MSc Business Economics, Finance track
Master Thesis
Why zero leverage? The effect of managerial characteristics on
zero-leverage firms
Liu, Chang
10604278
Supervisor: Dr. P.J.P.M. (Philippe) Versijp
ACKNOWLEDGEMENTS
First and foremost, I would like to express my sincere thanks to Dr. P.J.P.M. (Philippe) Versijp. As my supervisor, he has provided me with great patience and valuable guidance in the procedure of writing my thesis. I am grateful for his kind help and encouragement. I shall extend my gratitude to all the teachers who have helped me improve my proposal and fundamental of my thesis.
Secondly, I would also like to thank my parents and friends for their supports and warm help.
ABSTRACT
In this paper, I investigate the effect of managerial characteristics on zero-leverage firms facing distinct financial constraints. On average, 21% of U.S. public non-financial firms are unlevered over the period of 1996-2012. Based on their different capacities to obtain financing, zero-leverage firms are further classified into financially constrained and financially unconstrained zero-leverage firms, similar to the category in the study of Bessler et al. (2012). Using multivariate logit regression analysis, the results are consistent with the findings in the existing studies, as well as the financial constraint hypothesis. In my results, firms with young and female CEOs prefer to maintain zero leverage. Zero-leverage firms always grant more options and stock ownership, but less other compensation to their CEOs. Moreover, firms with CEOs or executive directors of long tenure end up unlevered. Nevertheless, male executive directors with implicit ethnicity have the preference of zero leverage. In conclusion, the results suggest that managerial characteristics indeed have more significant effects on financially unconstrained zero-leverage firms than constrained firms.
CONTENT
1. Introduction ... 1
2. Literature review ... 3
2.1. Zero-leverage phenomenon and plausible explanations ... 3
2.2. Managerial characteristics and leverage ... 6
2.3. Financial constraints and leverage ... 7
3. Methodology and hypothesis ... 9
3.1. Plausible explanations for zero-leverage phenomenon ... 9
3.1.1. Financial constraint hypothesis ... 9
3.1.2. Managerial characteristics ... 10
3.2. Hypothesis development ... 11
3.3. Multivariate logit regression ... 11
4. Data and descriptive statistics ... 13
4.1. Data, sample restriction and zero-leverage firm definition ... 13
4.2. Zero-leverage firms and managerial characteristics of CEOs ... 14
4.2.1. Variable construction ... 14
4.2.2. Zero-leverage firms selection and classification ... 15
4.2.3. Control group selection ... 17
4.2.4. Descriptive statistics ... 19
4.3. Zero-leverage firms and managerial characteristics of executive directors ... 22
5. Results ... 23
5.1. Managerial characteristics of CEOs and zero-leverage firms ... 23
5.2. Managerial characteristics of executive directors and zero-leverage firms ... 26
6. Robustness check and additional tests ... 31
7. Conclusion ... 33
References ... 37
Appendix A ... 39
1. Introduction
Relating to “low leverage puzzle”, zero-leverage phenomenon is one of the puzzling and unsolved topics in the corporate finance. Recently, stylized fact shows that firms choose extremely less or even zero leverage than predicted in the Pecking order theory and Trade-off theory. According to the market research on the CFO.com site, up to 126 S&P U.S. public companies maintain zero debt and more than $500 million revenue in 2012, such as Apple, MasterCard and Red Hat.
Zero-leverage firms usually have sufficient cash flows, as well as growing revenue. They pay large dividends instead of interest expenses. Some of these firms are large, profitable and pay out high dividends while other unlevered firms are small, young and less profitable (Bessler et al., 2012). Because they haven’t yet developed their reputation in the debt market, these firms may face the financial constraints to obtain debt. These features are consistent with the category of zero-leverage firms by Bessler et al. (2012). They classify the unlevered firms into financially constraint and unconstraint firms by debt capacities.
Recently, Strebulaev and Yang (2013) and Devoset al. (2012) study the mystery of zero-leverage phenomenon but have puzzling results. Firms usually prefer to keep leverage because of the agency benefits and tax shields. Managers can also be motivated to work more efficiently with less diluted ownership. Strebulaev and Yang (2013) find that the CEO of unlevered firms always has higher ownership and longer tenure. Nevertheless, Devos et al. (2012) reject the effect of management entrenchment and find these firms are mainly financially constrained. The effect of
managerial characteristics and financial constraints may be the potential explanations for these discrepancies. Firms could end up unlevered if managers benefit themselves or have personal preferences to use debt conservatively. If firms have varying levels of financial constraints to obtain external funds, managerial characteristics may have distinguished effects on these firms.
In order to better understand the zero-leverage phenomenon, this paper contributes to the existing studies by investigating not only the potential effects of managerial characteristics, but also distinct levels of financial constraints faced by zero-leverage firms. From the empirical perspective, studying the potential effects of managerial characteristics may help firms with different abilities of obtaining external financing to make proper capital structure and leverage choices. Capturing all of these, the research question of this paper is “How do managerial characteristics affect zero-leverage firms facing distinct financial constraints?” Accordingly, to explore the research question, the hypothesis of this paper is developed as “Managerial characteristics have more significant effects on zero-leverage firms facing few or no financial constraints”.
According to the distinct features of unlevered firms, these firms are classified as financially constrained and unconstrained unlevered firms. Moreover, several levered firms are chosen as a control group for each unlevered firm based on the benchmarks (industry, firm size and performance). Using the panel data from COMPUSTAT, Execucomp over the period of 1996-2012 and RiskMetrics data set from 2007 to 2012, multivariate logit regression analysis is conducted between zero-leverage firms and
their control groups. By comparing effects on each type of zero-leverage firms, we can examine the hypothesis and draw conclusions.
The outline of this paper is organized as follows. Section 2 provides some related existing literature and studies on this research topic. Section 3 presents the methodology and the hypothesis development. Section 4 presents the data, sample construction, control group selection and summary of descriptive statistics for key variables. Section 5 provides the results of the multivariate logit regressions analysis of managerial characteristics’ effects on zero-leverage firms. Section 6 provides the robustness check and additional tests. Section 7 concludes and discusses.
2. Literature review
Nowadays substantially large numbers of firms choose low leverage or even zero leverage. Although there are some tax benefits of debt financing, these firms maintain less debt than predicted in the Pecking order theory and Trade-off theory. This research is aimed to investigate the effects of managerial characteristics on zero-leverage firms facing distinct financial constraints. The related empirical evidence and literature can be categorized into the following three aspects.
2.1. Zero-leverage phenomenon and plausible explanations
Many empirical studies focus on the plausible reasons for the low leverage puzzle, but the empirical evidence for related zero-leverage phenomenon is scarce. In the studies of Strebulaev and Yang (2013), Minton and Wruck (2011), they conclude
that industry factor is not the reason why firms keep low leverage. Besides, existing literature has investigated other plausible explanations, such as managerial entrenchment, financial constraint and financial flexibility etc.
Empirical evidence shows that managerial preference of CEOs and directors may be an important explanation for zero-leverage phenomenon. Fama (1980) finds that entrenched managers may choose low leverage in order to reduce the specific firm risk or protect their own capital. Jensen (1986) also argues that entrenched managers may follow low leverage policy to keep away from the associated disciplinary pressure. Managers may eschew debt for their personal interest or better corporate governance. Strebulaev and Yang (2013) study the zero-leverage phenomenon and find that the preference of managers may be a plausible reason. Their results show that especially with small and less independent boards, unlevered firms always grant their CEOs with longer tenure and larger stock ownership.
Whereas, some scholars deny that the managerial preferences and entrenchment are not major explanations for unlevered firms. They try to explore plausible reasons from the financial constraint aspect. The financial constraint hypothesis is regarded as a plausible explanation for zero-leverage phenomenon. When capital market is imperfect, it’s not easy for some firms to gain external debt because of the asymmetric information. Hence in this situation, both demand side and supply side of capital are determinants of capital structure (Dang, 2013).
Moreover, financial flexibility can also be an important potential determinant of zero-leverage policy as shown in the existing research. According to the financial
flexibility hypothesis, in the presence of market frictions, firms will maintain low leverage for saving leverage capacities to invest in the future (DeAngelo and DeAngelo2007). This hypothesis implies that zero leverage may be the reason for firms to save financial flexibilities (Dang 2013).
As DeJong, Verbeek and Verwijmeren (2012) find in their recent study, financial flexibility can be an important explanation for low leverage puzzle, as well as zero leverage phenomenon. Meanwhile, Mura and Marchica (2010) demonstrate that financial flexibility as an untapped borrowing power may be the most critical factor of CFO’s leverage decision. Similarly, in the study of Singh and Hodder (2000), financial flexibility can be one of the important determinants of capital structure, especially when firms have low leverage.
Based on the mixed results in the existing studies on the plausible reasons for zero-leverage phenomenon, the main contribution of this paper is to find out how managerial characteristics of different financially-constrained zero-leverage firms affect their zero-leverage choices. Also according to the classification of unlevered firms in the existing literature, this study aims to use the comparative study to investigate the managerial characteristics’ effects on financially constrained and unconstrained unlevered firms. Considering the significant effect of financial constraints concluded in the existing studies, this paper assumes that managerial characteristics may affect more significantly on the financially unconstrained unlevered firms.
2.2. Managerial characteristics and leverage
Extensive empirical evidence shows that managerial characteristics can be important determinants for capital structures and leverage choices. These studies provide a solid empirical foundation for my investigation of the causal relation between managerial characteristics and zero leverage.
Some scholars find that CEOs’ personal behavior may affect the financing behavior of the companies they manage. Cronqvist et al. (2012) show that CEOs with less personal preferences of debt lead firms they manage to less leverage. Parsons and Titman (2008) also find that executive’s preference may affect the firms’ capital structure, as well as leverage choices. Besides, Bertrand andSchoar (2003); DeYoung, Peng and Yan (2010) study the CEO’s age and preference of leverage but have opposite results. Bertrand and Schoar (2003) conclude that older CEOs tend to take low leverage while DeYoung, Peng and Yan (2010) find oppositely.
Executive compensation may also be a key explanation for managers’ financing incentives and decisions. Lewellen (2006) finds that debt financing can be costly for executives’ stock-based compensation. He also shows that the volatility cost of debt increases with higher option ownership. This may indicate that CEOs with higher option ownership tend to choose less debt, which can also be an indication of the effect of managerial characteristics on zero-leverage firms. Coles, Daniel and Naveen (2005) show that managerial compensation is strongly causal related to firm’s debt financing. Duru, Lyengar and Zampelli (2012) also examine the causal relation between executive’s compensation and leverage choice. In the results, firms with
fewer debt contracts are more likely to pay higher executive compensation, such as stock options. Moreover, Bruslerie and Latrous (2012) conclude empirical results on French firms that the relation between shareholder’s ownership and debt level tends to be an inverted U-shape.
The related existing studies indicate that managerial features indeed have significant effects on firms’ leverage choices. These effects build a good foundation for this research topic. However, as Frank and Goyal (2009) conclude, there are still some important managerial characteristics to be explored. Hence the contribution of this paper is to extend the effect of managerial characteristics on zero-leverage firms and find more plausible untapped characteristics.
2.3. Financial constraints and leverage
Financial constraint, as one of the plausible explanations for zero-leverage phenomenon, has been supported by many scholars in the existing research. Devos et al. (2012) reject the managerial entrenchment hypothesis and argue that unlevered firms are all financially constrained. Bessler, Drobetz, Haller and Meier (2012) also find that, except for few firms maintaining unlevered deliberately, most of firms keep unlevered because of their constrained capacities.
Based on these features, they divide these unlevered firms into financially constraint and unconstraint firms. Similarly, Dang (2013) divides the zero-leverage firms with distinct financial constraints into two groups according to their dividend payments. Dang (2013) also argues that these two distinct groups have different
incentives to avoid debt and can be the main consequences of lacking external funds. These divisions provide important supports for my comparative study of zero-leverage firms facing distinct constraints.
Bessler et al. (2012) find that most of these unlevered firms are financially constrained by their abilities to obtain external debt. Nevertheless, Ferrando and Mulier (2013) find that firms with high profitability and lower leverage tend to face less financial constraints. By distinguishing the perceived and actual financial constraints, their results show that low-leverage firms are more likely to be financially unconstrained.
Kasseeah (2008) studies the financial decisions of UK and Chinese firms with different financial constraints. Unlike the large listed firms, small and medium-sized enterprises (SMEs) are found to maintain low leverage because of their financial constraints. Moreover, the results for both UK and Chinese firms suggest that most of firms prefer external debt and remain leverage. However, when firms gain increased experience in the external funding, they tend to choose low leverage. Similarly, Forbes (2007) assesses the relation between the increased financial constraints for distinct-sized public firms and Chilean capital controls. His results suggest that financial constraint decreases with firm size. This also indicates that smaller sized and less levered firms tend to be financially constrained.
Based on the results and features of firms with distinct financial constraints in the existing studies, there is indeed a significant relation between financial constraints and leverage. Hence contributing to the existing research, this paper is intended to
examine the effects of the managerial characteristics on zero-leverage firms by also taking the financial constraints into consideration. By comparing the effects of managerial characteristics on zero-leverage firms facing distinct constraints separately, it would be better to understand why firms maintain zero leverage.
3. Methodology and hypothesis
Several plausible reasons for firms maintaining zero leverage are briefly reviewed in this section. According to these explanations, the main hypothesis of this paper is developed. In order to test the hypothesis, this section also presents the multivariate logit regression analysis and type of data used.
3.1. Plausible explanations for zero-leverage phenomenon
As many scholars have found in their studies, there are many plausible explanations for zero-leverage phenomenon. The financial constraint hypothesis is one of the potential explanations. Some managerial characteristics of Chief Executive Officers (CEOs), managers and executive directors can also be plausibly determinants to zero-leverage firms. Hence this study is mainly focused on the potential effects of managerial characteristics on zero-leverage firms facing distinct financial constraints.
3.1.1. Financial constraint hypothesis
Both the demand side and supply side of capital are determinants of capital structure in imperfect markets (Dang, 2013). From the supply side, firms need to raise external fund from capital markets. When under the asymmetric information, it’s not
easy for lenders to evaluate the qualities of undistinguished firms and the qualities of their investments (Stiglitz and Weiss, 1981). Because of the credit rationing, it would be difficult for firms to raise enough debt from imperfect capital markets. These firms are always regarded as financially constrained firms. Because of their undistinguished qualities, it’s also difficult for them to borrow from banks or issue corporate bonds. Similarly, lacking of mature reputations in debt markets, these firms may not easily get direct debt.
Conversely, some firms may face less financial constraints. It’s more flexible for them to obtain both external and internal funding. These firms are regarded as financial unconstrained firms. Hence these firms are more likely to be unlevered than their constrained peers. Based on the financial constraint hypothesis, zero-leverage firms are further classified into financially constrained and unconstrained unlevered firms in this paper, using the Size-age index (Hadlock and Pierce, 2010).
3.1.2. Managerial characteristics
Many empirical studies show that managerial characteristics of managers can be one of important explanations for zero-leverage phenomenon. Some personal characters of managers, CEOs and executive directors are related to firms’ capital structures and leverage choices. CEO’s personal behavior and personal preference of leverage may affect company’s capital structure and financing behavior (Bertrand and Schoar, 2003; DeYoung, Peng and Yan, 2010). Moreover, Executive’s compensation may also have significant effect on leverage choice and capital structure (Duru, Lyengar and Zampelli, 2012). Besides, there are still other types of managerial
characteristics to be explored. Hence to investigate the influences of managerial characteristics on zero-leverage firms, executive’s age, gender, tenure, compensations, stock ownership and options granted are chosen as key variables.
3.2. Hypothesis development
According to the financial constraint hypothesis and plausible effects of managerial characteristics, this study is aimed to investigate the potential effects of some representative characteristics of executives on zero-leverage firms. Specially, considering financial constraints, zero-leverage firms are classified into financially constrained and unconstrained unlevered firms. By comparative studies on zero-leverage firms facing distinct financial constraints, whether the managerial characteristics have more influences on financially unconstrained zero-leverage firms can be tested. Studying this help managers make better leverage choices based on varying levels of financial constraints of the firms they manage.
Based on the plausible explanations and contributions of this study, the hypothesis is developed as follows. Managerial characteristics have more significant effects on financially unconstrained unlevered firms than their constrained peers.
3.3. Multivariate logit regression
To investigate the potential effect of managerial characteristics on zero-leverage firms facing distinct levels of financial constraints, this study is conducted on U.S. non-financial firms by choosing the annually panel data from the
COMPUSTAT-North American, Execucomp from 1996 to 2012 and RiskMetrics-director dataset over the period of 2007-2012. After the standard sample selection, independent variables of managerial characteristics are further constructed. Moreover, control groups of zero-leverage firms, financially constrained and financially unconstrained zero-leverage firms are chosen separately using three benchmarks (industry, firm size and performance).
In this study, multivariate logit regression analysis is chosen to test the hypothesis. To examine the significance of each managerial characteristic, logit regression is conducted for each variable. Then use multivariate logit regressions to test the significance of the effects of all characteristics in the model. The main forms of logit regression (1) and multivariate logit regression model (2) are as follows.
(1)
(2)
In these two models, the dependent variable yi, also as a dummy variable, equals to 1 if the firm observation is zero-leverage firm. Independent variables Xi is managerial characteristics variables, such as gender, age, tenure, compensations, stock
ownership, options granted, ethnicity and number of shares owned. βi is the coefficient for each variable Xi while α is the constant. When conducting regressions, other firm-specific accounting variables, and industry are chosen as control variables.
exp( ) ( 1 | ) 1 exp( ) ( ) ln( ) 1 exp( 1 1 ... ) ( 1| 1,..., ) 1 exp( 1 1 ... ) ( ) ln( ) 1 1 ... 1 i x i i i P P y xi x i i i p Logit p i i ix i p x k kx P P yi x xk x k kx P Logit P x k kx P α β α β α β µ α β β α β β β β ε + = = = + + = = + + − + + + = = = + + + + = =∂+ + + + −
Each regression is conducted between zero-leverage firms, financially constrained, unconstrained firms and their control groups. Control groups are chosen by constructing each firm-year observation a control group with similar industry, firm size and performance. Different from other studies, control groups in this multivariate logit model are levered firms.
Using the multivariate logit regression analysis, compare the results of regressions on the subsamples of financially constrained and unconstrained unlevered firms. The hypothesis holds if the managerial characteristics indeed have a more significant effect on financially unconstrained zero-leverage firms. This also indicates that there are more significant coefficients for financially unconstrained zero-leverage firms than constrained firms if the hypothesis holds for true.
4. Data and descriptive statistics
4.1. Data, sample restriction and zero-leverage firm definition
To study the potential effects of managerial characteristics on zero-leverage firms, annually panel data of U.S. non-financial public firms are selected from the COMPUSTAT-North America, Execucomp and RiskMetrics-Director data set. In order to investigate the effect of managerial characteristics of Chief Executive Officers (CEOs), the main sample is constructed by merging annual COMPUSTAT and Execucomp data set over the period of 1996-2012. Except for CEOs, executive directors are also in charge of management. To study managerial characteristics of these executive directors, the subsample is constructed by merging COMPUSTAT and
RiskMetrics data set over the period of 2007-2012.The time period is selected because of the limited time period of Risk Metrics-directors data set.
Then the two samples are restricted by excluding the financial companies (SIC code between 6000 and 6999), non-US firms (FIC code is not USA), as well as utilities (SIC code between 4900 and 4999). The non-publicly traded companies (STKO=1 or 2) are also excluded. The samples are further restricted by dropping the firm years with missing values on total assets, stock price (PRCC_F), long-term and short-term debt. Specially, the sample of directors is further selected by keeping firm-year observations with the employment title of “Executive”. After this selection, the sample of directors is only constructed by executive directors.
There are many definitions for leverage and the zero-leverage firm in the existing studies. In this paper, a firm-year observation is defined as zero-leverage firm if both short-term debt (DLC=0) and long-term debt (DLTT=0) equal to zero. This definition is simple, but most commonly used in the empirical studies.
4.2. Zero-leverage firms and managerial characteristics of CEOs
To examine how managerial characteristics of the CEOs affect zero-leverage firms, it starts with the main sample of merged COMPUSTAT and Execucomp data set over the period of 1996-2012. There are about 35743 observations in the sample.
4.2.1. Variable construction
After the data restriction, the CEOs’ sample is constructed with 86036 observations. Then some key firm-specific variables are constructed, such as firm-age
(number of years since the IPOdate), cash holdings, dividend payout ratio, Tobin’s Q, tangibility, profitability, EPS, ROE, ROA, and growth opportunities. Specially, firm size is the logarithm of inflation-adjusted total assets. Using the annual U.S.
Consumer Price Index (CPI) by “freduse, CPIAUCSL” command contained in the STATA, the inflation-adjusted total asset is total assets adjusted by converting into year-2012 dollar value.
Moreover, key variables of managerial characteristics of CEOs are further constructed, such as CEO’s age, gender, tenure, stock ownership, salary, bonus, option
granted, total compensation, cash compensation and restricted stock grantedetc.
There are two variables needed to be noticed, option_granted and
restricted_stock_granted. In the Execomp dataset, there are missing values of option
granted and restricted stock granted after 2006 because of the change of regulation. Hence the option_granted variable is constructed by combining the “option_awards_blk_val” (before 2006) and “option_awards_fv” (after 2006, FAS 123) in the sample. Similarly, therestricted_stock_granted variable is constructed by combining the “rstkgrnt” (before 2006) and “stock_awards_fv” (after 2006, FAS 123) in the Execomp dataset. The definitions of all variables are shown in the Appendix A.
4.2.2. Zero-leverage firms selection and classification
Zero-leverage firms are defined as firms with both zero long-term and short-term debt. Similar to the category of zero-leverage firms by Bessler et al. (2012), financially constrained and unconstrained zero-leverage firms are classified by using the benchmark of Size-age index (Hadlock and Pierce, 2010).
4.2.2.1. Size-age index
Size-age index is one of important benchmarks for categorizing the zero-leverage firms. As mentioned in the study of Hadlock and Pierce (2010), Size-age index, as a financial constraint index, is used to classify zero-leverage firms facing distinct financial constraints. The size-age index is defined as follows (Hadlock and Pierce, 2010).
Size − age index = −0.737 ∗ size + 0.043 ∗ size2− 0.040 ∗ age
The size is defined as the logarithm of inflation-adjusted total assets, which is also the firm size already constructed in the sample. The age is defined as the number of years that firm does not have missing value of stock prices in the COMPUSTAT. Specifically, by selecting the sample from the COMPUSTAT since 1950 and dropping the firm-years with missing values of the stock price (PRCC_F), the age is defined as the difference of current firm year and the minimum of total firm years.
4.2.2.2. Zero-leverage firm selection and classification
Zero-leverage firms are selected by keeping firm observations with both zero long-term and short-term debt. There are 9745 total observations of zero-leverage firms after the selection. Then classify zero-leverage firms further into financially constrained and unconstrained firms by Size-age index and equally Quintiles.
Based on the Size-age index (Hadlock and Pierce, 2010), first divide the total zero-leverage firms equally into five groups, marked as 0-4. Larger index is consistent with the greater financial constraint. Hence groups 0-1 are defined as financially unconstrained zero-leverage (FUZL) firms, while groups 3-4 are defined as the
financially constrained zero-leverage (FCZL) firms. The middle group is dropped from the whole groups. The frequencies of these zero-leverage firms are shown in the Table 1.
As shown in the Column N in the Table 1, over the period of 1996-2012, on average, there are 21% of zero-leverage firms in each year. And it also shows that fewer firms maintain zero-leverage policy as time goes by. The amount of unlevered firms in 2012 is almost the half of the amount in 1996. As shown in the Row Mean (1996-2012) of Table 1, unconstrained firms and unconstrained firms take the 8.99% and 7.86% of total firms on average, separately. Reported in Column FUZL firms in Table 1, total amount of FUZL firms in each year is greatly increased since 2000. And the percentage of the total sample is gradually increased during the past 16 years. On the contrary, as reported in the Column of FCZL firms, the amount and the percentage of FCZL firms gradually decreases over the period. However, seen from the whole period of 1996-2012, most of unlevered firms are financially unconstrained firms.
4.2.3. Control group selection
To study the effect of managerial characteristics on these unlevered firms, control groups need to be selected to compare the differences between unlevered firms and levered firms. Then it would be more convincible if the effect of managerial characteristics is significantly distinguished for zero-leverage firms. Control groups are chosen based on the similar industry (same three-digit SIC code), similar size (0.5 and 2 times of zero-leverage firm’s size) (Strebulaev and Yang, 2013) and similar performance (90% and 110% of zero-leverage firm’s ROA) (Devos et al., 2012).
Normally zero-leverage firms can also be included as peers. But in this study, control groups mainly consist of levered firms.
Table 1
Frequencies of zero-leverage (ZL) firms, financially constrained (FCZL) and unconstrained (FUZL) zero-leverage firms.
Zero-leverage (ZL) firms have both zero long-term debt and short-term debt (DLTT=0 and DLC=0, DLTT is the COMPUSTAT long-term debt while DLC is the debt in current liabilities). Financially unconstrained (FUZL) firms are zero-leverage firms with smaller Size-age index, while financially constrained (FCZL) firms are zero-leverage firms with larger Size-age index. Column N shows the number of total firms for each year in the sample. Columns ZL, FUZL, FCZL show the number of ZL, FUZL and FCZL firms in the sample. Column %-N of each type of firm reports the fraction of firms relative to total firms of each year.
Year N
ZL firms FUZL firms FCZL firms ZL %-N FUZL %-N FCZL %-N 1996 3302 553 16.,75% 80 2.42% 348 10.54% 1997 3548 571 16.09% 99 2.79% 340 9.58% 1998 3570 594 16.64% 137 3.84% 326 9.13% 1999 3759 633 16.84% 152 4.04% 334 8.89% 2000 3758 664 17.67% 201 5.35% 312 8.30% 2001 3366 634 18.84% 213 6.33% 290 8.62% 2002 3048 602 19.75% 220 7.22% 269 8.83% 2003 2792 595 21.31% 253 9.06% 233 8.35% 2004 2738 651 23.78% 276 10.08% 242 8.84% 2005 2644 639 24.17% 295 11.16% 208 7.87% 2006 2568 599 23.33% 293 11.41% 185 7.20% 2007 2535 592 23.35% 284 11.20% 179 7.06% 2008 2335 534 22.87% 269 11.52% 168 7.19% 2009 2144 488 22.76% 285 13.29% 128 5.97% 2010 2035 491 24.13% 282 13.86% 124 6.09% 2011 1940 476 24.54% 288 14.85% 120 6.19% 2012 1879 429 22.83% 271 14.42% 92 4.90% Total 47961 9745 20.32% 3898 8.13% 3898 8.13% Mean(1996-2012) 2821.24 573.24 20.92% 229.29 8.99% 229.29 7.86%
The control group selection is started with similar firms in the industry. First, identify control firms having the same 3-digit of SIC code as each zero-leverage firm,
financially unconstrained and constrained firm in the observation year over the period of 1996-2012 (Strebulaev and Yang, 2013). Of all firms in the same industry, the control groups are further chosen by similar size (log of inflation-adjusted total assets), selecting firms with size between 0.5 and 2 times of each zero-leverage firm observation (Strebulaev and Yang, 2013). By same industry and similar size, control groups are then chosen from the firms with similar performance, selecting firms with ROA between 0.9 and 1.1 times of each zero-leverage firm observation (Devos et al., 2012). After selecting control groups by these three benchmarks, each zero-leverage firms can have several proxies as its control group.
4.2.4. Descriptive statistics
The summary of descriptive statistics for key firm-specific variables as well as managerial characteristics variables between zero-leverage firms and their control groups are shown in the Table 2. In order to show whether each variable is significantly different between zero-leverage firms and their control groups, several T-tests are also conducted for each key variable.
Generally, CEOs of zero-leverage firms have significantly more salaries, bonuses and total compensation than CEOs of levered firms as shown in the Panel A of the Table 2. Financially unconstrained unlevered firms have larger capital expenditures than constrained firms. As for the firm performance, it seems that unconstrained firms have better performance than constrained firms.
As shown in the Table 2, CEOs of financially unconstrained firms get more salaries but fewer bonuses than CEOs of constrained firms. Besides, CEOs of
financially unconstrained firms have fewer stock ownership and granted options than CEOs of financially constrained unlevered firms. It also shows that CEOs of financially constrained firms have shorter tenures than those of unconstrained firms.
Table 2
Summary of descriptive statistics for zero-leverage (ZL) firms, financially unconstrained (FUZL), financially constrained (FCZL) zero-leverage firms and their control groups.
The ZL firms have both long-term and short-term zero debt. FUZL firms are zero-leverage firms with smaller Size-age index, while FCZL firms are zero-leverage firms with larger Size-age index. Panel A, Panel B and Panel C report the mean and standard deviation of variables for ZL, FUZL and FCZL firms and their control group, separately. Definitions of all variables are shown in the Appendix A. T-statistics and average differences of t-test for significant differences of each variable between unlevered firms and their control groups are reported in the last two columns. The values with *, ** and *** indicate the significance of 10%, 5% and 1% level separately.
Variable ZL firms Control group t-statistics difference-mean
Mean sd mean sd
Panel A: zero-leverage firms and control group
Firm age 11.78 6.43 10.75 5.84 6.15 0.889*** Firm size 6.96 1.33 6.09 1.25 28.63 0.862*** Cash 0.21 0.20 0.37 0.21 -35.16 -0.163*** Dividend 0.01 0.07 0.01 0.07 -2.24 -0.00357* Tobin's Q 2.37 4.00 3.11 4.44 -7.44 -0.727*** Tangibility 0.24 0.22 0.16 0.16 16.17 0.0756*** Profitability 0.07 0.17 0.08 0.21 -1.63 -0.00689 Capital expenditures 149.17 701.32 53.64 260.37 6.81 92.52*** ROA 0.01 0.26 0.02 0.26 -2.81 -0.0169** Growth opportunities 2.39 3.12 3.11 4.39 -8.59 -0.714*** Salary 553.30 289.09 448.21 244.34 15.38 97.91*** Bonus 308.99 1391.10 203.88 462.12 3.78 101.4*** CEO stock ownership 24.99 64.95 29.69 65.07 0.86 791.1 CEO option granted 25.92 62.51 28.76 74.23 0.84 1157.7 CEO cash compensation 6.43 1.09 6.13 1.36 10.68 0.290***
CEO Gender 0.97 0.17 0.96 0.20 2.14 0.00873* CEO's age 53.46 7.77 52.67 8.65 3.72 0.691*** CEO tenure 7.43 7.39 7.80 7.81 -2.14 -0.371* Total compensation 4295.05 11600.34 3522.98 12721.01 2.66 708.3** Number of observations 6268 2432
Panel B: Financially unconstrained zero-leverage firms and control group Firm age 12.22 6.42 11.46 5.90 3.98 0.652*** Firm size 6.96 1.34 6.21 1.25 21.89 0.751*** Cash 0.23 0.20 0.38 0.20 -27.64 -0.146*** Dividend 0.01 0.07 0.01 0.06 -1.68 -0.00302 Tobin's Q 2.46 4.47 3.18 4.77 -5.94 -0.717*** Tangibility 0.21 0.20 0.15 0.14 11.93 0.0570*** Profitability 0.07 0.16 0.09 0.16 -4.01 -0.0166*** Capital expenditures 129.90 617.84 60.67 288.80 4.88 67.46*** ROA 0.01 0.17 0.04 0.17 -6.28 -0.0286*** Growth opportunities 2.48 3.45 3.17 4.75 -6.67 -0.682*** Salary 553.46 283.57 459.80 253.09 12.53 89.34*** Bonus 288.33 1455.93 201.79 488.14 2.55 81.54*
CEO stock ownership 22.77 60.08 28.59 64.71 0.88 1049.5 CEO option granted 25.37 62.91 28.39 72.23 0.86 1535.0 CEO cash compensation 6.41 1.15 6.12 1.42 8.63 0.280***
CEO Gender 0.97 0.17 0.96 0.20 2.53 0.0115* CEO's age 53.56 7.71 52.70 8.70 3.69 0.773*** CEO tenure 7.48 7.45 7.96 7.88 -2.27 -0.447* Total compensation 4481.90 12844.20 3698.73 13432.65 2.18 727.4* Number of observations 4860 1964
Panel C: Financially constrained zero-leverage firms and control group
Firm age 11.49 6.45 10.09 5.92 7.81 1.253*** Firm size 6.86 1.35 5.98 1.29 26.12 0.890*** Cash 0.24 0.21 0.40 0.21 -30.34 -0.160*** Dividend 0.01 0.08 0.01 0.07 -1.20 -0.00236 Tobin's Q 2.53 4.54 3.28 4.84 -6.06 -0.740*** Tangibility 0.21 0.21 0.15 0.15 11.79 0.0580*** Profitability 0.06 0.26 0.07 0.22 -0.89 -0.00561 Capital expenditures 117.32 541.85 53.38 285.22 5.38 64.05*** ROA -0.01 0.33 0.01 0.28 -2.31 -0.0185* Growth opportunities 2.56 3.52 3.28 4.79 -6.88 -0.713*** Salary 533.38 282.61 412.00 214.35 16.71 112.6*** Bonus 298.77 1019.07 199.22 452.96 4.38 96.88*** CEO stock ownership 24.72 67.34 30.60 66.57 0.90 1040.4 CEO option granted 27.45 63.79 32.36 80.71 0.88 1522.7 CEO cash compensation 6.40 1.09 6.05 1.39 10.47 0.323***
CEO Gender 0.97 0.17 0.96 0.20 2.43 0.0108*
CEO's age 52.95 7.76 52.07 8.34 3.44 0.703***
CEO tenure 7.28 7.06 7.35 7.17 -0.38 -0.0696
Total compensation 4489.37 12983.29 3694.59 13978.15 2.32 764.0*
4.3. Zero-leverage firms and managerial characteristics of executive
directors
Executive directors are also in charge of daily management. Some characteristics of executive directors may also have potential effects on maintaining zero leverage policy. To study the potential effect of managerial characteristics of executive directors on zero-leverage firms, the sample is constructed by merging COMPUSTAT and RiskMetrics-Director data set over the period of 2007-2012. Specially, the sample is further restricted to those director’s observations with the employment title of “Executive”. There are 7370 observations after the merge and data restrictions.
Key variables of managerial characteristics of executive directors are then constructed. The gender of directors, constructed as a dummy variable, equals to 1 if the observation is male. The ethnicity of directors equals to 1 if the ethnicity in the sample is not “UNKNOWN”. Similarly, if the executive director is also employed as “CEO”, the employment title-CEO equals to 1. The construction of other variables such as director’s age, tenure, number of shares hold and employment title as CEO are shown in the Appendix A.
Different from the sample of managerial characteristics of CEOs, the sample of executive directors is limited. Zero-leverage firms are chosen as defined, but not further classified into unconstrained and constrained firms. Hence the effect of managerial characteristics of executive director is mainly examined between zero-leverage firms and levered firms. Table 3 reports the summary of statistics of key managerial characteristics of executive directors.
Table 3
Summary of descriptive statistics for zero-leverage (ZL) firms in the Director’s sample.
This table shows the descriptive statistics for ZL firms and levered firms in the sample of directors. The ZL firms have both long-term and short-term zero debt. Column N reports the mean and standard deviation of key variables of total firm-observations. Column ZL firms reports the mean and standard deviation of key variables of zero-leverage firms. Definitions of all variables are shown in the Appendix A.
5. Results
5.1. Managerial characteristics of CEOs and zero-leverage firms
Using the merged dataset of the Execucomp and COMPUSTAT data set over the period 1996-2012, the multivariate logit regressions are used to test the effects of CEOs’ managerial characteristics on zero-leverage firms. The regressions are conducted between zero-leverage firms and their control groups. Control variables are firm size, age, cash, dividend payout ratio and other firm-specific variables. Table 4, Table 5 and Table 6 report the regression results of zero-leverage firms, financially unconstrained and constrained unlevered firms. To avoid the outlier problem, all independent variables are “winsorizing”-adjusted and lagged for one year.
Variable
N ZL firms
mean sd mean sd Director’s gender 0.92 0.27 0.94 0.23
Ethnicity 0.68 0.47 0.63 0.48
Employment title: CEO 0.53 0.50 0.53 0.50 Number of shares hold (million) 1.65 8.49 1.42 4.48 Director’s age 57.03 7.95 56.87 8.68 Director’s tenure 11.69 27.73 12.00 9.32 Number of observations 5795 1324
From all results of multivariate logit regressions for all zero-leverage firms shown in Table 4, Table 5 and Table 6, there are some significant conclusions. Managerial characteristics such as CEO gender, age, salary, bonus, all other compensation and total compensation, are negative related to zero leverage. It reports that female and younger CEO tends to choose zero leverage. This is almost similar to the existing study of DeYoung, Peng and Yan (2010).They find that old CEO tends to choose more leverage which also indicates that young CEO tends to choose lower leverage even zero leverage. However Bertrand and Schoar, (2003) has the opposite conclusion regarding the CEO’s age. They find that older CEOs are more likely to select lower leverage. Besides, as shown in the Table 4, 5 and Table 6, CEO with longer tenure also tends to choose zero leverage.
When it comes to CEO’s compensation, unlevered firms always grant their CEOs with high stock ownership and options. This is consistent with the study of Lewellen (2006). He finds that more stock based compensations lead to less debt of the companies. Similarly, Duru et al. (2012) also show that CEOs of the companies with fewer debts tends to have higher executive compensations. Besides, CEOs of zero-leverage firms in the tables are shown to own fewer other type of compensations. Within all the results of the regressions, CEO cash compensation does not significantly determine zero-leverage choices. However, within all variables, CEO
gender and cash compensation have larger effects on the each kind of zero-leverage
firms than other variables. This may reflect that female CEOs may be a determinant role for a firm to maintain unlevered.
When comparing the financially unconstrained and constrained zero-leverage firms, results show that managerial characteristics of CEOs have more significant effects on the financially unconstrained firms. Coefficients of most variables in the Table 5 are larger than the coefficients in the Table 6. For example, the coefficient of
CEO tenure of unconstrained firms (0.0208) is much larger than the coefficient of CEO tenure of constrained firms (0.0090). This indicates that there is about 2%
probability of financially unconstrained firms to be unlevered than constrained firms. Similarly, the managerial characteristics of CEO gender, age, salary, bonus, stock ownership and all other compensations also have stronger effects on financially unconstrained firms. Moreover, the effect of CEO stock ownership is significant (0.00183***) for the financially unconstrained firms, but insignificant (0.000610) for the constrained firms. This difference also indicates that managerial characteristic of CEO stock ownership affects more significantly on financially unconstrained firms. Because of fewer financial constraints, it’s more flexible for executives of financially unconstrained firms to make their own decisions on zero-leverage policy.
As shown in the Table 4, Table 5 and Table 6, the coefficients of CEO cash
compensation, restricted stock granted and total compensation are all financially
significant when testing the significance of single variable separately. However, when test the significance of the multivariate logit model, these variables are not significant in the results, as shown in the Column (12) of each table. This indicates that managerial variables such as cash compensation, restricted stock granted and total compensation have little effect on leverage policies of zero-leverage firms.
In conclusion, the results of the multivariate logit regressions of zero-leverage firms with distinct constraints indeed prove that the hypothesis holds for true. Managerial characteristics of CEOs indeed have a more significant effect on financially unconstrained unlevered firms.
5.2. Managerial characteristics of executive directors and
zero-leverage firms
Except for Chief executive officers (CEOs), some directors entitled as “Executive” are also in charge of the daily operation and management of companies. They manage company affairs and making business plans. Hence managerial characteristics of executive directors are also taken into consideration as the plausible determinants of zero leverage. Based on the merged dataset of Risk Metrics and COMPUSTAT from 2007 to 2012, the managerial characteristics of executive directors are studied by using the multivariate logit regressions.
The sample is restricted to the director whose employment title is “Executive”. The variables of directors are defined in the Appendix A. In the directors’ sample, control groups cannot be chosen because of the limitation of the number of observations. Hence the multivariate logit regressions are conducted between zero-leverage firms and levered firms. Firm-specific variables are regarded as control variables. Dependent variable, also as a dummy variable, equals to one if the observation is zero-leverage firm. All independent variables are lagged for one year. The results of the multivariate logit regressions are shown in the Table 7.
Table 4
Determinants of managerial characteristics: CEOs and zero-leverage (ZL) firms.
The table reports the results of multivariate logit regressions on the effects of managerial characteristics (CEOs) on zero-leverage firms. Zero-leverage firms are firms with both zero long-term and short-term debt. The dependent variable, also as a dummy variable, equals to 1 if the firm observation is zero-leverage firm. All independent variables are lagged for one year. All the variables are defined in the Appendix A. Control variables are firm size, firm age, cash, dividend payout ratio, capital expenditure, tangibility, Tobin’s Q and other firm-specific variables. The coefficients, number of observations, t statistics (in parentheses), economic significance and clustered R-squared are reported. The coefficients with *, ** and *** indicate the significance of 10%, 5% and 1% level separately.
Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Gender -0.348*** -0.456*** (-3.69) (-4.13) CEO tenure 0.0172*** 0.0260*** (7.17) (8.71) CEO age -0.0144*** -0.0117*** (-5.59) (-3.97) Salary -0.00206*** -0.00169*** (-24.58) (-10.33) Bonus -0.000554*** -0.000283*** (-12.08) (-3.52)
CEO stock ownership 0.00334*** 0.00135***
(11.14) (3.81)
CEO options granted 0.00170*** 0.00173**
(3.46) (2.72)
CEO cash compensation -0.614*** 0.0367
(-21.89) (0.54)
Restricted stock granted -0.000196*** -0.0000223
(-11.61) (-1.13)
All other compensation -0.00197*** -0.000720***
(-12.28) (-5.49) Total compensation -0.0000891*** -0.0000137 (-14.97) (-1.71) Constant -0.893*** -1.346*** -0.433** -0.121** -1.043*** -1.298*** -1.243*** 2.763*** -1.084*** -1.029*** -0.892*** 0.426 (-9.66) (-46.49) (-3.12) (-2.67) (-44.74) (-57.79) (-52.97) (15.33) (-49.26) (-44.32) (-32.57) (1.06) Number of observations 17059 13572 13572 13572 13572 13572 13505 13572 13555 13572 13491 13491 Pseudo R-squared 0.0007 0.0034 0.0021 0.0495 0.0126 0.008 0.0008 0.0347 0.0129 0.0181 0.0211 0.0669
Table 5
Determinants of managerial characteristics: CEOs and financially unconstrained zero-leverage (FUZL) firms.
The table reports the results of multivariate logit regressions of the effect of managerial characteristics (CEOs) on the financially unconstrained zero-leverage (FUZL) firms. Zero-leverage (ZL) firms are firms with both zero long-term and short-term debt. FUZL firms are ZL firms with few financial constraints. The dependent variable, also as a dummy variable, equals to 1 if the firm observation is FUZL firm. All independent variables are lagged for one year. All the variables are defined in the Appendix A. Control variables are firm size, firm age, cash, dividend payout ratio, capital expenditure, tangibility, Tobin’s Q and other firm-specific variables. The coefficients, number of observations, t statistics (in parentheses), economic significance and clustered R-squared are reported. The coefficients with *, ** and *** indicate the significance of 10%, 5% and 1% level separately.
Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Gender -0.398*** -0.498*** (-3.80) (-4.02) CEO tenure 0.0208*** 0.0284*** (7.77) (8.43) CEO age -0.0108*** -0.0113*** (-3.70) (-3.34) Salary -0.00197*** -0.00154*** (-20.99) (-8.72) Bonus -0.000476*** -0.000181* (-9.46) (-2.12)
CEO stock ownership 0.00396*** 0.00183***
(11.06) (4.35)
CEO options granted 0.00182** 0.00191*
(3.16) (2.56)
CEO cash compensation -0.572*** -0.00909
(-18.19) (-0.13)
Restricted stock granted -0.000179*** -0.0000147
(-10.06) (-0.71)
All other compensation -0.00196*** -0.000707***
(-10.57) (-4.67) Total compensation -0.0000768*** -0.0000154 (-12.42) (-1.85) Constant -0.735*** -1.268*** -0.518*** -0.0463 -0.963*** -1.201*** -1.139*** 2.598*** -0.978*** -0.928*** -0.816*** 0.720 (-7.17) (-38.84) (-3.30) (-0.89) (-36.91) (-47.40) (-42.90) (12.84) (-39.21) (-35.42) (-26.60) (1.68) Number of observations 12843 10062 10062 10062 10062 10062 10017 10062 10045 10062 10004 10004 Pseudo R-squared 0.001 0.0052 0.0012 0.0463 0.0095 0.0104 0.0009 0.0311 0.0121 0.0172 0.0182 0.0655
Table 6
Determinants of managerial characteristics: CEOs and financially constrained zero-leverage (FCZL) firms.
The table reports the results of multivariate logit regressions of the effect of managerial characteristics (CEOs) on the financially constrained zero-leverage (FCZL) firms. Zero-leverage firms are firms with both zero long-term and short-term debt. FCZL firms are ZL firms with more financial constraints. The dependent variable, also as a dummy variable, equals to 1 if the firm observation is FCZL firm. All the variables are defined in the Appendix A. All independent variables are lagged for one year. Control variables are firm size, firm age, cash, dividend payout ratio, capital expenditure, tangibility, Tobin’s Q and other firm-specific variables. The coefficients, number of observations, t statistics (in parentheses), economic significance and clustered R-squared are reported. The coefficients with *, ** and *** indicate the significance of 10%, 5% and 1% level separately.
Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Gender -0.347** -0.544*** (-3.09) (-4.04) CEO tenure 0.00900** 0.0222*** (3.02) (6.08) CEO age -0.0210*** -0.0129*** (-6.89) (-3.75) Salary -0.00229*** -0.00208*** (-21.86) (-10.59) Bonus -0.000588*** -0.000276** (-10.74) (-3.01)
CEO stock ownership 0.00244*** 0.000610
(6.68) (1.41)
CEO options granted 0.00217*** 0.00189**
(4.23) (2.83)
CEO cash compensation -0.617*** 0.0674
(-19.24) (0.91)
Restricted stock granted -0.000155*** 0.0000119
(-8.41) (0.55)
All other compensation -0.00169*** -0.000604***
(-9.67) (-4.27) Total compensation -0.0000676*** -0.0000115 (-11.32) (-1.44) Constant -0.831*** -1.222*** -0.0421 -0.00466 -0.981*** -1.216*** -1.209*** 2.805*** -1.058*** -1.003*** -0.897*** 0.586 (-7.54) (-35.94) (-0.26) (-0.09) (-35.75) (-46.37) (-43.40) (13.71) (-41.22) (-37.45) (-28.67) (1.32) Number of observations 12297 9534 9534 9534 9534 9534 9480 9534 9518 9534 9467 9467 Pseudo R-squared 0.0007 0.0009 0.0046 0.0567 0.0143 0.004 0.0016 0.0379 0.0087 0.0155 0.0162 0.0686
Table 7
Determinants of managerial characteristics: Executive directors and zero-leverage (ZL) firms.
The table reports the results of multivariate logit regressions of the effects of managerial characteristics (Executive directors) on the zero-leverage firms. Zero-leverage firms have both zero long-term and short-term debt. The dependent variable, also as a dummy variable, equals to 1 if the firm observation is zero-leverage firm. All independent variables are lagged for one year. All the variables are defined in the Appendix A. Control variables are firm size, firm age, cash, dividend payout ratio, capital expenditure, tangibility, Tobin’s Q and other firm-specific variables. The coefficients, number of observations, t statistics (in parentheses), economic significance and clustered R-squared are reported. The coefficients with *, ** and *** indicate the significance of 10%, 5% and 1% level separately.
Variable (1) (2) (3) (4) (5) (6) (7) Director’s gender 0.350** 0.338* (2.79) (2.11) Ethnicity -0.180** -0.174* (-2.80) (-2.07) Employment title:CEO -0.0170 -0.0728 (-0.28) (-0.95)
Number of shares -7.39e-09 -7.69e-09
(-1.37) (-1.40) Director's age -0.00105 -0.000955 (-0.23) (-0.20) Director tenure 0.000271 0.000301 (0.25) (0.28) Constant -1.815*** -1.373*** -1.481*** -1.499*** -1.459*** -1.526*** -1.604*** (-14.96) (-26.75) (-33.45) (-38.72) (-5.52) (-38.44) (-5.32) Number of observations 7057 7057 7057 4810 4863 4857 4802 Pseudo R-squared 0.0012 0.0012 0 0 0 0.0014 0.0039
As shown in the Table 7, only two characteristics of director’s gender (0.350**) and ethnicity (-0.180**) have significant effects on zero-leverage firms. Besides, the coefficient of director tenure (0.00972) is significant at the 10% level. Firms are more likely to maintain unlevered if their executive directors have longer tenure, as reported in the Table 7. This is consistent with the multivariate logit regressions results of CEOs’ sample. A CEO with longer tenure tends to select zero leverage. However, coefficients of number of shares hold and
director’s age are not significant. Different from the results of the regressions of CEO’s
sample, firms with male executive directors are more likely to maintain unlevered. In the sample of CEO’s managerial characteristics, female CEOs are more likely to choose zero leverage. Moreover, if the executive director does not have an explicit ethnicity, the firm they
managed tends to be unlevered. The effects of executive director’s gender and ethnicity have not been studied in the existing empirical research. However, the results of the regressions of executive directors in the Table 7 cannot strongly support the hypothesis.
6. Robustness check and additional tests
In the results of the multivariate logit regression analysis, coefficients of CEO cash
compensation, restricted stock granted and total compensation are all insignificant for
zero-leverage firms, financially constrained and unconstrained unlevered firms. Several additional tests are conducted in order to verify the robustness of the multivariate logit regression results. First, the variable of CEO cash compensation is dropped to control for the possible side-effect of this insignificant variable. Second, we drop both cash compensation and restricted stock granted from the model. Third, all three insignificant variables are excluded from the regressions to avoid the possible negative influence on the results. These tests are conducted separately for each subsample of zero-leverage firms.
After excluding of these insignificant variables, the multivariate logit regressions end up better results. There are two significant findings after conducting these robust tests. First, when dropping both CEO cash compensation and restricted stock granted, the coefficient of
total compensation changes from insignificant to significant. Except for financially
constrained firms, this change is significant for both zero-leverage firm and financially unconstrained firms. As shown in the original results, managerial characteristics have more significant effects on financially unconstrained firms. Hence this change indicates that the insignificance of total compensation is mainly under the side-effects of other two insignificant variables, especially for financially unconstrained firms. Second, when dropping all three insignificant variables from the model, the regression results tend to be much better.
This happens especially significant for financially unconstrained zero-leverage firms. Regression results of robustness tests are shown in the Table 9, Table 10 and Table 11 in the Appendix B.
The results of robustness tests support the hypothesis again. Significant changes and better results, especially for financially unconstrained firms, indicate that managerial characteristics of CEOs have a much more strong influence on financially unconstrained unlevered firms.
Table 7 reports the original results of multivariate logit regression of managerial characteristics of executive directors on zero-leverage firms. As shown in the results, only gender and ethnicity are reported significant. It shows that rests of the characteristics have little influences on zero-leverage choice of executive directors. One of the possible reasons is the potential outliers of these insignificant variables. In order to additional test this result, all of independent are “winsorizing”-adjusted in the STATA to exclude the potential outliers. Then logit regressions conducted on the “winsorizing”-adjusted variables ends up having better results than before.
Table 8 reports the multivariate logit regression results of “winsorizing”-adjusted executive directors’ variables. Different from the original results, the coefficient of director
tenure changes from insignificant (0.000301) to significant (0.0125*) at the 10% level after
handling outliers. Although coefficients of three insignificant variables still remain unchanged, the whole results are better than before. All results in the Table 7 and Table 8 indicate that managerial characteristics such as the employment title as CEO, number of
shares hold and director’s age, have slight effects on the leverage choices of executive
Table 8
Determinants of managerial characteristics: Executive directors and zero-leverage (ZL) firms.
The table reports the results of multivariate logit regressions of the effects of managerial characteristics (Executive directors) on the zero-leverage firms over the period of 2007-2012. Zero-leverage firms have both zero long-term and short-term debt. The dependent variable, also as a dummy variable, equals to 1 if the firm observation is zero-leverage firm. The independent variables are “winsorizing”-adjusted and lagged for one year. All the variables are defined in the Appendix A. Control variables are firm size, firm age, cash, dividend payout ratio, capital expenditure, tangibility, Tobin’s Q and other firm-specific variables. The coefficients, number of observations, t statistics (in parentheses), economic significance and clustered R-squared are reported. The coefficients with *, ** and *** indicate the significance of 10%, 5% and 1% level separately.
Variable (1) (2) (3) (4) (5) (6) (7) Director’s gender 0.350** 0.305 (2.79) (1.90) Ethnicity -0.180** -0.172* (-2.80) (-2.06) Employment title:CEO -0.0170 -0.0773 (-0.28) (-1.01)
Number of shares hold 1.12e-09 -6.01e-09
(0.13) (-0.67) Director's age -0.000822 -0.00748 (-0.17) (-1.38) Director tenure 0.00972* 0.0125* (2.40) (2.55) Constant -1.815*** -1.373*** -1.481*** -1.515*** -1.472*** -1.637*** -1.354*** (-14.96) (-26.75) (-33.45) (-37.87) (-5.46) (-26.57) (-4.24) Number of observations 7057 7057 7057 4810 4863 4857 4802 Pseudo R-squared 0.0012 0.0012 0 0 0 0.0012 0.0035
7. Conclusion
In the existing research, firms maintain less or zero leverage than predicted in the capital structure theories. Characteristics of managers and executives, as well as financial constraints hypothesis are regarded as plausible explanations for the zero-leverage phenomenon. Nevertheless, the zero - leverage phenomenon, related to the low leverage puzzle, cannot be fully explained. There are still missing potential explanations to be explored. In order to study the plausible determinants of zero leverage, this paper is intended to examine the potential effects of managerial characteristics on zero-leverage firms. Moreover, in the presence of
market friction, zero-leverage firms have different capacities to obtain external financing. Because of different levels of financial constraints, managerial characteristics may have distinguished effects on constrained and unconstrained unlevered firms. Hence the main research question of this paper is to explore the effects of managerial characteristics on financially unconstrained unlevered firms as well as constrained unlevered firms.
Using the data from COMPUSTAT, Execucomp and RiskMetrics over the period of 1996-2012, on average, there are about 21% zero-leverage firms. Moreover, after the classification of zero-leverage firms, there are about 8.9% of total firms with low financial constraints and 7.86% of firms with high financial constraints. These firms are regarded as financially unconstrained and constrained unlevered firms. These firms are classified using the Size-age index mentioned in the study of Hadlock and Pierce (2010). For each zero-leverage firm, several proxies are chosen as control group based on the benchmarks of similar industry, firm size and performance.
Multivariate logit regression analysis is chosen as the methodology to study the effect of managerial characteristics of CEOs and executive directors. In this paper, managerial characteristics are assumed to have more significant effects on financially unconstrained zero-leverage firms than constrained firms. To test this hypothesis, several logit and multivariate logit regressions are conducted using the sample of CEOs and executive directors. Some of the results are consistent with the findings in the existing empirical studies. Besides, other results have contributed to the former research by finding the plausible effects of several untapped managerial characteristics.
In the multivariate logit regression results, most of managerial characteristics are reported to be financially significant. In the sample of managerial characteristics of the CEOs, female and young CEOs tend to prefer zero-leverage policy. This finding is consistent with the result of DeYoung, Peng and Yan (2010). Their result also shows that firm with older
CEOs ends up having more leverage. And a CEO with long tenure is more likely to choose zero leverage policy. Referring to CEOs’ compensation, zero-leverage firms tend to grant less salary, bonus and all other compensations to their CEOs. Nevertheless, CEOs in zero-leverage firms are always granted more options and stock ownership. Similarly, Duru et al. (2012) also find that firm with fewer debt grants more executive compensations to their CEOs. When comparing the effects on financially constrained and unconstrained unlevered firms, most of the effects of managerial characteristics are more significant for financially unconstrained firms than their constrained peers. Overall, the results of the CEOs’ sample suggest that managerial characteristics of CEOs have more significant effects on financially unconstrained firms. Hence the hypothesis proves to be true.
Except for CEOs, characteristics of the executive directors also have potential effects on the leverage of firms. In the results of multivariate logit regressions of the executive directors’ sample, some characters are found to be plausible explanations for zero leverage firms. Different from the results of the CEOs’ sample, executive director’s gender has the opposite effect on zero-leverage firms. Male executive directors prefer maintaining zero leverage. Moreover, if an executive director has specific ethnicity, the firm is more likely to be unlevered. And executive directors with longer tenure tend to choose zero-leverage policy.
In the regression results, most of the managerial characteristics of CEOs and executive directors are proven to be plausible explanations for firms without leverage. The findings of this paper also have some implications for firms to make leverage choices. For example, as shown in the results, female CEO or male executive director prefers zero leverage. As female CEOs take more important roles of corporate governance, they may take risky strategies of capital structure and leverage choices. Consistent with the hypothesis, managerial characteristics affect more on financially unconstrained zero-leverage firms. Because of few financial constraints, unconstrained firms are more flexible to obtain funds externally and
internally. Hence in order to choose a proper capital structure, firms should evaluate not only their financial constraints but also the characteristics of their executive directors and managers.
Because of the limited number of observations in the Directors’ sample, some variables such as director’s age and number of shares hold don’t have significant effects on zero-leverage firms. There are also limitations to conduct more effective robustness check for the regression results in this paper. Besides, as one of the puzzling phenomenon in the corporate finance, there is still much potential evidence needed to be explored. Hence future studies can choose more sufficient sample to investigate the untapped explanations for the zero-leverage phenomenon. In this case, the capital structure and leverage policy can be better understood.
