The impact of CEO power on corporate capital structure: New evidence from dynamic panel threshold analysis

https://doi.org/10.1016/j.iref.2017.05.010Get rights and content

Highlights

  • This study examines the non-linearity between CEO power and corporate capital structure.

  • We employ an innovative dynamic panel threshold model to estimate the nexus.

  • We find that CEO power has a strong positive influence on leverage in “low-CEO power” firms.

  • However, for “high-CEO power” firms, the impact is negative.

  • The results are robust to alternative measures of leverage and CEO power.

Abstract

This study examines the non-linearity between CEO power and corporate capital structure. Previous studies show that firm leverage responds differently to CEO power changes. In order to capture this non-linear relationship, we employ an innovative dynamic panel threshold model, this novel method allows the estimation of threshold effects with panel data even in case of endogenous regressors. Using a panel-dataset of Chinese SMEs from 2009 to 2013, we find that a CEO power threshold exists in the CEO power-firm leverage association. CEO power has a strong positive and statistically significant determinant of firm leverage, in the “low-CEO power” firms. However, at “high-CEO” regime, the impact is negative but insignificant determinant of leverage. The results are robust to alternative measures of leverage and CEO power, as well as additional explanatory variables.

Introduction

The determinants of corporate capital structure have been extensively investigated in the literature. Among them, the majority of existing evidence shows that the corporate capital structure is not only be affected by firm-, industry-, and market-level characteristics, but also by the personal traits of top managers. Recently, studies have started to pay particular to how decision making power of Chief Executive Officer (CEO) shapes capital structure decisions (Chintrakarn et al., 2014, Jiraporn et al., 2012). However, the empirical research on this topic still remains scarce and there is no consistent conclusion about the relationship between CEO power and firm leverage. In addition, although we have learned much from those of prior studies, most models in the literature are static, making it difficult to develop tests of the association between CEO power and capital structure dynamics. The purpose of our study is to investigate the CEO power-firm leverage nexus taking account for capital structure dynamics.

The start of modern capital structure research can be traced back to Modigliani and Miller (1958). In this study, they show that subject to certain conditions, the value of a firm is independent of its capital structure choice. Since then, financial economists have devoted significant effort to studying the determinants of capital structure and several theories have been developed to show that market frictions and imperfections do matter in shaping capital structure. One theory that has been broadly employed to interpret the relationship between manager behaviour and firm leverage is agency theory. The central theme of agency theory is that corporate capital structure is determined by agency costs which arise from the divergence of ownership and control (Berle & Means, 1932) and the imperfect alignment of interests between managers and owners (Jensen & Meckling, 1976). Due to this, the prevalent view presents that self-serving managers do not make capital structure decisions that maximise owner wealth (Morellec, Nikolov, & Schürhoff, 2012).

Yet, although the agency theory predicts that agency costs can lead to firm leverage deviate from the optimal level for owners, it is still unclear whether agency costs can result in too much or too little leverage (Jiraporn et al., 2012). On the one hand, managers might voluntarily use more than the optimal amount of debt to consolidate their equity voting power and avoid takeover threats (Harris & Raviv, 1990). On the other hand, managers may pursue lower levels of leverage to avoid the disciplining role of debt. For instance, Grossman and Hart (1982) and Jensen (1986) argue that debt is a disciplining instrument that can be applied to mitigate agency problems by reducing the free cash flow availability of managers. Moreover, the use of high leverage can increase the probability of bankruptcy and job loss (Friend and Lang, 1988, Jensen, 1986). In this case, managers have incentives to protect their under-diversified human capital, thus prefer less debt than the optimal level.

Likewise, building on the abovementioned predictions and extensive management literature on managerial discretion (see, Bertrand and Schoar, 2003, Adams et al., 2005, Cronqvist et al., 2012), previous studies document that CEO power has a significant impact on capital structure. More specifically, earlier study by Jiraporn et al. (2012) show that as CEO power increases, firms use significantly lower level of leverage. However, this finding is challenged by Chintrakarn et al. (2014), who argue that the effect of CEO power on leverage is complex and the simple linear relation is spurious. They also suggest that to investigate the reliable relationship between CEO power and leverage, researchers should use non-linear models. Through testing ad hoc non-linear models, they find that the relationship between CEO power and firm leverage is hump-shaped. They argue that firms with relatively weak CEOs, that is CEOs that hold less decision-making power, appear to be in favour of higher leverage. This is because corporate capital structure choices are more influenced by other stakeholders, such as board of directors (BOD). As a result, firm tends to use more debt to reduce the agency costs arising from conflicts between CEO and owners. However, as CEO continues to have higher power and grows beyond a certain threshold, he/she is more likely to manipulate corporate leverage in order to pursue their own benefits. In this case, CEO tends to pursue lower debt levels to avoid the disciplining role of debt.

This paper provides new evidence that sheds light on the impact of CEO power on firm leverage. Specifically, we explore whether there exist threshold levels of CEO power in the power–leverage relationship. One of the most interesting forms of nonlinear regression models with wide applications in economics is the threshold regression model. The importance of this model stems from the fact that it treats the sample split value (threshold parameter) as unknown. Tong (1983) first proposes threshold regression models for time series data. Hansen (1999) extends the threshold regression to static panel data structure and derives the corresponding asymptotic theory for threshold parameters and regression slopes. Therefore, a natural starting point for the empirical analysis of CEO power thresholds is the panel threshold model suggested by Hansen (1999). However, the application of the Hansen (1999) threshold model to the empirical analysis of a CEO power–firm leverage relationship is not without problems. The most important limitations of this method are that the model is a static setup and all regressors are required to be exogenous. In reality, a firm's capital structure decisions are inherently dynamic and the past financing decisions may proxy for some unobservable firm characteristics that influence the current decisions (Florackis and Ozkan, 2009, Guney et al., 2011, Morellec et al., 2012). Therefore, to account for the dynamic process in corporate capital structure, it is essential to employ a more advanced technique. The Caner and Hansen (2004) threshold model is able to deal with the dynamic issue, but this technique is based on cross-section analysis (Law & Singh, 2014). Since the data we employed in this study is panel data, which can provide more information and mitigate multicollinearity as well as control for cross firm heterogeneity, it is therefore more appropriate to use other estimation methods. To this end, we apply the dynamic panel threshold model proposed by Kremer, Bick, and Nautz (2013), building on Hansen (1999) and Caner and Hansen (2004). In the dynamic model, the endogeneity of important regressors is no longer an issue, thus it provides robust results.

To the best of our knowledge, this technique has not been employed before in analysing a CEO power–firm leverage nexus. In comparison with those of ad hoc non-linear methods employed in prior studies on CEO power–leverage relationship, the Kremer et al. (2013) methodology has three distinctive advantages which are summarised as follows: 1) the threshold model does not require any specified functional form of nonlinearity, such as previous research by Chintrakarn et al. (2014) that add a quadratic term in the regression; 2) the number and location of thresholds are (endogenously) determined by the data, that is, it internally sorts the data, on the basis of some threshold determinant, into groups of observations each of which obeys the same model; 3) asymptotic theory applied in the threshold model can be used to construct appropriate confidence intervals and a bootstrap method can be employed to determine the statistical significance of the thresholds. Therefore, this methodology allows us to examine the threshold effects of the CEO power–leverage link in a more adequate and flexible way than prior studies.

To large extent, this paper extends the existing literature in respect of this research method. Moreover, our study also contributes to the understanding of determinants of capital structure by investigating the impact of personal characteristics of the firm's top executive, the CEO. Since the seminar work by Modigliani and Miller (1958), economists have devoted significant effort to studying the determinants of capital structure. The focus of most empirical work has been on firm, industry, and market characteristics. Nevertheless, the findings show that firms that are similar in terms of these fundamentals often choose very different corporate leverage. Thus, this has emphasized the importance of studying the impact of the personal traits of a CEO (Cronqvist et al., 2012). In particular, this study investigates the impact of CEO power on capital structure in Chinese SMEs based on agency theory, managerial power theory and corporate governance theory.

Furthermore, although there are many prior empirical studies on financing decisions, much less attention is paid to the small and medium sized enterprises (SMEs), especially in emerging markets, given that their growth and prosperity is subjected to different contingencies and constraints (Mateev, Poutziouris, & Ivanov, 2013). This paper, therefore, adds to the existing empirical literature by employing a sample of SMEs in transition economies, specifically China. Finally, the findings from the present study provide an important implication for the growth of the firm, thus furthering Chinese economic development. China has become the largest emerging market and the second largest economy in the world. The great success of the economic development is driven primarily by SMEs which make up the vast majority of firms and contribute more GPD, jobs and production than those of large companies (Huang et al., 2016, Li et al., 2016). Understanding the specific determinants of financing policy decisions might be vitally important for SMEs' shareholders as well as policymakers to further improve firm performance, thus promoting China's economic growth. For instance, Chinese SMEs may optimize and improve corporate governance furtherly, thereby reducing the CEOs' improper behaviours that can harm owners' benefits.

The remainder of this paper proceeds as follows: Section 2 describes our empirical methodology and Section 3 presents the dataset and all the variables used in the econometric model. In Section 4 we present the main results of the paper. Some concluding remarks are offered in the last section.

Section snippets

Empirical model and methodology

To investigate the relationship between CEO power and capital structure, we start with the traditional linear regression model which can be briefly described as follows:yit=βPOWERit+λXit+εitwhere, yit represents our proxy for corporate capital structure, POWERit is the CEO's decision-making power in the firm, X is a set of other explanatory regressors (control variables) while εit is an error term, i = 1, …, N represents the firm and t = 1, …, T represents the time. In order to examine whether

Sample selection

In this study, we choose to focus on China's SMEs. There are several reasons why we do so. First, as aforementioned, the great success of Chinese economic development is driven primarily by SMEs which account for the majority of all firms and contribute to the growth of GDP, employment opportunities and fiscal revenues more heavily than large-size firms. Second, private owned firms account for most of Chinese SMEs, which prefer to appoint top executives based on their ability and performance

Descriptive statistics and correlation matrix

Table 3 summarises the key descriptive statistics over the sample period. We observe that the average book leverage for Chinese listed SMEs is 36.8% over the period 2009 to 2013, which is similar to that of the Chinese companies reported in Huang and Song (2006). As the indicators obtained from PCA are normalized, the mean value of CEO Power is 0. Table 3 further presents the descriptive statistics of firm-level control variables. The reported results are comparable to those described in prior

Conclusion

The main objective of this paper is to re-examine the non-linear relationship between CEO power and firm leverage using 231 Chinese SMEs with 1155 firm-year observations over the period of 2009–2013. Previously Chintrakarn et al. (2014) suggest that the effect of CEO power on firm leverage varies across firms with different levels of CEO power. To this end, in the present study we apply the Kremer et al. (2013) dynamic panel threshold model. The main advantages of this methodology can be

References (58)

  • S.H. Law et al.

    Does too much finance harm economic growth?

    Journal of Banking & Finance

    (2014)
  • Y. Liu et al.

    The impact of CEO power on bond ratings and yields

    Journal of Empirical Finance

    (2010)
  • M. Mateev et al.

    On the determinants of SME capital structure in Central and Eastern Europe: A dynamic panel analysis

    Research in International Business and Finance

    (2013)
  • S. Pathan

    Strong boards, CEO power and bank risk-taking

    Journal of Banking and Finance

    (2009)
  • S. Ren et al.

    How do marketing, research and development capabilities, and degree of internationalization synergistically affect the innovation performance of small and medium-sized enterprises (SMEs)? A panel data study of Chinese SMEs

    International Business Review

    (2015)
  • J. Sun et al.

    Ownership, capital structure and financing decision: Evidence from the UK

    British Accounting Review

    (2016)
  • H.I. Ting

    CEO turnover and shareholder wealth: Evidence from CEO power in Taiwan

    Journal of Business Research

    (2013)
  • E. Veprauskaitė et al.

    Do powerful chief executives influence the financial performance of UK firms?

    British Accounting Review

    (2013)
  • R.B. Adams et al.

    Powerful CEOs and their impact on corporate performance

    Review of Financial Studies

    (2005)
  • E.I. Altman et al.

    Corporate financial distress diagnosis in China

    (2007)
  • M.S. Bartlett

    A note on the multiplying factors for various χ2 approximations

    Journal of the Royal Statistical Society

    (1954)
  • P. Berger et al.

    Managerial entrenchment and capital structure decisions

    Journal of Finance

    (1997)
  • A.A. Berle et al.

    The modern corporation and private property

    (1932)
  • M. Bertrand et al.

    Managing with style: The effect of managers on firm policies

    Quarterly Journal of Economics

    (2003)
  • A. Bick

    Pitfalls in panel threshold models: The role of regime dependent intercepts

    (2007)
  • T.J. Brailsford et al.

    On the relation between ownership structure and capital structure

    Journal of Accounting and Finance

    (2002)
  • M. Caner et al.

    Instrumental variable estimation of a threshold model

    Econometric Theory

    (2004)
  • J. Chen et al.

    What determine firms' capital structure in China?

    Managerial Finance

    (2014)
  • P. Chintrakarn et al.

    Powerful CEOs and capital structure decisions: Evidence from the CEO pay slice (CPS)

    Applied Economics Letters

    (2014)
  • Cited by (27)

    • Powerful female CEOs and the capital structure of firms

      2024, Journal of Behavioral and Experimental Finance
    • CEO succession and corporate innovation: A managerial myopic perspective

      2023, North American Journal of Economics and Finance
    • Impacts of social trust on corporate leverage: Evidence from China

      2022, International Review of Economics and Finance
    View all citing articles on Scopus
    View full text