disagreement parameters are less likely to issue (public) equity and more likely to issue (public) debt. This result is consistent with the findings in Dittmar & Thakor (2007). Lastly, their results show that firms that are riskier and have more growth opportunities are more likely to issue equity, perhaps due to the associated potential agency problems.

Both Gilchrist et al. (2005) and Chemmanur et al. (2010) study how disagreement amongst investors affects security issuance decisions. Gilchrist et al. (2005) argue that dispersion in investor beliefs can cause stock price bubbles in the presence of short-selling constraints. Thus, according to the market-timing theory of Baker & Wurgler (2002), one would expect an increase in equity issues. Using dispersion of analysts’ earnings forecasts as a measure of disagreement they find higher dispersion of investor beliefs increases the probability of an equity issuance. Similar results are found by Chemmanur et al. (2010), who study how investor optimism and dispersion of investor beliefs affects the probability of firm issuing equity. They show that both higher optimism and higher dispersion increase the probability of an equity issue. Furthermore, they highlight the importance of investor heterogeneity on a firm’s financing decision. Huang & Thakor (2013) study the relationship between investor heterogeneity and share repurchases. They find evidence, in line with Dittmar & Thakor (2007) that firms tend to repurchase shares when agreement between managers and investors is lower and subsequently find that this action increases manager-investor agreement. They also find that dispersion of opinion between investors is unable to explain repurchase decisions once they control for the stock price and manager-investor agreement.

Research related to manager-investor disagreement, but to a lesser extent to security issuance are the papers by Bakke & Whited (2010) & Thakor & Whited (2011). Both papers study the effect of manager-investor disagreement on corporate investments and find a negative relationship between

1_{Gomes & Philips (2004) use analyst earnings forecast error (forecast – actual EPS) and dispersion as two}

(13)

13

disagreement and corporate investments. This relationship is also found by Dittmar & Thakor (2007). Although these papers are not directly related to this research it could be worthwhile to take into consideration as disagreement, security issuance behaviour and corporate investments all seem to affect each other. Furthermore, these papers can help in establishing a suitable proxy for manager-investor disagreement.

3. Methodology

This section will formulate the hypotheses which are derived from the theories described in the previous section. Subsequently, the empirical model that is used to test these hypotheses is discussed. The purpose of this model is to test whether a causal effect between disagreement and security issuance decisions exists.

Debt Issue Equity Issue

Trade-off theory • Until optimal debt level is _reached • When stock prices rise

• When optimal debt level is reached

• When stock prices decrease

Pecking order theory • Until debt capacity is reached

• When debt capacity is reached and internal funds are depleted

Market-timing theory • When stock prices are low and decreasing

• When stock prices are high and rising

Disagreement theory

• Disagreement is high and the value assets in place is low

• When the value assets in place is high

• Disagreement is low and the value of assets in place is low

Figure 1: Predictions of Security Issuance Theories

A. Hypotheses

The objective of this research is to establish whether lower disagreement between managers and investors leads to a higher probability of an equity issue. This follows from the theory developed by Dittmar & Thakor (2007). It states that a manager values autonomy in project choice and this autonomy is more firmly established by issuing equity when disagreement is low. A similar hypothesis has been suggested by Gomes & Philips (2004), which states the probability of issuing more information sensitive securities, such as equity, is decreasing in the degree of asymmetric

(14)

14

information2. It should be noted that this prediction is conditional upon issuing in a public market. Figure 1 helps in distinguishing between the different theories of security issues. It serves as a tool to clearly distinguish the formulated hypothesis from the different theories. The first hypothesis can thus be formulated as follows:

Hypothesis 1: Firms with less disagreement between managers and investors have an increased probability of issuing equity.

Both Dittmar & Thakor (2007) and Gomes & Philips (2004) find evidence that aligns with this hypothesis. As figure 1 shows, the stock price is an important determinant in all security issuance theories. Dittmar & Thakor (2007) argue that stock prices tend to be higher when disagreement is low. Thus, solely examine and test the disagreement theory and exclude all other issuance theories motives, this paper controls for several pricing control variables.

The second hypothesis is constructing following Boot & Thakor (2011). They propose a similar model to Dittmar & Thakor (2007) but allow for certain situations where the hypothesis does not hold. In particular, they consider a situation where the current value of the assets in place is sufficiently high for debtholders to be indifferent about project choice. This means that a manager would achieve greater autonomy over project choice by issuing debt, rather than equity. The following hypothesis arises from this prediction:

Hypothesis 2: Firms with high valued assets in place have an increased probability of issuing debt, regardless of the level of disagreement.

The following subsection will discuss how these hypotheses will be tested and the subsequent result will be presented in section 5.

B. Method

This research will use binary choice model to establish a relationship between disagreement and the probability of issuing equity. Specifically, a logit model with the following basic form is used:

𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃( 𝑌𝑌 = 1⃓ 𝑋𝑋 ) = 𝐹𝐹(𝑋𝑋 𝛽𝛽) (1)

where Y is a binary variable, X the independent variables and F represents the cumulative logistic distribution function. The model is estimated using the maximum likelihood method. This method has

2_{Fulghieri & Lukin (2001) show that, in contrast to the pecking order theory, firms may sometimes prefer to}

(15)

15

been employed by several papers, including Dittmar & Thakor (2007), Gomes & Philips (2004), DeAngelo et al. (2010) and Leary & Roberts (2010). Chemmanur et al. (2010) and Dong et al. (2012b) use a maximum likelihood probit model in a similar manner.

To control for factors that might affect the probability of an equity issue a multivariate logit regression framework with specifically designed control variables is used. Following Dittmar & Thakor (2007) the regression will looks as follows:

𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃( 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑖𝑖 = 1 ) = 𝐹𝐹(𝛽𝛽0+ 𝛽𝛽1 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐸𝐸𝐸𝐸𝐷𝐷𝐸𝐸𝐷𝐷𝐷𝐷𝑖𝑖+ 𝛾𝛾1𝐶𝐶𝐶𝐶𝐷𝐷𝐷𝐷𝐷𝐷𝐶𝐶𝐶𝐶𝐷𝐷𝑖𝑖+ 𝜀𝜀𝑖𝑖) (2)

where the dependent variable Equity is a dummy variable that equals 1 for an equity issue and 0 for a debt issue. All disagreement parameters are represented by DISAGREEMENT. Variables that act as a proxy for pricing variables and all other control variables are encompassed by CONTROLS. Each variable specification and construction is discussed in the next section. The first hypothesis is tested controlling for specific pricing variables and control variables as used in Dittmar & Thakor (2007). To verify the robustness of results additional controls are added and an extended sample period is used. The regression is repeated to check whether there is a difference between the sample period used in Dittmar & Thakor (2007) and the additional years added to the sample. The second hypothesis can be tested in a similar way, however, some adjustments are made to make this feasible. Specifically, a distinction will be made between firms with high valued assets in place and low valued assets in place. This distinction will be made at the median of the ratio of intangible assets divided by total assets measured in the fiscal year prior to the security issue. Firms with a low value for this ratio are considered to have a high value for the assets in place while firms with a high value for this ratio are expected to have a low value for the assets in place. This distinction will help in testing the second hypothesis. To confirm the robustness of the model the same test can be conducted using different subsamples of the dataset.

4. Data & Descriptive Statistics

A. Data & Sample Selection

This study uses a sample of U.S. listed firms for the period 1993 – 2014. Four data sources are used to construct the dataset, which are all available through Wharton’s Research Data Services. Data on security issues is collected from the Thomson One New Issues Database. To construct the disagreement parameters, analyst EPS forecast data is obtained from the Institutional Brokers Estimates System (I/B/E/S) database. The Centre for Research in Security Prices (CRSP) is used to gather monthly data on stock prices and subsequent returns. Lastly, annual accounting data is retrieved from Compustat’s North America Fundamentals database.

(16)

16

This research will analyse firms that issued nonconvertible debt or seasoned equity in the sample period. Any firms that issued equity and debt in the same year are excluded. If a firm has multiple similar security issues in a year only the first firm-year observation is kept. This data is merged with the data on analyst forecast and security prices. Observations that miss data on either of these are excluded from the study. The sample is then matched with accounting data from Compustat. Firms for which total assets or sales is either missing, negative or zero are deleted. In line with related research, firms whose Standard Industrial Classification (SIC) is between 4900 and 4999 (utilities), between 6000 and 6999 (financial), or greater than 9000 (public) are dropped as to avoid regulated or quasi-public firms. All issues of preferred stock, perpetual debt, tracking stock and depository shares are also excluded to provide a clean sample. To reduce the impact of outliers all variables are winsorized at a 1% level if necessary. The final sample consists of 6,403 seasoned equity issues and 3,416 nonconvertible debt issues made by 3,960 different firms.

B. Variables

B.1. Disagreement Parameters

Several proxies are employed to distinguish disagreement between managers and investors. The first disagreement measure is constructed following Dittmar & Thakor (2007) and Thakor & Whited (2011). It is computed as the difference between a firm’s actual EPS and the mean of the analyst forecasted EPS no more than 60 days before the announcement of the actual EPS normalized by the actual EPS. This measure is calculated in the quarter prior to the security issue. This variable is referred to as Actual – Forecast EPS throughout this paper. The amount by which the actual EPS exceeds the forecast is negatively related to disagreement between managers and investors. In other words, investors are more likely to agree with a manager if there is a positive earnings surprise. While on could argue that even a positive earnings surprise is a sign of disagreement, as is apparent from the diverging priors, this paper focusses on the autonomy that a positive earnings surprise attains. Thus, the expectation is that the higher this measure is the more likely firms are to issue equity.

Dittmar & Thakor (2007) also provide a second disagreement measure, which is more frequently used in related literature. It is defined as the standard deviation of the analyst forecasts in the quarter prior to the security issue normalized by book equity. This measure is able to measure disagreement among investors. The idea is that disagreement among investors connotes to disagreement between managers and investors in a similar way. This implies the expectation that the lower this measure is, the lower disagreement between managers and investors is, the more likely firms are to issue equity. Throughout the paper the variable will be referred to as Standard Deviation of Analyst Forecast. It should be noted that there is no consensus in the literature regarding the effect of this measure on the propensity to issue equity as both Chemmanur et al. (2010) and Gilchrist et al. (2005) find that higher analysts’ dispersion leads to a higher probability of an equity issue.

(17)

17 B.2. Pricing Variables

To rule out the effect of the market-timing hypothesis on security issuance behaviour I control for various stock price variables, similar to those described in Dittmar & Thakor (2007). Raw returns are calculated for the 3, 6, 9, and 12 months prior to the security issue. Market adjusted returns (raw return – market return) are calculated for the same periods. Lastly, the market-to-book ratio for the fiscal year prior to the issue date is used. The market-to-book ratio is calculated as the market value of equity plus the book value of debt divided by book value total assets.

B.3. Control Variables

Firm-specific accounting variables are calculated using the data from Compustat. These are used to control for factors that could affect the issuance decision. All variables are measured at the fiscal year-end preceding the security issue. Many of the control variables used in this paper are commonly used in papers related to this topic. Studies have shown that larger firms, on average, have lower asymmetric information problems as they tend to receive more analyst coverage and tend to be more transparent (Chemmanur et al. 2010). This could lower the cost of debt for larger firms and tilt their security issue preference towards debt. Therefore, firm size is controlled for, as the natural logarithm of total assets. To control for a firm’s profitability, operating income before depreciation divided by total assets is used. Dittmar & Thakor (2007) argue that firms with higher profitability tend to have lower leverage and this in turn affects the security issuance decision. Thus, firms with higher profitability are expected to favour issuing debt. Additionally, firms with higher profitability also tend to generate more cash. Cash-rich firms can rely more internal financing which eases the pressure on their need for external financing. To address this issue financial slack is controlled for. Financial slack is defined as cash and cash equivalents over total assets. Growth options of firms are measured by research and development expense. It has been shown that firms with more growth options are more likely to issue equity. To control for this the measure is normalized by total assets. Rajan & Zingales (1995) show that a firm’s tangibility affects their leverage ratio. Firms with higher tangibility can post more collateral and are able to reduce their cost of debt. They might therefore prefer debt issuance over equity issuance. Tangibility is calculated as the total of property, plant and equipment divided by total assets. Naturally, leverage is controlled for. Highly levered firms are more likely to issue equity than low levered firms. Leverage is measured as total book value of debt divided by total book value of assets.

In addition to the commonly used control variables used in Dittmar & Thakor (2007) this paper will include several other control variables that might affect security issuance decisions. Firstly, DeAngelo et al. (2010) show that more mature firms are less likely to issue equity and more likely to issue debt. More mature firms are usually in a more stable phase of their corporate lifecycle and have fewer growth options. Therefore, their cost of debt decreases making issuing debt more favourable. To control for this the age of a firm is included in the regression. Firm age is defined as the number of

(18)

18

years since the IPO date. The Altman Z-score (Altman, 1968) is included to control for the probability of financial distress. Firms with a higher probability of distress are riskier which will increase their cost of debt. They are therefore more likely to issue equity. The Altman Z-score is calculated according to Altman (1969), who shows that firms with higher z-scores have a lower probability of default. Whether a firm has rated debt on its balance sheet is also included. Elliot et al. (2008) shows that firms with a credit rating tend to be more levered and have lower costs of debt. On the other hand, having a credit rating might increase information production and subsequently decrease disagreement between investors and management. This could induce management to issue equity instead of debt. The effect of credit rating is therefore ambiguous. Ratings dummy is a variable that equals 1 if a firm has a credit rating and 0 otherwise. Capital expenditures ratio is added as an additional firm growth indicator. Capex ratio is calculated as total capital expenditures over total assets. Lastly, cash flow volatility is controlled for which also measures the riskiness of a firm. This is defined as the standard deviation of cash flows (operating income before depreciation) using a maximum ten years of data prior to the issue. Leary & Roberts (2010) show that, on average, equity issuers tend to have more volatile cash flows. Summary statistics regarding all these variables are presented in the next subsection.

C. Descriptive Statistics

This section will describe the final sample and summarize firm characteristics. The final sample consists of 3,960 unique firms with a total of 9,819 securities issues. Table I presents the yearly distribution of debt and equity issues. As can be seen from table I the final sample is made up out of 3,416 debt issues and 6,403 equity issues. It shows that, on average, the number of equity issues exceeds the number of debt issues each year. Only in 1998 have there been more debt issues. It also shows that the number of equity issues has increased over the years while the number of debt issues has been relatively stable. Interesting to see is that in the first year of the financial crisis (2008) the number of equity issues has decreased significantly. In the years just prior to the crisis a huge misbalance can be seen with a near average of four times as many equity issues than debt issues.

Table II presents the firm characteristics for debt issuers and equity issuers. The difference between the two subgroups is listed in the last column and all differences are statistically significant. As can be seen from the table equity issuers are on average smaller, have lower profitability, and have fewer tangible assets ratio compared to debt issuers. These findings are in line with the results from other studies. The table also shows that firms who issue equity tend to be younger than firms who issue debt. Also in line with expectations is the fact equity issuers have a higher R&D ratio and are riskier as measured by the Altman Z-score. Interestingly, the ratings dummy is higher for debt issuers giving favour to the theory that having rated debt decreases the cost of debt. Furthermore, table II shows that equity issuers have lower leverage and are more cash-rich than debt issuers. When looking

(19)

19

at cash flow volatility it appears that debt issuers are riskier firms than debt issuers. This is interesting as the expectation is that riskier firms prefer to issue equity. Lastly, the table shows that debt issuers have higher capital expenditures than equity issuers although the difference is minimal. It thus appears that debt issuers are able to spend more freely than equity issuers.

Table I

Distribution of Equity and Debt Issues for the Sample Period test

This table presents the distribution of nonconvertible debt issues and seasoned equity offerings for US listed firms in the period 1993 – 2014. Data on disagreement parameters, pricing variables and control variables is available for all firms. Security issues data is obtained from the Thomson One New Issues Database. All financial and quasi-public firms are excluded from the sample. All types of preferred stock, perpetual debt and depository shares are omitted from the sample as well as firms that issued equity and debt in the same year. The final sample consists of 3,416 debt issues and 6,403 equity issues.

t test e

Year Debt Issues Equity Issues Total

1993 137 215 352 1994 118 174 292 1995 134 271 405 1996 155 294 449 1997 182 287 469 1998 209 197 406 1999 145 235 380 2000 105 270 375 2001 135 358 493 2002 104 341 445 2003 119 353 472 2004 98 405 503 2005 91 323 414 2006 129 321 450 2007 136 331 467 2008 122 182 304 2009 186 339 525 2010 227 283 510 2011 186 259 445 2012 257 293 550 2013 239 350 589 2014 202 322 524 Total 3416 6403 9819

The second part of the table distinguishes between firms with high valued assets in place and firms with low valued assets in place. It shows that firms with low valued assets in place are, on average, larger and older. Whereas these results might raise some questions other results are more in line with expectations such as the fact that, on average, firms with higher valued assets in place have more tangible assets and a higher leverage. This indicates that firms with higher valued assets in place are more likely to issue debt, regardless of disagreement.

(20)

20 Table II

Descriptive Statistics - Control Variables

test

Table II presents the control variables for the full sample and illustrates the differences in firm characteristics between equity and debt issuers. The means are provided in the middle two columns while the last column presents the difference in mean between the two subgroups. Stars indicate whether the difference is statistically significant using a standard t-test. Variables are defined as follows: Firm Size is defined as the natural logarithm of total assets; Profitability is operating income before depreciation divided by total assets; Financial Slack is defined as cash & cash equivalents divided by total assets; R&D Ratio is research & development costs divided by total assets; Tangibility is defined as property, plant & equipment divided by total assets; Book leverage is total book debt over total assets; Firm Age is defined as the number of years since the IPO date; Altman Z-score is defined as in Altman (1968); Ratings Dummy is a dummy variable that equals 1 if a firm has a credit rating and 0 otherwise; Cash flow Volatility is the standard deviation of the operating income before depreciation calculated using ten years of observations; Capex Ratio is capital expenditures divided by total assets. All variables are calculated in the fiscal year prior to the issue. The first section of the table details the full sample. The second section of the table distinguishes between firms with high value of assets in place and firms with low valued assets in place. The distinction between firms is made at the median of the ratio of intangible assets divided by total assets.

Test

Debt Issuers Equity Issuers Difference

Full Sample Firm Size 8.38 5.23 3.151*** Profitability 0.15 -0.023 0.173*** Financial Slack 0.078 0.31 -0.229*** R&D Ratio 0.016 0.13 -0.116*** Tangibility 0.37 0.24 0.129*** Book Leverage 0.31 0.23 0.0880*** Firm Age 16.8 12.0 4.802*** Altman Z-score 4.09 3.07 1.023*** Ratings Dummy 0.87 0.20 0.674***

Cash Flow Volatility 420.9 46.1 374.8***

Capex Ratio 0.073 0.067 0.00556***

Observations 3416 6403

High & Low Value of Assets in Place

High Value Assets in Place

Firm Size 8.12 4.84 3.283*** Profitability 0.15 -0.079 0.230*** Financial Slack 0.078 0.41 -0.337*** R&D Ratio 0.011 0.18 -0.165*** Tangibility 0.52 0.28 0.243*** Book Leverage 0.32 0.19 0.124*** Firm Age 15.8 11.5 4.277*** Altman Z-score 3.89 2.96 0.931*** Ratings Dummy 0.86 0.14 0.720***

Capex Ratio 0.11 0.083 0.0265***

(21)

21

Low Value Assets in Place

Firm Size 8.59 5.77 2.818*** Profitability 0.15 0.043 0.109*** Financial Slack 0.079 0.17 -0.0948*** R&D Ratio 0.018 0.076 -0.0576*** Tangibility 0.27 0.19 0.0788*** Book Leverage 0.31 0.27 0.0393*** Firm Age 18.1 12.6 5.478*** Altman Z-score 4.21 3.04 1.171*** Ratings Dummy 0.89 0.29 0.604***

Capex Ratio 0.048 0.045 0.00267*

*

p < 0.10, ** p < 0.05, *** p < 0.01

Table III presents the summary statistics of the disagreement parameters for debt issuers and equity issuers. It shows that for the variable Actual – Forecast EPS the mean of equity issuers is 0.00828 higher than for debt issuers. As this variable is a proxy for disagreement it implies that, on average, there is less disagreement amongst equity issuers. The difference between debt and equity issuers is statistically significant at a 1% level. The difference in the standard deviation of analyst forecast proxy is also significant.

Table III

Measures of Disagreement test

Table III presents the mean of the disagreement parameters used in this research for debt issuers and equity issuers. The last column presents the difference in mean between the two subgroups. Stars indicate whether the difference is statistically significant using a standard t-test. The number of observations is given for each parameter. Actual – Forecasts EPS is defined as the actual EPS minus the mean of the analyst forecasts divided by the actual EPS measures in the quarter preceding the issue. St. Dev. Forecast measures the standard deviation of the analyst forecast in the quarter preceding the issue. This measure is normalized by book equity.

Debt Issuers Equity Issuers Difference Actual-Forecast EPS -0.022 0.061 -0.0828***

[0.7] [0.8] (-5.15)

Obs. 3358 5913

Std. Dev. Forecast/Book Equity 0.0079 0.035 -0.0276***

[0.0] [0.1] (-18.33) Obs. 3230 4962 Standard Deviation 0.050 0.038 0.0127*** [0.1] [0.1] (9.64) Obs. 3230 4962 Observations 3416 6403

t statistics in parentheses, standard deviation in brackets

*

(22)

22

Table III shows that equity issuers have a lower value for this parameter which is in line with the prediction that firms with less disagreement are more likely to issue equity. Interestingly, when this measure is normalized by the book value of equity are different result emerges. Equity issuers now have a 0.0276 higher mean than debt issuers. According to this measure, it now appears that firms with more disagreement are more likely to issue equity. However, no causal relationship between disagreement and security issuance behaviour can be established from table III as these results do not control for any variables that might influence issuance decisions. It should also be noted that the number of observations for the standard deviation of analyst forecasts decreases by nearly a 1000 for equity issuers compared the first disagreement parameter.

Table IV

Descriptive Statistics – Pricing Variables test

Table IV shows that equity issuers have higher stock returns than debt issuers. The means of the pricing variables are presented in the middle two columns where a distinction is made between debt issuers and equity issuer. The last column shows the difference between the two with stars indicating whether this difference is statistically significant using a standard t-test. The 3-month Return variable is defined as the average three month stock return in the months preceding the issue. The 6-month, 9-month, 12-month Returns are calculated in a similar way. Market adjusted returns are calculated by subtracting the return of the S&P500 from the firm specific stock returns in the months preceding the issue. The Market-to-Book Ratio is defined as the market value of equity plus the book value of debt divided by the book value of total assets.

Test

Prior 3-month Return 0.037 0.18 -0.144***

[0.2] [0.4] (-18.44)

[0.3] [0.8] (-22.46)

[0.4] [1.1] (-22.78)

[0.5] [1.3] (-21.71)

Market-to-Book Ratio 1.76 3.27 -1.503***

[1.0] [4.0] (-20.71)

Market Adj. Prior 3-month Return 0.023 0.15 -0.132***

[0.2] [0.4] (-14.44)

[0.3] [0.8] (-16.93)

[0.4] [1.1] (-16.62)

[0.6] [1.4] (-15.46)

t statistics in parentheses, standard deviation in brackets

*

p < 0.10, ** p < 0.05, *** p < 0.01

Table IV presents the average stock returns and market-to-book ratio for debt and equity issuers. It provides a clear picture of the differences between the two subgroups; equity issuers have

(23)

23

higher stock prices and higher subsequent returns compared to debt issuers. To illustrate this the table shows the 3, 6, 9, and 12-month returns prior to the issue. Market adjusted returns are also presented for the same months. In all cases equity issuers exceed debt issuers in stock returns with statistically significant differences. Furthermore, equity issuers have an average market-to-book ratio of 3.27 while debt issuers have an average market-to-book ratio of a little more than half of that. These results argue in favour of the market-timing hypothesis, which states that firms are more likely to issue equity when their stock prices are high. Thus, by including these pricing measures into the multivariate regression the effect of the market-timing hypothesis can be controlled for and a possible causal interpretation can be given to the effect of disagreement on security issuance decisions.

Lastly, tables A.I and A.II in the appendix provide additional information on the distribution of the sample. Table A.I in the appendix presents the distribution of security issues by industry. From this table it can be seen that equity issues dominate the manufacturing and service industry and that debt issues are more frequent in the mining, construction, transportation, and retail industry. These results could indicate that the manufacturing and service industries have higher agreement parameters and lower values of assets in place. Similarly, it predicts that the mining, construction, transportation, and retail industries have higher valued assets in place. Evidence for this is presented in table A.II. This table presents the distribution of firms with high valued assets in place and low valued assets in place by industry. Firms with high valued assets in place are represented more in the mining, construction, and retail industry, while firms with low value assets in place are represented more in the manufacturing and services industry. Both tables combined provide additional evidence that firms with higher valued existing assets are more likely to issue debt, regardless of the level of disagreement. Table A.III and A.IV present the distribution of security issues by the age of a firm. From these tables it can be seen that younger firms are more likely to issue equity as well as have lower valued assets in place.

5. Results

The following section presents the outcome of the empirical analysis to test both hypotheses formulated in section 3.

A. Hypothesis 1

Table V presents the regressions results of the logit model that tests the effect of disagreement on security issuance behaviour. The dependent variable is a dummy that equals 1 if a firm issued equity and 0 if a firm issued debt. To check for multicollinearity a Collin test is performed on all the independent control variables. The outcome of the test shows that the regressions do not suffer from multicollinearity that could alter the regression results. Further details are presented in the appendix. The results support the prediction made in hypothesis 1 and clearly establishes a causal relationship between disagreement and the decision to issue a certain security.

(24)

24

The Actual - Forecast EPS disagreement proxy is used in all regressions in table V. The first three regressions are done using the same controls as Dittmar & Thakor (2007) and different pricing variables are used to control for market-timing effects. The table shows a positive relationship between manager-investor agreement and the decision to issue equity. The strongest significant effect is observed in regression three where the pricing variable that is controlled for is the market-to-book ratio. This result provides evidence in favour of hypothesis 1 and shows that firms who experience less disagreement are more likely to issue equity. In addition to the controls used in Dittmar & Thakor (2007) extra controls are added in regression 4, 5, and 6. The pricing variable that is used is the market-to-book ratio. As can be seen from regression 4, the significance disappears when all the controls are added. However, when cash flow volatility is omitted in regression 5, and subsequently capex ratio in regression 6, the significance level increases to 10%. This change in significance could be explained by the fact that a firm’s risk is controlled by the Altman Z-score control.

Thus, adding another firm risk control, such as cash flow volatility, could adversely affect the regression. The relatively small effect of omitting capital expenditures in the regression can be explained by the fact that capital expenditures in itself could capture some of the disagreement. Capital expenditures tend to fluctuate a lot which can be driven by disagreement. Firms with higher capital expenditures are usually able to do so because there is more agreement between managers and investors. This would allow managers to issue and spend funds more freely. Thus, it can be seen how disagreement affect capital expenditures as well as issuance decisions. The first six regressions are conducted using the full sample period from 1993 – 2014. However, when the sample period is adjusted some interesting results emerge. In the column 7, regression 3 is replicated but for an adjusted sample period that matches the sample period of Dittmar & Thakor (2007). A stronger and more significant effect of disagreement on security issuance decisions now emerges. Even when additional controls are added the results remains significant at a 1% level. When the same regression is conducted on the sample period 2002-2014 the effect disappears. It appears that for this period disagreement has little explanatory power over issuance decisions. A possible explanation for this difference could be that the recent financial crisis affected security issuance motives. To illustrate this two additional regressions are performed. Columns 10 and 11 employ the full sample period but omit security issues made in 2008 until 2010. In these years the financial crisis is considered to have had its biggest impact. Both regressions show stronger significant results for the disagreement parameter when these crisis years are omitted. It not only shows how the financial crisis affected issuance decisions but also the fragility of security issuance theories. They illustrate the conditionality of these theories.

(25)

Table V

Testing Hypothesis 1: Actual – Forecast EPS Measure test

Table V presents the logit regression results using the Actual – Forecast EPS disagreement proxy. It illustrates the effect of disagreement on the security issuance decisions. The dependent variable is a dummy that equals 1 for an equity issue and 0 for a debt issue. Actual – Forecast EPS is calculated as the actual EPS minus the forecasted EPS divided by the actual EPS no more than 60 days prior to the issue. All other variables are calculated as described in tables I through IV and are measured in the fiscal year prior to the issue. Regressions 1 through 6 are done using the full sample. Regressions 7, 8 and 9 use the sample period as indicated at the top of each column. The last two regressions use the full sample period with the exception of the years 2008 until 2010. All regressions use similar controls as Dittmar &Thakor (2007) except regressions 4 through 6, 8, 10, and 11, where additional controls are added. Heteroskedastic robust standard errors are used in all regressions.

test

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

1993 - 2014 1993-2002 2002-2014 No Crisis Years

Actual – Forecast EPS 0.089* 0.079* 0.122** 0.024 0.129* 0.139* 0.235*** 0.326*** 0.015 0.163** 0.173** (1.9) (1.7) (2.6) (0.3) (1.8) (1.9) (3.0) (3.0) (0.3) (2.1) (2.2) Prior 3-month Return 0.988***

(8.0)

Prior 12-month Return 0.439*** (7.0) Market-to-Book Ratio 0.226*** 0.217*** 0.211*** 0.190*** 0.260*** 0.303*** 0.154*** 0.223*** 0.203*** (5.7) (3.4) (4.0) (3.9) (4.4) (2.8) (3.1) (3.8) (3.8) Firm Size -0.931*** -0.918*** -0.935*** -0.766*** -0.665*** -0.637*** -0.866*** -0.458*** -1.088*** -0.648*** -0.621*** (-39.8) (-36.9) (-37.7) (-11.5) (-14.4) (-14.0) (-23.8) (-6.3) (-29.8) (-13.5) (-13.1) Profitability -3.112*** -3.711*** -4.677*** -4.554*** -2.702*** -2.363*** -2.448*** 0.855 -5.737*** -2.426*** -2.038** (-8.4) (-9.3) (-8.7) (-4.9) (-3.3) (-2.8) (-3.0) (0.6) (-8.8) (-2.7) (-2.2) Financial Slack 1.960*** 2.165*** 1.268*** 0.320 -0.245 -0.139 1.433** 0.511 0.924** -0.501 -0.370 (6.2) (6.5) (3.8) (0.7) (-0.6) (-0.3) (2.5) (0.7) (2.3) (-1.2) (-0.8) R&D Ratio 2.781*** 2.661*** 2.149** 2.787 5.070*** 4.512*** 4.716*** 8.974*** 1.251 5.763*** 5.101*** (3.1) (2.8) (2.3) (1.6) (3.2) (2.9) (3.0) (3.4) (1.0) (3.2) (3.0) Tangibility -0.194 -0.029 -0.019 -0.151 -0.211 -0.787*** -0.572*** -1.328*** 0.258 -0.438 -0.958*** (-1.5) (-0.2) (-0.1) (-0.4) (-0.7) (-3.7) (-2.7) (-3.8) (1.5) (-1.4) (-4.1) Book Leverage 0.553*** 0.457*** 0.557*** 1.350*** 1.358*** 1.409*** 0.017 0.534 0.901*** 1.310*** 1.363*** (3.3) (2.6) (3.1) (4.1) (4.6) (4.8) (0.1) (0.9) (3.9) (4.1) (4.3)

(26)

26 Firm Age -0.060*** -0.056*** -0.053*** -0.017* -0.050*** -0.047*** (-7.1) (-7.9) (-7.5) (-1.7) (-6.7) (-6.4) Altman Z-score -0.039 -0.054** -0.061** -0.098* -0.046* -0.056** (-1.4) (-2.2) (-2.4) (-1.8) (-1.8) (-2.0) Ratings Dummy -1.294*** -1.462*** -1.482*** -1.178*** -1.428*** -1.450*** (-8.6) (-10.7) (-10.9) (-5.4) (-9.7) (-9.9) Capex Ratio -1.782* -2.039*** -1.765** (-1.9) (-2.7) (-2.2)

Cash Flow Volatility -0.000

(-0.8) Constant 6.840*** 6.703*** 6.744*** 7.810*** 6.965*** 6.766*** 5.867*** 4.565*** 8.308*** 6.730*** 6.543*** (34.6) (32.0) (32.3) (15.1) (18.4) (18.3) (19.7) (8.5) (27.2) (16.9) (16.7) Pseudo R2 0.460 0.464 0.462 0.482 0.448 0.441 0.422 0.341 0.507 0.433 0.427 Observations 9066 8276 8256 3427 4490 4523 3317 1841 5323 3931 3964 t statistics in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01

How disagreement affects corporate decisions : an analysis of security issues