1 University of Amsterdam
Amsterdam Business School
MSc Finance in Quantitative Finance
Master Thesis
Impact of Industry Competition on Dynamic Capital Structure
Author: Jiani Geng
Student number: 10830685
Thesis supervisor: Prof. Stefen Arping Date: December 2018
2
Statement of Originality
This document is written by Student Jiani Geng who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are 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 is responsible solely for the supervision of completion of the work, not for the contents.
3
Abstract
This paper examines how industry competition influences target capital structure and speed of adjustment toward target leverage. The sample consists of U.S. public firms across nine industries from 1997 to 2016. The findings show that industry competition decreases target debt ratio. In addition, industry competition is positively related with adjustment speed towards target capital structure. However, these results apply to market leverage only.
4
Table of Content
1. Introduction ... 5
2. Literature Review ... 8
2.1 Capital Structure Theories ... 8
2.1.1 Static Trade-off Theory ... 8
2.1.2 Pecking Order Theory ... 9
2.1.3 Market Timing Theory ... 9
2.1.4 Dynamic Trade-off Theory ... 10
2.2 Determinants of Optimal Capital Structure ... 10
2.3 Adjustment Speed to Target Capital Structure ... 13
2.4 Determinants of Speed of Adjustment ... 15
2.4.1 Firm-specific factors ... 15
2.4.2 Macroeconomic factors ... 16
2.5 Hypotheses ... 18
3. Methodology ... 20
3.1 Regression Model for Hypothesis 1 ... 20
3.1.1 Measurements for dependent variable ... 20
3.1.2 Measurements for independent variables ... 21
3.2 Regression Model for Hypothesis 2 ... 23
4. Data and Descriptive Statistics ... 25
5. Results ... 27
5.1 Estimation of Target Capital Structure ... 27
5.2 Estimation of Adjustment Speed ... 28
6. Robustness checks ... 32
7. Conclusion ... 36
5
1. Introduction
Capital structure is the focus of a firm’s financial decision. A lot of empirical studies have been conducted to investigate target capital structure in the past decades. Apart from static capital structure, dynamic capital structure is of growing importance as well. In presence of transaction costs, firms actually deviate from the optimal leverage that minimizes the cost of capital and maximizes the firm value. Empirical evidence shows that firms do rebalance their capital structures frequently but not continuously due to adjustment costs (Fischer, Heinkel, & Zechner, 1989). This dynamic behavior leads to the question of the speed of adjustment toward optimal capital structure. Flannery and Rangan (2006) employ a dynamic partial adjustment model including firm fixed effects and estimate that firms converge to the target leverage at the speed of about 30% per year. However, a major drawback of Flannery and Rangan’s methodology is the assumption of same speed of adjustment which is challenged by cross-sectional heterogeneity because of the adjustment costs. Drobetz and Wanzenried (2006) analyze the influence of firm-specific characteristics on the speed of adjustment to the optimal financial structure of Swiss firms, and they document that firms adjust faster if they have greater growth opportunities or their deviations from the target ratio are larger. Further evidence of the asymmetric adjustment speed is shown between under-levered firms and over-levered- firms (Byoun, 2008). In particular, the most significant adjustments take place when over-levered firms have financial surplus. Faulkender et al. (2012) demonstrate that positive cash flows enable firms to adjust at lower costs leading to a positive relationship between cash realization and speed of adjustment.
In addition to the transaction costs, financial distress costs play a vital role in the adjustment speed as well. Titman and Tsyplakov (2007) find that financial distress costs have a first-order impact on both target capital structure and adjustment speed toward the target. Specifically, financial distress costs tend to decrease target leverage and increase the speed of adjustment.
The adjustment speeds are heterogeneous across industries as well (Elsas & Florysiak, 2011). They classified U.S firms into 43 industry groups based on Fama and French (1997) industry classification. Results show that precious metals industry has fastest
6
speed of adjustment while textiles industry has slowest speed of adjustment. They suggest that the heterogeneity might be caused by industry-specific adjustment costs. However, what factors are responsible for heterogeneous speeds between industries is still unknown.
The objective of this paper is to analyze how industry competition influences the speed of adjustment toward target capital structure. Past research pays little attention to the impact of industry competition on speed of adjustment. Industry characteristics affect firms’ choice of financial structure (MacKay & Phillips, 2005). In concentrated industries, firms without strong competitors tend to be larger, and their profits are higher and more stable. On the other hand, firms in competitive environment have lower and less stable profits. These firms have weak capital structure and are easily attacked by predatory behavior by financially strong firms (Brander & Lewis, 1986). Therefore, financial distress costs for such firms are costly and they are expected to have lower debt ratio and faster adjustment speed toward the target leverage.
The model employed in this paper builds on the dynamic partial adjustment model of Flannery and Rangan (2006). The difference is that a heterogeneous speed is indicated instead of assuming homogeneous speed. In addition, the estimation method is Generalized Method of Moments (GMM) which solves the endogeneity problem and provides consistent estimates. Same as Drobeta and Wanzenried (2006), the determinants of adjustment speed toward target capital structure are endogenized into the model. The coefficients on the industry competition denoting the determinants of speed are of primary interest.
The data are collected from COMPUSTAT for U.S. listed firms between fiscal year 1997 and 2016. Consistent with empirical studies, utilities and financial firms are deleted from the sample, leading to nine industry groups. After several adjustments, the final sample is an unbalanced panel data of 119,418 firm-year observations for 13,429 firms. For comparison, both book leverage and market leverage are used as proxies for capital structure (i.e. Flannery & Rangan, 2006; Frank & Goyal, 2009).
7
The main results are in line with the hypotheses. First, firms in competitive industries have lower market debt ratio. The same conclusion does not apply to book debt ratio. Second, firms in competitive industries have more rapidly adjustment speed toward target leverage. The result is not reliable to book leverage either. As a robustness check, the full sample is split into two dataset from 1997 to 2008 and from 2009 to 2016. The findings are robust to two samples.
This paper contributes to the existing literature on how industry competition affects the target capital structure and adjustment speed. Recent studies only report that adjustment speed varies across industries, but no further investigation on industry effects is conducted. This paper identifies the effect of industry competition and quantitatively examines the impact.
The remainder of this paper has the following structure. In section 2, relevant capital structure theories are first reviewed. In addition, determinants of target capital structure and speed of adjustment toward the target leverage are discussed. Finally, hypotheses are proposed. In section 3, regression models employed and estimation method are explained. Section 4 describes data used and Section 5 presents the main results. Section 6 reports robustness checks and Section 7 draws the conclusion.
8
2. Literature Review
This chapter will formulate hypotheses regarding the influence of industry competition, on the basis of the existing theoretical and empirical literature. Section 2.1 reviews the capital structure theories and section 2.2 discusses the key determinants of target capital structure. Section 2.3 discusses the empirical findings regarding the speed of adjustment toward target ratio and section 2.4 considers what factors impact the speed of adjustment. Finally, section 2.5 proposes the hypotheses.
2.1 Capital Structure Theories
The modern capital structure started with capital structure irrelevance proposition that capital structure choice does not influence firm value in a perfect capital market (Modigliani & Miller, 1958). This proposition holds only under strict assumptions, which makes it impossible to apply in reality. In addition, empirical results highlight that financing does matter. Several classical theories, taking financial frictions into consideration, have been developed since then. This section explains the four commonly accepted capital structures deciding the firms’ financial structures: static trade-off theory, pecking order theory, market timing theory and dynamic trade-off theory.
2.1.1 Static Trade-off Theory
The static trade-off theory states that optimal capital structure is determined by comparing the costs to benefits of debt (i.e. Kraus & Litzenberger, 1973; Myers, 1984; Jensen & Meckling, 1976). Tax shield on interest payments is a major benefit of debt (Kraus & Litzenberger, 1973). However, excessive debt increases the probability of bankruptcy since the tax benefit of high leverage is balanced by the increased financial distress costs.
Apart from the tax effects and costs of financial distress, agency costs play a significant role as well in the static trade-off theory (Jensen & Meckling, 1976). Conflicts between shareholders and managers raise agency problems. Managers, for their own interest, attempt to expand business and build their reputation and influence. This action, known as “empire building”, impairs the interest of shareholders as managers may invest in projects that do not maximize shareholders value. Debt repayment can reduce the free
9
cash available (Jensen, 1986). Therefore, the alleviation of incentive conflicts between shareholders and managers constitutes an advantage of debt financing.
On the other hand, debt financing causes the conflicts between debtors and shareholders (Jensen & Meckling, 1976). Shareholders have an incentive to shift the risk to debtors by investing in highly risky projects. If a risky project succeeds, shareholders capture the excess returns and debtors receive only the fixed interest payments. However, in the case of failure, it is debtors who endure the losses due to the limited liability. This behavior of risk transfer results in “asset substitution effect” and shareholders receive less than what they expected because debtors anticipate this behavior. Thus, agency costs increase with debt.
2.1.2 Pecking Order Theory
Pecking order theory states that firms have preference for capital sources in existence of issuance costs and asymmetry information between managers and investors (i.e. Myers, 1984; Myers & Majluf, 1984; Lucas & McDonald, 1990). Because of information asymmetry, managers are better informed about the true value of the firm and the business risk than outside investors (Myers & Majluf, 1984). Suppose a firm’s equity value is underpriced and the firm has a new project to be financed. If equity issuance is used to finance the project, the new investors will grab most of the profits because of underpricing. Therefore, managers will forgo the project so as to protect current shareholders, leading to an underinvestment problem. To aviod this problem, managers will choose cash or debt to finance the project. In addition, the issuance costs of new securities overwhelm other costs. Therefore, pecking order theory explains the hierarchy of internal funding and external financing. Costless cash is the first choice to fund investment opportunities. Debt is used if cash is not availiable and equity is the last resort due to the riskiness. Hence, optimal debt ratio does not exist according to pecking order theory. The capital structure reflects simply the cumulative result of external financing (Myers, 1984).
2.1.3 Market Timing Theory
Market timing theory maintains that a firm’s management waits for the market timing opportunities to change the capital structure. Intuitively, when managers find the firm’s
10
current stock price is overvalued, they will take advantage of this mispricing by issuing securities. Managers are more likely to repurchase securities when the stock price is cheap. To test the theory, Baker and Wurgler (2002) study the extent to which the firm issues equity when market-to-book ratio is high and find significant evidence that managers incline to raise capital when market values exceed book values. In other words, leverage is negatively related to market-to-book ratios. Therefore, the capital structure is the result of historical financing decisions.
2.1.4 Dynamic Trade-off Theory
Static trade-off theory focuses on a single-period and predicts that the firm always stays at optimal capital structure. An extension of static model to multi-periods leads to dynamic trade-off theory, which suggests that it is costly to keep the debt ratio at the target all the times because of frequent rebalancing of debt and equity. Fischer et al. (1989) suggests that the existence of transaction costs drives capital structure away from the optimal. Firms have a range of leverage ratio within which debt ratio is allowed to float. Once the debt ratio crosses the upper or lower limit, a firm rebalances its leverage to the optimal level. Because of transaction costs the convergence toward the optimal capital structure is slow. Graham and Harvey (2001) conduct a survey on CFOs and ask them about their financing decisions. The result shows that the majority of CFOs have a range for target debt ratio or a strict target, providing support for dynamic trade-off theory.
Leland (1994) presents a dynamic trade-off model as well. In his model, the changes in debt ratio reflect accumulated profits or losses. Further, the firms do not move leverage toward the optimum as long as the costs of adjustments exceed the costs of being at suboptimal capital structure.
2.2 Determinants of Optimal Capital Structure
Based on the capital structure theories, empirical research has been done to identify the factors that may drive firms’ capital structures. Several stylized facts have been found although studies vary in terms of sample selection, variable definition and methodology. Rajan and Zingales (1995) analyze the financing decisions of listed firms across G-7 countries and recognize four factors which are asset tangibility, growth opportunities,
11
profitability and firm size. Frank and Goyal (2009) evaluate the leverage choices of U.S public firms and indentify two more factors that are industry median leverage and expected inflation. This section will discuss abovementioned six factors in detail.
Profitability
According to trade-off theory, profitability results in higher level of debt financing as tax benefits increase with leverage and agency costs decrease with leverage (i.e. Kraus & Litzenberger, 1973; Jensen, 1986). However, pecking order theory predicts that low debt ratio should be associated with high profitability (Myers, 1984). Retained earnings are the first choice to finance positive NPV projects, followed by debt and equity. Empirical findings that profitable firms tend to have less leverage, supports pecking order theory (i.e. Frank & Goyal, 2009; Rajan & Zingales, 1995).
Growth opportunity
Firms with more growth opportunities are prone to asset substitution, worsening debt-related agency problems (i.e. Jensen & Meckling, 1976; Myers, 1984). Alternatively, growth reduces the free cash available and there is no need to issue debt in order to prevent free cash flow problem (Jensen, 1986). Thus, trade-off theory predicts that growth decreases leverage. In contrast, the prediction of pecking order theory is ambiguous. Debt increases when investment exceeds retained earnings and falls when investment is less than retained earnings. Empirical evidence suggests that leverage is negatively related to growth opportunities, in line with trade-off theory (i.e. Frank & Goyal, 2009; Rajan & Zingales, 1995).
Firm size
Large firms tend to be more diversified and their default risk is low (Titman & Wessels, 1988). In addition, expected bankruptcy costs are relatively lower for larger firms (Ang, Chua, & McConnell, 1982). Therefore, large firms are expected to take on more leverage under trade-off theory. However, pecking order theory has the opposite prediction. Large firms tend to disclose more information than small ones, resulting in less information asymmetry and lower equity issuance costs. Empirical studies document that firm size is
12
positively related to debt ratio, supporting the prediction of trade-off theory (i.e. Frank & Goyal, 2009; Rajan & Zingales, 1995).
Asset tangibility
Under the trade-off theory, tangible assets are positively related to leverage level. First, the tangibility of asset can be regarded as a measure for collateral, used to raise loans or sold to raise cash in emergency. A higher level of asset tangibility provides debtors with higher security as their repayment can be guaranteed if the firm is bankrupt. On the contrary, a low level of asset tangibility increases the risk of debtors since their losses cannot be fully recovered in case of bankruptcy. Second, tangibility makes debt less risky as debtors are secured, results in lower financial distress costs. Moreover, tangible assets are difficult to be replaced with high-risk assets, which decrease agency costs caused by conflicts between shareholders and debtors (Jensen & Meckling, 1976). On the other hand, information asymmetry decreases in asset tangibility, which makes equity less costly (Harris & Raviv, 1991). According to pecking order theory, a negative relation between asset tangibility and debt ratio is expected. Empirical evidence shows that firms with more tangible assets tend to have high leverage, providing support for trade-off theory (i.e. Frank & Goyal, 2009; Rajan & Zingales, 1995).
Industry median
It is evident that capital structure varies across industries since firms within the same industry could have similar business environment. However, empirical results show that there is significant variation in leverage choices even after controlling industry fixed effects (MacKay & Phillips, 2005). More specifically, a firm’s capital structure depends on its position within its industry as well. There is evidence that firms actively move their debts ratio toward industry median (Hovakimian, Opler, & Sheridan, 2001). Industry median leverage is often used as a benchmark for managers to decide their own firm’s optimal leverage. French and Goyal (2009) prove that firms in industries in which the median debt ratio is high incline to have high debt ratio.
13
The real value of tax shield on interest payment is higher when inflation is expected to be high (Taggart, 1985). Therefore, trade-off theory implies that expected inflation increase debt ratio as the firm benefits more from tax effect. Ritter and Warr (2002) demonstrate that the existence of expected inflation causes stocks to suffer from severe undervaluation. Accordingly, market timing theory predicts that leverage is positively related to expected inflation. Consistent with the theories, empirical results show that firms tend to use more debt when inflation is expected to be high (Frank & Goyal, 2009).
2.3 Adjustment Speed to Target Capital Structure
Graham and Harvey (2001) find that the majority of U.S firms have a target range for the leverage or even a specific target debt ratio. Current literature reveals that leverage is mean-reverting (Lemmon, Roberts, & Zender, 2008). This section will discuss empirical studies of the speed of adjustment (SOA) at which firms move toward their target capital structure.
According to Leary and Roberts (2005), who explore the factors affecting the capital structure change, SOA can show support for capital structure theories. A zero SOA could be supportive of either pecking order theory or market timing theory since it implies target capital structure may not exist. A positive but less than one SOA indicates that firms actively rebalance their leverage ratios with the presence of adjustment costs, providing support for dynamic trade-off theory.
The estimates of SOA vary over a large range, possibly due to different estimation strategies, sample selections and variable definitions. Fama and French (2002) examine the financial decisions of U.S publicly traded firms with pooled OLS regression and find a low SOA of 7%. Flannery and Rangan (2006) argue that differences in target capital structure are driven largely by firm-fixed and time-invariance factors which result in downward biases of earlier estimates. They take fixed effects into consideration and document that the SOA of U.S firms is 34.3%. However, firm-fixed effects could be correlated with lagged variable causing an upward bias of SOA. Recognizing this problem, Flannery and Rangan (2006) try to fix it by demeaning variables to sweep out the firm-fixed variable, which makes little difference. Lemmon et al. (2008) build on this work, taking a different approach to solve endogeneity problem. They find a speed of 25%
14
estimated with generalized method of moments (GMM). In addition to the estimation methods, the instruments used in these models differ as well, leading to difficulty in reaching a consensus on the consistent estimates.
There are several ways for firms to rebalance their capital structures. A firm can make internal adjustments by dividend payout or keeping net income as retained earnings. Alternatively, a firm can use external sources to change capital structure. Debt issuance or share repurchase is commonly used when the firm is under levered. Early retirement of debt or new security issuance is used when the firm is over levered.
A large body of literature suggests that transaction costs play a key role in SOA (i.e. Fischer et al., 1989; Leary & Roberts, 2005; Byoun, 2008; Faulkender et al., 2012). Evidently, a firm will not adjust its leverage if the transaction costs outweigh the costs of being off the target. Conceivably, firms differ in adjustment costs and a low SOA is associated with relatively high costs of adjustment. Byoun (2008) and Faulkender et al. (2012) show that the SOA is a decreasing function of adjustment costs in a partial adjustment model, supporting the costly adjustment view. This leads to another drawback of Flannery and Rangan’s (2006) methodology that SOA is assumed to be homogenous across firms.
On the other hand, researchers point out that the influences of adjustment costs alone on financial decisions are too small to explain the slow SOA. Titman and Tsyplakov (2007) shed light on debtor-shareholder conflicts and financial distress costs. They construct a dynamic capital structure model in which firm value and investment choices are endogenized. Their results show that conflicts between shareholders and debtors and financial distress costs profoundly affect not only the optimal leverage ratio but also the movements of debt ratio. More specifically, conflicts of interest between debtors and shareholders slow down the movement toward the target capital structure, while financial distress costs increase the SOA. Their results support the prediction of dynamic trade-off theory as costs of financial distress are part of opportunity costs of being at suboptimal capital structure. In addition to the agency problem between shareholders and debtors, Morellec et al. (2012) highlight the importance of shareholder-manager conflicts and
15
demonstrate that the SOA decreases in agency costs. This is consistent with the result of Titman and Tsyplakov (2007).
2.4 Determinants of Speed of Adjustment
A growing literature has provided specific determinants of SOA from firm-specific factors to macroeconomic factors. The empirical evidence on macroeconomic conditions is consistent, whereas findings into firm characteristics are inconsistent. This section will introduce these determinants of adjustment speed.
2.4.1 Firm-specific factors
Firm size
According to the theory, large firms are expected to have a quicker SOA. They tend to have high profitability, stable cash flows as well as less information asymmetry, resulting in low transaction costs. Further, recapitalization involves substantial fixed costs, which is relatively smaller for large firms. Surprisingly, the empirical results contradict the theory. Flannery and Rangan (2006) group American firms into deciles based on market value of equity and find a negative relation between SOA and firm size. Their explanation is that larger firms tend to use debt which is more expensive to adjust. Elsas and Florysiak (2011) replicate Flannery and Rangan’s (2006) data with different estimation have the same result, but argue that the core reason is the opportunity costs of deviation being low for large firms. However, Drobetz and Wanzenried (2006) do not obtain significant relation between size and SOA in Swiss firms.
Growth opportunity
The effect of growth opportunity on SOA is empirically ambiguous. Growing firms tend to be young and have lots of positive NPV projects. To prevent under-investment problems, they may rely on external sources to fund opportunities (Myers, 1984). Information asymmetry associated with external financing tend to increase recapitalization costs. Therefore, SOA is expected to have a negative relation with growth opportunities, consistent with the findings of Dang et al. (2012). By contrast, Drobetz and Wanzenried (2006) find a positive impact. They argue that high-growth firms frequently visit capital markets, making it easier to change the combination of debt and equity. Even
16
with the costs of asymmetric information, firm value may not decrease because of the positive effect of future growth opportunities.
Distance from the target capital structure
As predicted by dynamic trade-off theory, firms do not adjust when their debt ratios are within the boundaries of the target. Therefore, firms are expected to have higher SOAs when their deviations from the target leverage are larger. This concurs with the empirical estimates (i.e. Drobetz & Wanzenried, 2006; Elsas & Florysiak, 2011).
Byoun (2008) further examines the adjustment asymmetry for U.S firms above and below target leverage, controlling for financial surplus and deficit. He finds that firms make most of the adjustments when they are above target with a financial surplus or below target with a financial deficit. In contrast, Dang et al. (2012) document that UK firms with high SOAs are over-levered with a financial deficit because their adjustments rely heavily on equity.
Default risk
Default risk increases the probability of bankruptcy and financial distress costs. Under dynamic trade-off theory, a firm will rebalance its capital structure when the financial distress costs which are the costs of being at suboptimal leverage ratio exceed transaction costs (Leland, 1994). In addition, financial distress costs have a first-order impact on SOA (Titman & Tsyplakov, 2007). Elsas and Florysiak (2011) denote credit ratings as a measure for default risk, with high credit ratings indicating low default risk and vice versa. Their estimates reveal that SOA increases with default risk. More specifically, the SOA of firms with highest default risk is almost double the speed of average.
2.4.2 Macroeconomic factors
Other than the firm characteristics abovementioned, it has conclusively been shown that economic prospects have a positive effect on SOA. The study of Cook and Tang (2010) on the U.S. market and the study of Drobetz and Wanzenried (2006) on the Swiss market considered the effect of macroeconomic variables on the adjustment speed toward optimal capital structure. This section will explain these variables in detail.
17
Term spread
Term spread, the difference between short-term and long-term interest rates, is generally interpreted to be a predictor of future economic prospects. It is well established that a high term spread can be seen as an strong indicator of favoriable economic prospects (Estrella & Hardouvelis, 1991). Drobetz and Wanzenried (2006) construct the term spread as the difference between the yield on long-term government bonds and the three-month money market interest rate. Instead, Cook and Tang (2010) use three-three-month Treasury-bill interest rate when calculating the term spread. In accordance with the prediction, empirical results show that SOA is faster when term spread is higher (Drobetz & Wanzenried, 2006; Cook & Tang, 2010).
GDP growth rate
Gross Domestic Product (GDP) is a traditional method of business cycle. An increase in GDP growth rate indicates an economic expansion, while a decrease in GDP growth rate shows an economic recession. Cook and Tang (2010) calculate yearly GDP growth rate based on quarterly data and their results suggest that SOA is quicker when GDP growth rate is higher.
Default spread
Default spread, regarded as a proxy for default risk, is the difference between the yield on low-grade and high-grade bonds with same maturity. It can be considered as an indicator of the current economic state. Default spread is low during economic expansion and high during economic recession. Cook and Tang (2010) define default risk as the difference between the average yield of Baa bonds and the average yield of Aaa bonds, with a maturity between 20 and 25 years. Albeit focus on Swiss market, Drobetz and Wanzenried (2006) use default spread of U.S. bonds because they proxy it for global default risk. In line with the predicion, SOA is more readily when default risk is lower (i.e. Drobetz & Wanzenried, 2006; Cook & Tang, 2010).
18 2.5 Hypotheses
In addition to firm characteristics and business cycle, industry effects may also have implications for the capital structure choice of firms.
Istaitieh and Rodriguez‐Fernandez (2006) argue that leverge is inversely related to industry competition.In terms of strategic behavior, unlevered or less-levered firms have healthy capital structure and they can take advantage of the situation by using aggressive behavior to drive out highly levered competitors. In anticipation of this predatory behavior firms in competitive industries tend to have a low debt ratio because financial distress costs are costly if they have high leverage (Brander & Lewis, 1986).
Furthermore, Chhaocharia et al. (2009) studied the effect of competition on agency conflict between shareholders and management. They figure out agency conflict is negatively related to competition among U.S firms. In a competitive industry, the profits of rivals can be used as a benchmark for measuring the manager’s performance. Shareholders know whether or not management makes an effort. Thus, the incentive conflicts are alleviated and debt is not necessarily needed to mitigate free cash flow problem (Jensen, 1986). Based on the arguments above, the following hypothesis is proposed.
Hypothesis 1. Firms in competitive industries have lower debt ratios.
Elsas and Florysiak (2011) report adjustment speed varies across industries. They use dynamic panel fractional estimation to replicate Flannery and Rangan’s (2006) study on U.S firms from 1965 to 2009. Based on Fama and French (1997) industry classifications, they find the average SOA is 25.8% for all industry groups. Textiles industy has the lowest adjsutment speed due to high transaction costs driven by low capital market transactions. However, precious metals industry has the quickest adjustment speed because frequent capital market transactions decreases transaction costs.
Apart from transaction costs, financial distress costs have effect on adjustment speed as well. To gain market share, firms without enough internal funds tend to invest in new projects by debt financing. Firms will not be able to repay debt if the new investment fails, which leads to substantial financial distress costs. Since financial distress costs play
19
a vital role in speed of adjustment compared to transaction costs, speed of adjustment is expected to be faster for firms with high financial distress costs.In competitive industries, firms try to avoid over leverage as they need to maintain a healthy capital structure in case of predatory price war by rivals. Thus, second hypothesis is proposed.
20
3. Methodology
In this section, methodologies employed to test hypotheses will be described. The regression models for target capital structure and adjustment speed toward the target are explained in detail as well as variable definitions.
3.1 Regression Model for Hypothesis 1
Target capital structure is unobservable and can be only estimated. As described in literature, researchers have examined different factors to estimate target debt ratio. To test hypothesis 1 this paper adopts mostly used factors to predict target leverage with fixed effects regression model below:
(1)
where is firm i’s target leverage ratio at time t and is a vector of explanatory variables . These explanatory variables determining the optimal financial structure are those discussed in the previous literature, which are industry competition, profitability, growth opportunity, firm size, asset tangibility, industry median leverage and expected inflation. and are firm-specific and time-invariant factors to control fixed effects (MacKay & Phillips, 2005).
3.1.1 Measurements for dependent variable
Market debt ratio is a common proxy for leverage (e.g., Flannery & Rangan, 2006; Leary & Roberts, 2005; Fama & French, 2002). Following Flannery and Rangan (2006), market debt ratio is defined as
where stands for book value of both short- and long-term debt for firm i at time t, denotes the number of common shares outstanding at time t and is the share price at time t.
Book debt ratio is also used for comparison (Flannery & Rangan, 2006; Frank & Goyal, 2009),
21
where denotes book value of total debt for at time t and equals total assets at time t. 3.1.2 Measurements for independent variables
Industry Competition
Industry competition is hardly to calculate directly and many researchers use Herfindahl-Hirschman Index (HHI) instead to proxy for competition. HHI is a method to calculate industry concentration which is defined as the sum of squared market share of each company:
Market share is calculates by firm i’s sales as a proportion of total sales of firms in the industry. If there are only a few firms within one industry, HHI will be high because this industry is highly concentrated. However, a less concentrated industry where there are hundreds or even thousands of firms has a low HHI. Firms in such an industry are very competitive in order to survive. Therefore, HHI can be used to measure industry competition. According to U.S department of justice, industries in which HHI exceeds 2,500 are highly concentrated, industries in which HHI is between 1,500 and 2,500 are moderately concentrated, and industries where HHI is below 1500 are less concentrated. In this paper, a dummy variable is employed which is set to one if HHI is larger than or equal to 1500 and set to zero if HHI is smaller than 1500. Thus, an industry is highly competitive with dummy equal to zero and less competitive with dummy equal to one. Industry competition is predicted to be negatively related to firm leverage.
Profitability
Profitability is measured by earnings before interest and taxes (EBIT) as a percentage of total assets (Flannery & Rangan, 2006). Interest expenses and income taxes are added back to income before extraordinary items to calculate EBIT first, and then divide EBIT by total assets to get profitability. Profitability is expected to have negative impact on debt ratio.
22
Firm Size
Asset size is a proxy for firm size because larger firms tend to have more assets. Total assets are first transformed to 1983 dollars to avoid impact of inflation and then take natural logarithm (Flannery & Rangan, 2006). Firm size is supposed to have a positive correlation with leverage.
Growth Opportunity
Firms with high growth opportunities usually have high market-to-book ratio of assets because of future expectation. Therefore, market-to-book ratio (MB) is used as a proxy for growth opportunity. Market value of assets is calculated as total debt plus value of preferred stock plus number of common shares outstanding times price per share. Market value of assets divided by book value of assets is market-to-book value. Growth opportunity is expected to have negative effect on debt ratio.
Asset tangibility
Fixed assets can be converted to cash more quickly than intangible assets so fixed assets as a proportion of total assets is used as a proxy for asset tangibility. Fixed assets are usually long-term assets which are property, plant and equipment (PP&E). Therefore, PP&E divided by total assets is asset tangibility. Asset tangibility is expected to have a positive relation to leverage level.
Industry median leverage
Industry median leverage is the median industry market leverage, measured as market debt ratio, based on 2-digit industry classification. Firm’s leverage is expected to be positively related to the industry median leverage.
Expected inflation
Percentage change in Consumer Price Index (CPI) is a common method to proxy for expected inflation (Frank & Goyal, 2009). CPI measures the price level of a basket of consumer goods and services. Thus, the percentage change in CPI estimates the price
23
change from consumers’ perspective, which represents expected inflation. Inflation is predicted to be positively related to debt ratio.
3.2 Regression Model for Hypothesis 2
A dynamic capital structure model is employed based on the partial adjustment model of Flannery and Rangan (2006). The adjustment process toward the optimal leverage ratio depends on the changes in the leverage that is partially absorbed by the difference between the target leverage and lagged leverage :
(2)
Where i i is the observed change of debt ratio, i i is the gap between target debt ratio this year and debt ratio observed last year, and i captures the speed of adjustment toward desired leverage, which relies on the ratio of marginal cost of being off target to marginal cost of adjustment. Intuitively, the higher the adjustment costs relative to the benefits, the slower is the pace of adjustment. If there are no adjustment costs, firms immediately close the gap in the next period indicating i . However, the transaction costs do exist in reality and because of which firms can make only partial adjustment in the next period. Accordingly, it leads to i . Another extreme situation is i , which suggests that firms do not make any movements. The possible reason could be that the cost of being off target is too small and firms are unwilling to change current debt ratio.
As predicted in the second hypothesis, the speed of adjustment is determined by industry competition, which is measured by HHI:
(3)
Substituting (2) and (3) into (1) and taking into consideration of panel data gives
=( )
24
where is time-fixed effect, is firm-fixed effect and is the error term. The primary focus is , which is the coefficient on the interaction term between industry concentration and lagged leverage. As it is predicted that the adjustment speed increases in industry competition, the hypothesis formulated for the regression model is
H0: H1:
The reason is that the dummy variable HHI is one when a industry is highly concentrated which is expected to be less competitive.
Endogenity happens because dependent variable leverage is regressed on its own lagged value. Recent studies adopt instrumental variables to mitigate the bias, which is the correlation between the independent variable and the error term. In addition, Generalized Method of Moments (GMM) estimation is applied to solve endogenity (Arellano & Bond, 1991). GMM takes first differences of variables in equation 4, which removes possible correlations between unobservable firm-related factors and explanatory variables. Empirical research suggests OLS estimation tends to underestimate SOA and fixed effects estimation tends to overestimate SOA. According to Lemmon et al. (2008), the estimates of GMM lie within the boundaries between OLS and fixed effects estimation. This paper employs two-step system GMM estimators. First, target debt ratio is estimated. Different from regression model for hypothesis 1, industry competition is removed from independent variables to avoid multicollinearity in step two. Second, fitted values from first step are used to estimate speed of adjustment. All coefficient estimates are adjusted for robust standard error to control for heteroskedasticity.
25
4. Data and Descriptive Statistics
The firm-level data are retrieved from the annual COMPUSTAT files for all U.S. public firms. The sample period ranges from 1997 to 2016. Following previous studies, utilities (SIC 4900-4999) and financial firms (SIC 6000-6999) are excluded from the sample because their capital structures are highly regulated. The other firms are classified into 9 different industries based on 2-digit SIC. Furthermore, firms with less than two consecutive years of data are removed because regression model includes lagged variables. A few adjustments are made to avoid data errors and extreme outliers. First, observations with missing values for total assets, short-term debt and long-term debt are dropped. Second, variables with asset size smaller than one million dollars are excluded. This is because asset size is the denominator when calculating market-to-book ratio and it causes extreme variables without this step. Third, market and book leverages are dropped if they are larger than 1. Fourth, other variables are winsorized at 1% and 99% percentiles. As a result, the final dataset is an unbalanced panel of 119,418firm-year observations for 13,429 firms in 9 industries between 1997 and 2016.
Table 1. Summary Statistics
The sample includes U.S. publicly traded firms’ data with at least two consecutive years from 1997 to 2016. BDR and MDR are manually winsorized by exluding values larger than 1. Other variables are winsorized at 1% and 99% percentiles to prevent the effect of extreme outliers.
N mean Standard
Deviation min max
BDR 119400 0.2056 0.2119 0 0.8587 MDR 119400 0.2058 0.242 0 0.9325 EBIT_TA 110200 -0.081 0.3962 -2.335 0.383 MB 119400 1.9849 2.3821 0.2088 16.4192 lnTA 119400 19.7429 2.4448 14.2966 28.176 FA_TA 119300 0.2913 0.2695 0 1 Ind_Median 119400 0.1211 0.0833 0 0.4755 Inflation 104200 0.0218 0.0106 -0.0036 0.0384 HHI 119400 0.0194 0.1381 0 1 Definitions
MDR: market debt ratio= (book value of long-term and short-term debt)/ market value of assets
BDR: book debt ratio= (book value of long-term and short-term debt)/ book value of assets
26
MD: growth opportunity= (book value of liabilities + market value of equity)/book value of assets
lnTA: firm size= natural logarithm of total assets measure in 1983 dollars. FA_TA: asset tangibility= tangible assets/total assets
Ind_Median: industry median leverage= median industry market leverage Inflation: expected inflation= percentage change in CPI
HHI: industry concentration dummy variable which is 1 if HHI >1500 and 0 if HHI<1500.
Table 2.Difference of Firm Characteristics
The average values of firm-specific characteristics for competitive industries (HHI=0) are shown in column (1) and the average values of firm-specific characteristics for less competitive industries (HHI=1) are shown in column (2).
HHI=0 HHI=1 (1) (2) BDR 0.2059 0.189 MDR 0.2055 0.2235 EBIT_TA -0.0796 -0.1576 MB 1.9869 1.9256 lnTA 19.7609 18.7862 FA_TA 0.293 0.2065 Ind_Median 0.1216 0.0959
Table 1 provides summary statistics and definitions for all variables used in this paper. The difference of firm-specific characteristics between competitive industries and less competitive industries is shown in Table 2. The mean values of variables for firms in competitive industries are in column 1 (HHI=0) and the mean values of variables for firms in less competitive industries are in column 2 (HHI=1). It is found that firms in competitive industries tend to have higher book debt ratio and lower market debt ratio.
27
5. Results
5.1 Estimation of Target Capital Structure
Table 3. Regression Results of Target Capital Structure
Column (1) shows the regression results of target capital structure with market debt ratio (MDR) as dependent variable. Column (2) shows the regression results of target capital structure with book debt ratio (BDR) as dependent variable. Independent variables include HHI dummy variable, profitability, growth opportunity, firm size, asset tangibility, industry median leverage and expected inflation. Fixed effects are used to control for firm and year invariant factors.
MDR BDR VARIABLES (1) (2) HHI 0.0496*** 0.00292 -0.0114 -0.0106 EBIT_TA -0.0813*** -0.0649*** -0.00182 -0.00163 MB -0.0215*** -0.00333*** -0.000319 -0.000284 lnTA 0.0253*** 0.0198*** -0.000554 -0.000504 FA_TA 0.171*** 0.149*** -0.00398 -0.0036 Ind_Median 0.685*** 0.345*** -0.0106 -0.00956 inflation 0.175*** -0.214*** -0.0445 -0.0395 Constant -0.372*** -0.245*** -0.0115 -0.0104 Observations 96,304 96,304 Number of firms 13,429 13,429
Firm FE Yes Yes
Year FE Yes Yes
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 3 provides regression results of target capital structure with market debt ratio and book debt ratio as dependent variables proxy for leverage in column (1) and (2) respectively.
28
Independent variables applied to predict optimal capital structure are commonly used in other studies, which are described in literature before. In addition, dummy variable HHI is used to study the effect of industry competition on target capital structure. HHI = 1 represents concentrated industry with less competition and HHI=0 represents competitive industry with severe competition. Year and firm fixed effects are included to prevent potential bias.
As shown in table 3 the sign of coefficient estimates are overall consistent with previous studies. Profitability (EBIT_TA) and growth opportunity (MB) are negatively related to optimal leverage. Firm size (LnTA), asset tangibility (FA_TA) and industry medium (Ind_Medium) are all positively correlated to both market and book leverage. However, inflation has opposite signs across two columns. In line with Frank and Goyal (2009), expected inflation has positive and significant impact on market leverage. What differs from Frank and Goyal’s findings is that expected inflation has a negative and significant effect on book leverage while their result is insignificant. Since this study employs same proxy for book leverage as Frank and Goyal (2009), the difference could be time period which is from 1950 to 2003 in their study. They argue that book leverage is backward looking while expected inflation anticipates the firm’s future.
Furthermore, industry concentration has a significantly strong positive effect on market leverage, which is in line with hypothesis 1 that firms in competitive industries have lower debt ratio. More specifically, market leverage of firms in competitive industries is 4.96% less than that of firms in concentrated industries. For book leverage the effect of industry concentration is positive but not significant.
5.2 Estimation of Adjustment Speed
Table 4. Regression Results of Adjustment Speed
Column (1) shows the regression results of adjustment speed with market debt ratio (MDR) as dependent variable. Column (2) shows the regression results of adjustment speed with book debt ratio (BDR) as dependent variable. Independent variables include lagged leverage, interaction term between dummy HHI and lagged leverage, target leverage, interaction term between dummy HHI and target leverage and year dummy variables.
MDR BDR
29 L.LEV 0.747*** 0.759*** -0.0126 -0.0136 HHI#L.LEV 0.202*** 0.292 -0.078 -0.332 LEV* 0.354*** 0.391*** -0.0137 -0.0163 HHI#LEV* -0.176 -0.258 -0.343 -0.318 _Iyear_1998 0.0236*** 0.0125*** -0.00264 -0.00216 _Iyear_2000 0.0111*** -0.00931*** -0.00256 -0.00211 _Iyear_2001 -0.0192*** -0.00928*** -0.00267 -0.00212 _Iyear_2002 -0.0158*** -0.0221*** -0.00269 -0.00205 _Iyear_2003 -0.0517*** -0.0200*** -0.00246 -0.00202 _Iyear_2004 -0.0362*** -0.0216*** -0.00231 -0.00207 _Iyear_2005 -0.0250*** -0.0140*** -0.00248 -0.00216 _Iyear_2006 -0.0274*** -0.0120*** -0.00248 -0.00221 _Iyear_2007 -0.0144*** -0.00969*** -0.00263 -0.00221 _Iyear_2008 0.0317*** -0.0128*** -0.00358 -0.00225 _Iyear_2009 -0.0604*** -0.0362*** -0.00264 -0.00206 _Iyear_2010 -0.0409*** -0.0271*** -0.00251 -0.00206 _Iyear_2011 -0.0177*** -0.0204*** -0.00275 -0.00216 _Iyear_2012 -0.0274*** -0.0145*** -0.00265 -0.0022 _Iyear_2013 -0.0410*** -0.0173*** -0.00255 -0.00216 _Iyear_2014 -0.0242*** -0.0112*** -0.00268 -0.00223 _Iyear_2015 -0.0119*** -0.00876*** -0.00276 -0.00225 _Iyear_2016 -0.0390*** -0.0143***
30 -0.00257 -0.00221 Constant -0.002 -0.0197*** -0.00235 -0.00246 Observations 96,186 96,186 Number of firms 13,429 13,429
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
After estimating optimal capital structure, the speed at which firms adjust their actual leverage toward the target can be measured. Table 4 shows the regression results of adjustment speed toward target capital structure. Market debt ratio and book debt ratio are used as proxies for leverage in columns (1) and (2) respectively.
GMM is used to estimate dynamic panel regression model, which allows simultaneous estimation of coefficients (Arellano & Bond, 1991). For target capital structure, fitted values from estimating target leverage in equation 1 are used. Therefore, the estimation has two steps with estimating target debt ratio first and then speed of adjustment. Year dummy variables are included in the equations to control year fixed effects. Year 1999 is automatically omitted to avoid multicollinearity when running code in STATA.
The main interest focuses on the estimate for , which is the coefficient on the interaction term between lagged leverage and industry concentration. Note that in equation 4, there is a negative sign on and the coefficient on lagged leverage is reported as . Therefore, the estimated coefficients are interpreted accordingly.
Table 4 summarizes the effect of industry concentration on the speed of adjustment. First, the statistically positive coefficients on lagged leverage show that firms do dynamically adjust their debt ratio toward the target. Consistent with prior research, the results show rapid adjustment speed but slower than Flannery and Rangan’s (2006) result (34.4% per year). The difference arises from the determinants of target leverage and regression methodology used in this paper. Second, the coefficient on the interaction term between industry concentration and lagged market leverage is statistically positive. According to the model, it reveals that the speed of adjustment is 20.3% slower (-0.203) for firms in
31
concentrated industries. More specifically, the annual adjustment speed is 25.3% (1-0.747) for firms in competitive industries and 5% (1-0.747-0.203) for firms in concentrated industries. This result shows that the speed of adjustment varies a lot in terms of industry competition. Therefore, this result supports the hypothesis 2 that firms in competitive industries adjust more quickly toward optimal leverage than firms in concentrated industries. However, the impact of industry concentration on adjustment speed in terms of book leverage is not reliable.
32
6. Robustness checks
In this section, robustness checks are done for different samples. The results are robust to alternative sample periods.
Table 5. Regression Results of Target Capital Structure for Different Samples
Panel A depicts sample period from 1997 to 2008 and Panel B depicts sample period from 2009 to 2016. Column (1) shows the regression results of target capital structure with market debt ratio (MDR) as dependent variable. Column (2) shows the regression results of target capital structure with book debt ratio (BDR) as dependent variable. Independent variables include HHI dummy variable, profitability, growth opportunity, firm size, asset tangibility, industry median leverage and expected inflation. Fixed effects are used to control for firm and year invariant factors.
Panel A. Sample Period 1997-2008
MDR BDR VARIABLES (1) (2) HHI 0.0515*** 0.0115 -0.0128 -0.0116 EBIT_TA -0.0756*** -0.0639*** -0.00235 -0.00203 MB -0.0221*** -0.00462*** -0.000399 -0.000344 lnTA 0.0219*** 0.0181*** -0.000751 -0.000666 FA_TA 0.188*** 0.178*** -0.00535 -0.0047 Ind_Median 0.668*** 0.264*** -0.0122 -0.0106 inflation 0.610*** -0.562*** -0.0886 -0.0762 Constant -0.310*** -0.192*** -0.0148 -0.0131
Panel B. Sample Period 1997-2008
MDR BDR VARIABLES (1) (2) HHI 0.0737*** 0.00779 -0.0156 -0.0151 EBIT_TA -0.0810*** -0.0701*** -0.00268 -0.00246 MB -0.0205*** -0.00272***
33 -0.000517 -0.000475 lnTA 0.0275*** 0.0274*** -0.00084 -0.000798 FA_TA 0.144*** 0.102*** -0.0054 -0.00506 Ind_Median 0.756*** 0.501*** -0.021 -0.0197 inflation -0.0983* -0.307*** -0.0592 -0.0539 Constant -0.445*** -0.425*** -0.017 -0.0162
Firm FE Yes Yes
Year FE Yes Yes
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 6. Regression Results of Adjustment Speed for Different Samples
Panel A depicts sample period from 1997 to 2008 and Panel B depicts sample period from 2009 to 2016. Column (1) shows the regression results of adjustment speed with market debt ratio (MDR) as dependent variable. Column (2) shows the regression results of adjustment speed with book debt ratio (BDR) as dependent variable. Independent variables include lagged leverage, interaction term between dummy HHI and lagged leverage, target leverage, interaction term between dummy HHI and target leverage and year dummy variables.
Panel A. Sample Period 1997-2008
MDR BDR VARIABLES (1) (2) L.LEV 0.654*** 0.742*** -0.0347 -0.0235 HHI#L.LEV 0.395*** -0.396 -0.034 -0.536 LEV* 0.495*** 0.414*** -0.0347 -0.0258 HHI#LEV* -0.894*** -0.742*** -0.034 -0.054 _Iyear_1998 0.0384*** 0.0215*** -0.00335 -0.00223 _Iyear_1999 0.0207*** 0.00942***
34 -0.00277 -0.0021 _Iyear_2000 0.0281*** 0.0103*** -0.0028 -0.00212 _Iyear_2002 0.000236 -0.0136*** -0.00249 -0.00207 _Iyear_2003 -0.0344*** -0.0107*** -0.00266 -0.00202 _Iyear_2004 -0.0159*** -0.0124*** -0.00262 -0.00209 _Iyear_2005 -0.00632** -0.00527** -0.00278 -0.00224 _Iyear_2006 -0.00903*** -0.0028 -0.00293 -0.00226 _Iyear_2007 0.00422 -0.000881 -0.00303 -0.0023 _Iyear_2008 0.0542*** -0.00262 -0.00451 -0.0024 Constant -0.0321*** -0.0301*** -0.00339 -0.00296
Panel B. Sample Period 2009-2016
MDR BDR VARIABLES (1) (2) L.LEV 0.790*** 0.645*** -0.0343 -0.0432 HHI#L.LEV 0.436*** -0.429 -0.024 -0.632 LEV* 0.297*** 0.511*** -0.0333 -0.0494 HHI#LEV* -0.816*** -0.663*** -0.022 -0.021 _Iyear_2010 -0.00209 -0.0166*** -0.00246 -0.00363 _Iyear_2011 0.0225*** -0.00841*** -0.00264 -0.00297 _Iyear_2012 0.0145*** 4.35E-05 -0.00216 -0.00237 _Iyear_2014 0.0183*** 0.00590*** -0.0021 -0.002
35 _Iyear_2015 0.0301*** 0.0104*** -0.00236 -0.00247 _Iyear_2016 0.00126 0.0043 -0.0023 -0.00263 Constant -0.0357*** -0.0406*** -0.00284 -0.00607
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
The full sample is separated into two periods which are 1997-2008 and 2009-2016. The underlying intuition is that the adjustment speed may overall slow down after 2008 financial crisis due to the expensive adjustment costs. As shown in Table 5, the results of target capital structure are robust to shorter sample periods. The findings in Table 6 show that after financial crisis firms decrease the speed at which they move to target capital structure. The speed is 34.6% (1-0.654) before crisis and decreases to 21% (1-0.79) after crisis.
36
7. Conclusion
This section includes the conclusion of the paper, limitations and suggestions for future studies.
The primary purpose of this paper is to study the impact of industry competition on dynamic capital structure. More specifically, this paper investigates how competition affects target debt ratio and adjustment process toward the target leverage, which constitutes two hypotheses:
1. Firms in competitive industries have lower debt ratios.
2. Firms in competitive industries have a faster speed of adjustment towards the optimal leverage.
The sample consists of 13,429 U.S. listed firms across 9 industries from 1997 to 2016. All variables used are calculated based on fiscal year data from COMPUSTAT. Market leverage and book leverage are regarded as two proxies for debt ratio. A dummy for Herfindahl-Hirschman Index (HHI) is used as a measure for industry competition.
With fixed effects regression model, the findings show firms involved an intense competition indeed have lower target leverage. However, the results are significant for market debt ratio only. Although not primary interests, other determinants of target debt ratio are tested as well. These determinants include profitability, growth opportunity, firm size, asset tangibility, industry median leverage and expected inflation. The coefficients are all statistically significant and their effects are generally in line with precious research. Partial adjustment regression model is employed to test the adjustment speed. The results show that speed of adjustment toward the target leverage is more rapidly for firms in competitive industries, which is consistent with the hypothesis. Again, the results are significant for market leverage only.
The main contribution of this paper is that industry competition shows a positive relation with speed of adjustment. Recent evidence only points out that adjustment speed varies across industries, but no further research on industry effects is conducted. This paper delves into the industry factor and attempts to fill in the gap.
37
This paper has several limitations. First, HHI is not a perfect proxy for industry competition because it suffers a downward bias. Several firms, although run separately, could be owned by one parent company. However, the subsidiary firms are treated as independent firms when calculating market share and industry concentration is therefore underestimated. Second, GMM estimation is a biased approach still. Flsas and Florysiak (2011) argue that GMM estimator does not take the fractional nature of leverage into consideration.
Based on limitations abovementioned, a few suggestions for future studies are made. First, an independent and unbiased indicator for industry competition would be a better solution. Second, a better model or estimation method would be good to estimate more accurate results. Furthermore, other industry factors could be further investigated whether they have impact on dynamic capital structure.
38
8. References
Ang, J. S., Chua, J. H., & McConnell, J. J. (1982). The administrative costs of corporate bankruptcy: A note. The Journal of Finance , 37 (1), 219-226.
Antoniou, A., Guney, Y., & Paudyal, K. (2002). Determinants of corporate capital structure: Evidence from European countries. University of Durham: Working Paper. Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. The review of economic studies , 58 (2), 277-297.
Baker, M., & Wurgler, J. (2002). Market timing and capital structure. The Journal of Finance , 57 (1), 1-32.
Banjeree, S., Heshmat, A., & Wihlborg, C. (2004). The dynamics of capital structure. Research in Banking and Finance , 4, 275-297.
Bradley, M., Jarrell, G. A., & Kim, H. E. (1984). On the existence of an optimal capital structure: Theory and evidence. The Journal of Finance , 39 (3), 857-878.
Brander, J. A., & Lewis, T. R. (1986). Oligopoly and financial structure: The limited liability effect. The American Economic Review , 956-970.
Byoun, S. (2008). How and when do firms adjust their capital structures toward targets? The Journal of Finance , 63 (6), 3069-3096.
Chhaochharia, V., Grinstein, Y., Grullon, G., & Michaely, R. (2009). Product market competition and agency conflicts: Evidence from the Sarbanes Oxley Law. Johnson School Research Paper Series , 18-2012.
Cook, D. O., & Tang, T. (2010). Macroeconomic conditions and capital structure adjustment speed. Journal of Corporate Finance , 16 (1), 73-87.
Dang, V. A., Kim, M., & Shin, Y. (2012). Asymmetric capital structure adjustments: New evidence from dynamic panel threshold models. Journal of Empirical Finance , 19 (4), 465-482.
De Andres Alonso, P., Iturriaga, F., & Sanz, J. (2005). Financial decisions and growth opportunities: a Spanish firm's panel data analysis. Applied Financial Economics , 15, 391-407.
Drobetz, W., & Wanzenried, G. (2006). What determines the speed of adjustment to the target capital structure? Applied Financial Economics , 16, 941-958.
39
Elsas, R., & Florysiak, D. (2011). Heterogeneity in the speed of adjustment toward target leverage. International Review of Finance , 11 (2), 181-211.
Estrella, A., & Hardouvelis, G. A. (1991). The term structure as a predictor of real economic activity. The Journal of Finance , 46 (2), 555-576.
Fama, E. F., & French, K. R. (1997). Industry costs of equity. Journal of Financial Economics , 43 (2), 153-193.
Fama, E. F., & French, K. R. (2002). Testing trade-off and pecking order predictions about dividends and debt. The Review of Financial Studies , 15 (1), 1-33.
Faulkender, M., Flannery, M. J., Hankins, K. W., & Smith, J. M. (2012). Cash flows and leverage adjustments. Journal of Financial Economics , 103 (3), 632-646.
Fischer, E. O., Heinkel, R., & Zechner, J. (1989). Dynamic capital structure choice: Theory and tests. The Journal of Finance , 44 (1), 19-40.
Flannery, M. J., & Rangan, K. P. (2006). Partial adjustment toward target capital structures. Journal of Financial Economics , 79 (3), 469-506.
Frank, M. Z., & Goyal, V. K. (2009). Capital structure decisions: which factors are reliably important? Financial Management , 38 (1), 1-37.
Frank, M. Z., & Goyal, V. K. (2007). Trade-off and pecking order theories of debt. Handbook of Empirical Corporate Finance , 2, 135-202.
Graham, J. R., & Harvey, C. R. (2001). The theory and practice of corporate finance: Evidence from the field. Journal of Financial Economics , 60 (2-3), 187-243.
Graham, J. R., & Leary, M. T. (2011). A review of empirical capital structure research and directions for the future. Annual Review of Financial Economics , 3 (1), 309-345. Harris, M., & Raviv, A. (1991). The theory of capital structure. The Journal of Finance , 46 (1), 297-355.
Hovakimian, A., Opler, T., & Sheridan, T. (2001). The debt-equity choice. Journal of Financial and Quantitative Analysis , 36 (1), 1-24.
Istaitieh, A., & Rodriguez‐Fernandez, J. M. (2006). Fac or‐produc marke s and firm's capital structure: A literature review. Review of Financial Economics , 15 (1), 49-75. Jensen, M. C. (1986). Agency costs of free cash flow, corporate finance, and takeovers. The American Economic Review , 76 (2), 323-329.
40
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics , 3 (4), 305-360. Kraus, A., & Litzenberger, R. H. (1973). A sta e‐preference model of op imal financial leverage. The Journal of Finance , 28 (4), 911-922.
Leary, M. T., & Roberts, M. R. (2005). Do firms rebalance their capital structures? The Journal of Finance , 60 (6), 2575-2619.
Leland, H. E. (1994). Corporate debt value, bond covenants, and optimal capital structure. The Journal of Finance , 49 (4), 1213-1252.
Lemmon, M. L., Roberts, M. R., & Zender, J. (2008). Back to the beginning: persistence and he cross‐sec ion of corpora e capi al s ruc ure. The Journal of Finance , 63 (4), 1575-1608.
Lucas, D. J., & McDonald, R. L. (1990). Equity issues and stock price dynamics. The Journal of Finance , 45 (4), 1019-1043.
MacKay, P., & Phillips, G. M. (2005). How does industry affect firm financial structure? The Review of Financial Studies , 18 (4), 1433-1466.
Maksimovic, V., & Zechner, J. (1991). Debt, agency costs, and industry equilibruim. The Journal of Finance , 46 (5), 1619-1643.
Modigliani, F., & Miller, M. H. (1958). The cost of capital, corporation finance and the theory of investment. The American Economic Review , 48 (3), 261-297.
Morellec, E., Nikolov, B., & Schurhoff, N. (2012). Corporate governance and capital structure dynamics. The Journal of Finance , 67 (3), 803-848.
Myers, S. C. (1984). The capital structure puzzle. The Journal of fFinance , 39 (3), 574-592.
Myers, S. C., & Majluf, N. S. (1984). Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics , 13 (2), 187-221.
Rajan, R. G., & Zingales, L. (1995). What do we know about capital structure? Some evidence from international data. The Journal of Finance , 50 (5), 1421-1460.
Ritter, J. R., & Warr, R. S. (2002). The decline of inflation and the bull market of 1982– 1999. Journal of Financial and Quantitative Analysis , 37 (1), 29-61.
