Adjustment speed of capital structure: A literature survey of empirical research

2.1. Theoretical overview

Since the seminal work by Modigliani and Miller (1958), the extensive literature on the capital structure area largely focuses on testing the static and dynamic trade-off theories (Dang et al., 2019). The static trade-off theory suggests the existence of an optimal capital structure that maximizes firm value, balancing tax benefits of debt financing against costs of financial distress resulting from a high debt level (Ahmad and Etudaiye-Muhtar, 2017; Jensen and Meckling, 1976). However, shocks to cash flows and stock prices frequently cause firms to deviate from their target leverage level (Chang et al., 2014; Ramalingegowda and Yu, 2018). Since being either under-leveraged or over-leveraged is inconsistent with firm value maximization, firms have incentives to move back to the target leverage level using transactions to raise or reduce capital sources (Cook and Tang, 2010; Dang et al., 2019).

Prior studies generally find unexpectedly low SOAs (Fama and French, 2002; Ramalingegowda and Yu, 2018) as leverage adjustments are not costless in a dynamic setting (Devos et al., 2017). Accordingly, the subsequent dynamic trade-off theory suggests the presence of a trade-off between costs as a result of deviation from the optimal leverage ratio and adjustment costs (An et al., 2015; Dang et al., 2019). Most empirical studies attribute incomplete capital structure rebalancing to leverage adjustment costs, namely, transaction costs or new security issuance and repurchase costs, which are related to information asymmetry and agency problems between managers and outside investors (An et al., 2015; Flannery and Rangan, 2006; Öztekin and Flannery, 2012).

However, no single theory can fully explain capital structure choices adopted by corporate entities (Byoun, 2008; Huang and Ritter, 2009). Although neither the pecking order theory nor the market timing theory predict the existence of target leverage ratios and adjustment towards these targets, these theories may be relevant in explaining managerial SOA decision-making (Byoun, 2008). Under the pecking order theory, adverse selection costs of issuing risky securities, because of information asymmetry or managerial optimism, lead to a preference for internal funds over costly external financing (Byoun, 2008; Myers and Majluf, 1984; Öztekin, 2015). Prior studies (Byoun, 2008; Leary and Roberts, 2005) argue that firms may have optimal leverage levels and simultaneously prefer internal funds in the presence of significant adverse selection or transaction costs. Accordingly, adjustments back to target leverage ratios likely occur when firms face cash flow imbalances (surpluses or deficits) depending on whether leverage ratios are above or below targets (Byoun, 2008; Devos et al., 2017).

Meanwhile, the market timing theory posits that firms consider the time-varying relative costs of issuing securities when making capital structure decisions (Baker and Wurgler, 2002; Dang et al., 2014; Öztekin, 2015). Accordingly, firms will alter their leverage position to exploit favourable pricing opportunities (Öztekin, 2015). Previous studies (Devos et al., 2017; Flannery and Rangan, 2006; Warr et al., 2012) suggest that different market timing can affect leverage adjustment costs. Thus, over- or under-valuation of firm equity will lead to differences in SOA (Flannery and Rangan, 2006; Frank and Goyal, 2004; Warr et al., 2012). Specifically, firms with over-valued (under-valued) equity are more likely to issue equity (debt) instead of debt (equity), thereby preventing (expediating) firms moving back to the optimal leverage level when they have below-target leverage ratios.

2.2. Models

2.2.1. Standard partial adjustment model

Due to the presence of leverage adjustment costs, firms may not fully adjust their capital structure towards the target (An et al., 2015; Öztekin and Flannery, 2012). Therefore, most previous studies estimate the SOA using the standard partial adjustment model as follows

( L i , t − L i , t − 1 ) = α 0 + λ ( L i , t * − L i , t − 1 ) + ε i , t

(1)

where λ is the magnitude of the SOA; L i , t , L i , t − 1 are actual leverage ratios of firm i in year t and t–1, respectively; L i , t * is the target leverage ratio for firm i in year t; and ε i , t is the error term.

The target leverage ratio L i , t * in equation (1) is not directly observed. It is typically estimated as a function of several factors as shown below

L i , t * = α F X i , t − 1 + α M F i r m + α N Y e a r

(2)

where X i , t − 1 is a vector of firm characteristics for firm i in year t–1. Firm is a vector of indicator variables capturing firm fixed effects. Year is a vector of indicator variables capturing annual fixed effects.

Substituting the target leverage ratio L i , t * of equation (2) into equation (1) yields the following dynamic panel model

L i , t = α 0 + ( 1 − λ ) L i , t − 1 + ( α F λ ) X i , t − 1 + ( α M λ ) F i r m + ( α N λ ) Y e a r + ε i , t

(3)

Then, determinants of SOA are incorporated into equation (3) as shown below

L i , t = α 0 + ( 1 − λ ) L i , t − 1 + ( α F λ ) X i , t − 1 + ( α M λ ) F i r m + ( α N λ ) Y e a r + ( γ K λ ) Z i , t − 1 + ε i , t

(4)

where Z i , t − 1 is the vector of SOA determinants. According to Flannery and Hankins (2013), ordinary least squares (OLS) estimates of the lagged dependent variable’s coefficient in a dynamic panel model are biased due to correlation with fixed effects. To correct estimation biases in the dynamic panel model, most previous studies apply the generalized method of moments (GMM) approach proposed by Blundell and Bond (1998) for the estimation of equation (4).

2.2.2. Modified partial adjustment model

Equation (4) assumes that all firms adjust their capital structure towards the optimal level at the same rate (λ) in all time periods. Recent studies have examined cross-section variations in SOA by employing the modified partial adjustment model because it is flexible and allows SOA to depend on the firm’s specific conditions (An et al., 2015; Dang et al., 2019; Öztekin and Flannery, 2012). Rather than estimating the target leverage ratio L i , t * through a static model in equation (2), these studies estimate the dynamic panel model as follows

L i , t = α 0 + ( 1 − λ ) L i , t − 1 + ( α F λ ) X i , t − 1 + ( α M λ ) F i r m + ( α N λ ) Y e a r + ε i , t

(5)

The estimation of the model in equation (5) will provide an initial set of estimated λ, α F , α M and α N , which are used to calculate the target leverage ratio L i , t * in equation (2). Then, a model is developed by regressing the SOA parameter against a variety of attributes that are hypothesized to be related to SOA as follows

λ i , t = γ K Z i , t − 1

(6)

Substituting equation (6) into equation (1) yields the modified partial adjustment model as shown below

( L i , t − L i , t − 1 ) = α 0 + ( γ K Z i , t − 1 ) ( L i , t * − L i , t − 1 ) + ε i , t

(7)

3.1. Fundamental and operational characteristics

Due to differences in leverage adjustment costs and benefits, SOA is heterogeneous across firms (Dang et al., 2012; Elsas and Florysiak, 2011; Zhou et al., 2016). Previous studies typically condition the SOA on a set of firm characteristics (Elsas and Florysiak, 2011; Fitzgerald and Ryan, 2019). Importantly, most selected firm-specific characteristics reflect fundamental or operational determinants of a firm’s capital structure itself, because it is logical that factors determining target leverage of a firm also determine the economic pressure to stay close to this target (Elsas and Florysiak, 2011).

Elsas and Florysiak (2011) conduct an extensive analysis to explore cross-sectional heterogeneity in SOA employing a set of firm fundamental characteristics for listed US firms. They provide evidence for pronounced heterogeneity, where SOA is faster for firms with high financing deficits, for smaller firms due to higher opportunity costs of deviating from target, when deviations from target leverage are large and for firms in financial distress. In contrast, González and González (2011) observe no differences in SOA between small and large Spanish firms. However, as per the study by Aybar-Arias et al. (2012), the degree of financial flexibility, growth opportunities and firm size are positively related to SOA, whereas the distance to the optimal debt ratio has a negative impact on SOA. Recently, Aderajew et al. (2019) examined 1500 Dutch farms (as distinct from standard firms) and found that SOA is slow and varies according to farm size and type. More specifically, SOA is relatively faster for horticulture farms and slower for livestock farms.

By developing a dynamic panel threshold model of capital structure, Dang et al. (2012) show that UK firms with large financing imbalances, large capital expenditures or low earnings volatility have a significantly faster SOA than those with the opposite characteristics. Fitzgerald and Ryan (2019) found that openly held firms adjust back to the target leverage level faster than closely held firms. Their results further suggest that smaller, higher growth and low dividend-paying firms adjust towards the target level faster than their larger, lower growth and high dividend-paying counterparts. Similar results are observed in the study by Fama and French (2002), who confirm that dividend payers have a slower SOA than non-payers. Specifically, they suggest that SOA is 7%–10% per year for dividend payers and 15%–18% per year for non-payers.

Differing with studies discussed above, Mukherjee and Wang (2013) focus on the leverage adjustment benefits and postulate that SOA is positively associated with the starting leverage deviation. The marginal tax shield is a decreasing function, while the marginal bankruptcy cost is an increasing function, of a firm’s leverage ratio. The net benefit of rebalancing expands at an increasing rate as the firm’s leverage moves farther away from the target. Consequently, the greater the distance from the target leverage ratio, the bigger the incentive to adjust and the faster the SOA. They indicate that SOA is positively related to the deviation from the target leverage level and the SOA sensitivity to leverage deviation depends on whether the current capital structure is below or above the target.

SOA asymmetry is also determined by the firm’s cash flow situation (Byoun, 2008; Faulkender et al., 2012; Zhang et al., 2020a). Particularly, Byoun (2008) expresses a need to incorporate the pecking order and trade-off behaviours in examining capital structure decision-making. Due to the preference for internal funds, adjustments back to target leverage likely occur when firms face imbalances in cash flows. Byoun (2008) documents that SOA is faster when firms face financial surpluses with above-target leverage ratios or financial deficits with below-target leverage ratios. However, the Byoun (2008) findings depend on an implicit assumption that firms adjust towards the target leverage by making changes in debt ratios. Dang and Garrett (2015) provide new insights into leverage adjustment mechanisms by analytically deriving the difference in SOA when leverage adjustments are restricted to changes in debt, and when they include changes in debt, equity and total assets. Their results indicate that the two approaches only deliver the same SOA under restrictive scenarios and over-levered firms with a financing deficit adjust towards their target leverage faster. Smith et al. (2015) extend Byoun’s (2008) model about the effect of financial deficits and surpluses on SOA to demonstrate how industry characteristics identified by Kayo and Kimura (2011), including industry concentration, industry munificence and industry dynamism, affect SOA heterogeneity. They find that New Zealand firms operating in less concentrated, highly munificent or highly dynamic industries and whose debt is well above target, adjust towards their target leverage faster.

Faulkender et al. (2012) recognize that cash flow realizations can provide opportunities for firms to adjust leverage at relatively low marginal cost, regardless of whether their transaction costs are high or low. As such, they show that firms with high absolute cash flows and high absolute leverage deviations make larger capital structure adjustments than firms with similar leverage deviations but cash flow realizations near zero. They also provide results about the impact of market conditions on SOA with under-levered firms moving less quickly towards higher target leverage levels when interest rates are high. Conversely, over-levered firms appear to increase their SOA significantly in the presence of higher equity valuations. Recently, Zhang et al. (2020a) demonstrated similar findings that Chinese firms with a larger absolute cash flow adjust towards their leverage targets significantly faster than those with a smaller absolute cash flow.

In a notable study, Zhou et al. (2016) exploit the leverage deviation to investigate capital structure dynamics from the perspective of the ex-ante cost of equity capital. They document that firms with costs of equity that are more sensitive to leverage deviation adjust towards the optimal capital structure faster. Rather than estimating SOA heterogeneity across firms, Elsas et al. (2014) examine whether managerial SOA decision-making is associated with major investments accompanied by external financing. They find that security types issued to finance a large investment significantly depend on the distance from the target leverage. Over-leveraged firms issue less debt and more equity when financing large projects, and vice versa. As such, they provide evidence that firms making large investments converge rapidly back to the optimal capital structure.

Previous studies (Leary and Roberts, 2005; Nguyen et al., 2021a) also document SOA heterogeneity across leverage levels. In particular, Leary and Roberts (2005) employ simulations that allow the dynamic capital structure to exhibit three scenarios of adjustment costs: fixed cost, proportional cost and fixed cost together with a weakly convex component. Through examining each scenario, they find that firms will raise (reduce) debt when their leverage is relatively low (high). Although Leary and Roberts (2005) document SOA heterogeneity across several leverage levels, the discrete segmentations used in their study uncover limited results. More importantly, the study does not test for SOA asymmetry corresponding to low- and high-levered firms.

Instead of focusing on a range for the target leverage, Nguyen et al. (2021a) analyse actual leverage to reveal a smooth pattern of SOA from low-levered firms to high-levered firms. They show that when total leverage and long-term leverage are considered, low- and high-levered firms are associated with a higher degree of SOA than are medium-levered firms. Furthermore, there is a difference in SOA between low- and high-levered firms, which points to SOA skewness. However, when short-term leverage is considered in the analysis, SOA becomes smaller at varying levels of short-term debt.

Although the literature on fundamental and operational factors driving SOA is extensive, previous studies often provide mixed results. Given the mixed conclusions in the literature, we call for more research to better understand how these characteristics affect SOA. In addition, it is surprising that research questions in most surveyed studies are pursued in only one country. While single market studies reduce econometric and estimation concerns related to correlation, endogeneity and sample selection bias, the use of a specific country context also comes with costs. For example, one cost is the limited generalizability of findings in these studies. Therefore, future research might examine interesting comparisons and cross-country predictions ² with relatively similar institutional settings, such as region or trade-bloc level settings.

3.2. Financial reporting and managerial incentives

The main objective of financial reports is to ‘provide financial information ³ about the reporting entity that is useful to existing and potential investors, lenders and other creditors in making decisions about providing resources to the entity’ (FASB, 2010). Therefore, it is logical that properties of the financial reporting system may affect firm capital structure decisions. Recently, Ramalingegowda and Yu (2018) employed accounting conservatism as an attribute of the financial reporting system to examine its impact on managerial SOA decision-making and documented that firms with more conservative financial reporting adjust their capital structure towards the target more quickly. The positive impact of accounting conservatism on SOA is concentrated in under-levered firms, which occurs through conservatism facilitating debt issuance.

Another important dimension of financial reporting is its readability. Despite its paramount importance, information in financial statements is plagued with readability concerns (Habib et al., 2018). As managers can obfuscate the quality of financial reports by making it harder for outside investors to understand and infer the future cash flow implications of current accounting information (Biddle et al., 2009), the lack of annual report readability reduces financial reporting quality (Lo et al., 2017), thereby increasing external financing costs (Rjiba et al., 2021). As such, financial report readability potentially plays a significant role in reducing leverage adjustment costs associated with information asymmetry and financing frictions. Thus, an avenue for future research would be to investigate whether the readability of financial reports is associated with SOA outcomes.

According to Hail et al. (2010), a key driver of financial reporting quality is managerial reporting incentives, which are influenced by several factors such as firm characteristics, corporate governance structure, managerial incentive-based compensation, product market competition, the strength of the enforcement regime and a country’s legal institutions. Brisker and Wang (2017) investigate the impact of one form of managerial incentive-based compensation, chief executive officer (CEO) inside debt, on SOA. Offering managers debt-based compensation exacerbates managerial risk-averse behaviours (Brisker and Wang, 2017). Liao et al. (2015) suggest that excessively risk-averse managers tend to avoid lifting the debt ratio to the level that shareholders desire. Consistent with these arguments, Brisker and Wang (2017) provide empirical evidence that a higher CEO inside debt ratio is associated with faster (slower) leverage adjustments towards the shareholders’ desired level for over-levered (under-levered) firms.

In contrast, offering managers equity-type compensation encourages risk taking and mitigates managerial conservatism (Brisker and Wang, 2017). After several high-profile accounting scandals in the early 2000s, a particular concern of regulators and investors is that stock-based compensation might induce managers to increase the short-term stock price through earnings management (Cheng and Warfield, 2005). An avenue for future research is to examine the association between capital structure SOA and the nature or extent of managerial equity-based compensation. However, if managerial incentive-based compensation discourages adjustments towards the target leverage, will compensation committees consider the re-design of executive compensation packages? Although this is not a simple question because reducing equity-based compensation could also motivate managers into forgoing risky but positive net present value (NPV) projects, for instance. Future research might consider addressing these issues.

Managerial personality traits or styles have recently been found to have significant effects on capital structure SOA. In particular, Lartey et al. (2020) find that firms managed by extrovert CEOs adjust towards the target leverage level faster. The positive relation between extraversion and SOA is enhanced for firms that are larger, have greater collateralizable assets and are more vulnerable to external shocks. As identified in prior research (Malmendier et al., 2011), CEO overconfidence and formative early-life experiences also play an important role in explaining observed variations in capital structure. A potential extension, therefore, is to examine the impact of such CEO characteristics on SOA. We further encourage future studies to investigate how other CEO characteristics, including CEO tenure, age, gender and education, also influence a firm’s SOA decision-making. Prior studies find that the board of directors and other members of the top management team such as CFOs also play significant roles in capital structure decision-making (Alves et al., 2015; Graham and Harvey, 2001). Therefore, future research can consider how company directors, the chief financial officer (CFO) and broader top management team, instead of only the CEO, also affect SOA decisions.

3.3. Corporate governance and monitoring structures

Corporate governance is a framework to build an environment of accountability, trust and transparency (Detthamrong et al., 2017). The extant literature suggests that corporate governance is an important tool to reduce the agency problems between managers and outside investors (Chang et al., 2014; Detthamrong et al., 2017; Liao et al., 2015), which are associated with distortions in corporate policy choices and lower firm performance (Morellec et al., 2012). Because debt limits managerial flexibility (Jensen, 1986), self-interested managers may not make capital structure decisions that maximize shareholder wealth. Morellec et al. (2012) suggest that an effective corporate governance system can discipline managers to use more debt and make more timely leverage adjustments towards the target level.

Complementing the theoretical findings of Morellec et al. (2012) and Chang et al. (2014) provide empirical evidence that both the disciplinary and takeover defence roles of debt provide motivations for managers to adjust firm leverage. They find that both over-levered and under-levered firms with weak governance adjust slowly towards their target debt levels, though with different motivations. Extending the study of Chang et al. (2014, 2015), show that product market competition increases the incentives for firms with weak governance structures to maximize the wealth of shareholders, thereby increasing SOA. The difference in SOA between firms with weak and strong governance structures is found to be smaller among firms operating in highly competitive industries.

Liao et al. (2015) find that a faster SOA is associated with better corporate governance quality defined by a more independent board featuring CEO–chairman separation and greater presence of outside directors, coupled with a larger institutional shareholding influence. Conversely, managerial incentive-based compensation discourages adjustment towards the target leverage level. Nguyen et al. (2021b) argue that corporate governance in emerging markets plays a vital role in alleviating agency problems and severe information asymmetry because legal systems, the rule of law and investor protection mechanisms are not effective. Using data from Vietnamese stock markets, they demonstrate that board size, board independence, gender diversity and managerial ownership increase SOA, whereas CEO duality decreases SOA.

Some other forms of ownership are also documented to be associated with SOA heterogeneity (An et al., 2021; Do et al., 2020; Pindado et al., 2015). Particularly, focusing on a selection of countries that are part of the Eurozone (Austria, Belgium, Germany, Spain, Finland, France, Ireland, Italy and Portugal), Pindado et al. (2015) find that family firms rebalance their capital structure faster than other firms. Family firms’ faster SOA is mainly driven by companies in which the family participates in management and by family businesses that are still in the first generation. Do et al. (2020) reveal that foreign investors in Taiwanese firms serve as a direct substitute for debt by enhancing corporate governance. As such, foreign investors help reduce leverage adjustment costs, which in turn increases SOA. Employing a large sample of 7246 firms across 38 countries from 2000 to 2013, An et al. (2021) provide similar evidence that foreign institutional ownership is positively related to firms’ SOA. This positive relation is concentrated for over-leveraged firms that need to decrease financial leverage to converge back to the target.

At the macro-level, regulatory oversight can also be considered as an effective monitoring mechanism. After several failed attempts, the China Securities Regulatory Commission (CSRC) launched the Split Share Structure Reform (SSSR) in April 2005 to enhance corporate governance by reducing agency problems between state owners and non-state shareholders via conversion of state-owned non-tradable shares into tradable shares, making state-owned shares sensitive to the stock market mechanism. He and Kyaw (2018) argue that this reform provides an excellent laboratory to re-examine capital structure dynamics in China and their evidence supports that SOA increases across all firms after the SSSR. Given that news media coverage can serve as an external corporate governance mechanism, Dang et al. (2019) analyse global news across 33 countries and suggest that greater news coverage and more positive news sentiment are associated with quicker SOA.

The association between corporate governance and SOA is only beginning to receive extensive attention from academics. As discussed above, there is some evidence that the existence of stronger governance and monitoring mechanisms is associated with an improvement in SOA decision-making. However, there is still a need for more research in this area to address different types of governance structures. In particular, the extant literature suggests that external auditors and financial analysts play important roles in mitigating agency costs arising from conflicts of interest between shareholders and managers (Fan and Wong, 2005; Habib et al., 2018; Kim et al., 2015). In addition, Yuan et al. (2016) argue that director and officer liability (D&O) insurance can serve as an effective external monitoring mechanism, which can improve corporate governance at the firm-level and investor protection at the country-level. Thus, it would be interesting for future research to investigate whether SOA is associated with other corporate governance and monitoring mechanisms including D&O liability insurance, external auditing and analyst coverage. It may also be relevant to consider the role of different governance systems, such as bank-based governance relationships in countries such as Japan and Germany, on SOA compared with that for more common-law country governance structures. Due to the relatively recent implementation of new corporate governance codes in many countries, regulatory changes in the environment may provide opportunities and natural experiment settings for future research to examine how corporate governance reforms are associated with firm capital structure SOA decisions on a longitudinal basis.

3.4. Informal institutions

The preceding section reviewed empirical studies on the association between capital structure SOA and formal governance and monitoring mechanisms. In this section, we survey empirical studies investigating the impact of informal institutions on SOA. Informal institutions refer to firm-specific norms and values, management ethos, codes of conduct in business, general norms and values in society and firm reputation from the perspective of competitors, suppliers and customers (Habib et al., 2018). Huang et al. (2021a) recently investigated the impact of trust on SOA heterogeneity and found that social trust has a positive effect on SOA. Furthermore, their results suggest that the positive effect of social trust on SOA is more pronounced for over-levered firms, firms with higher information asymmetry, firms with reduced ease of financing and firms located in countries with weaker governance quality.

Hogg and Abrams (1988) suggest that the value in sharing an identity and having a sense of being in a particular group has substantial influence on people’s behaviours. A growing literature empirically examines the impact of religious beliefs, an important aspect of culture, on a wide range of corporate decisions, namely, tax avoidance (McGuire et al., 2012), investment decisions (Hilary and Hui, 2009) and capital structure decisions (Alalmai et al., 2020). To extend this strand of research, future work might investigate the impact of religion on SOA decision-making. From a conceptual perspective, both religion and trust are embedded within the broader concept of social capital (Habib et al., 2018). These two positive attributes are in contrast to a culture of corruption, which adversely affects social capital. Thus, another avenue for future research could be to examine whether SOA is associated with corruption, which undermines the functioning of formal institutions.

Habib et al. (2018) argue that the role of informal institutions is of special significance in emerging economies, where formal institutions such as investor protection, accounting standards and corporate governance are weak. Accordingly, Li et al. (2019) believe that China, a transition economy with developing legal systems and relatively weak corporate governance, provides a suitable setting to examine whether the network of interlocked directors is associated with capital structure SOA. For the Chinese market environment, they document that firms in the central position of the director network have a faster SOA. Similarly, Li et al. (2017) identify that bank connections in Chinese-listed firms established through personal networks affect SOA. Specifically, when a firm’s leverage is below the optimal level, firms with bank connections adjust their leverage by 13% annually towards the target, compared with 9.2% for firms without bank connections.

In another study, Alnori and Alqahtani (2019) investigate the impact of sharia compliance status on SOA. Firms complying with sharia are subject to several restrictions stemming solely from three sources, namely, the Qur’an, Sunnah and Ijtihad. These restrictions exist at multiple levels, including financing, investments, transactions and risk management tools, but do not constrain non-sharia-compliant firms. As such, sharia-compliant firms may incur greater expected bankruptcy costs of debt and transaction costs. Using a sample of Saudi Arabian non-financial firms, Alnori and Alqahtani (2019) indicate that sharia-compliant firms have significantly slower SOAs.

Meanwhile, there is ongoing debate whether political connections are beneficial or detrimental to shareholders’ wealth (Habib et al., 2018). In particular, Fisman (2001) provides evidence that political connections increase firm value, among many other benefits. Conversely, politically connected firms usually undertake high rent-seeking activities that are harmful to minority interests (Faccio, 2006). Given competing views about the consequences of political connections, it would be interesting to investigate the impact of political connections on SOA as there are potential channels such as access to debt markets, bank and other forms of finance linked to the political sphere.

3.5. Financial market attributes

Stock liquidity is considered as one of the most essential characteristics of financial market efficiency (Atawnah et al., 2018; Nguyen and Muniandy, 2021). Weston et al. (2005) reveal that stock liquidity significantly influences costs of securities issuance. Accordingly, Ho et al. (2020) find that high-liquidity firms have a significantly faster SOA than less-liquid firms. However, the positive effect of liquidity on SOA exists only for over-levered firms and is less pronounced in strong institutional environments. Tekin (2020) investigates how market differences affect SOA. Using a quasi-natural experiment setting based on regulatory and institutional differences across the highly regulated Main market (MAIN) and the comparatively unregulated Alternative Investment Market (AIM) of the London Stock Exchange, he finds that firms listed on the AIM have a faster SOA than firms trading on the MAIN before the global financial crisis. However, the relationship is reversed after the global financial crisis. In a similar vein, Dang et al. (2014) demonstrate a negative impact of the global financial crisis on capital structure SOA. Over the pre-crisis period, more constrained firms with high growth, larger investment requirements, smaller size and volatile earnings, adjust their capital structures more quickly than their less constrained counterparts.

Meanwhile, the market timing theory posits that managers can minimize the cost of capital by timing the market (Baker and Wurgler, 2002). Accordingly, firms should take into account the time-varying relative costs of issuing securities when making capital structure decisions (Dang et al., 2014; Öztekin, 2015). Warr et al. (2012) argue that different market timing can affect leverage adjustment costs and over- or under-valuation of firm equity, which will lead to SOA heterogeneity. As expected, they find that over-levered firms adjust more rapidly towards their target when their equity is over-valued. However, when a firm is under-valued and needs to reduce leverage, SOA is much slower. In addition, Abdeljawad and Nor (2017) show that when the timing variable is accounted for, SOA is significantly higher and the timing role is lower for over-leveraged firms compared with under-leveraged firms.

Another strand of research investigates how credit market conditions and supply-side factors affect SOA. In particular, DeAngelo et al. (2011) demonstrate that firms prefer to operate under-levered to preserve debt capacity to fund the uncertain arrival of investment opportunities, at which point they might use ‘transitory’ debt to fund investment shocks and rebalance to the target leverage level in future years. As a credit line is the debt market’s solution to the borrower’s demand for financial flexibility, Lockhart (2014) extends the study of DeAngelo et al. (2011) and investigates the impact of credit lines on SOA. Their results suggest that the availability of credit lines is associated with cross-sectional variations in estimated SOA. Devos et al. (2017) examines the impact of debt covenants on SOA. They provide evidence that covenants lower the SOA by 10%–13%, relative to the SOA of firms without covenants. SOA is significantly lower, by 40%–50%, for firms with the most intense covenant provisions. However, SOA is reduced more for strict capital covenants than for strict performance covenants. Nieto and Rodriguez (2015) analyse the effects of dynamic correlations between the returns on stock and bonds issued by the same firm on SOA. They reveal that SOA is positively and significantly related to the stock–bond correlation.

In the United States, central to the credit supply chain are commercial banks, which are the largest source of external financing for corporate borrowers. Using the staggered deregulation of the US banking industry as an exogenous shock, Rahman (2020) indicates that interstate and intrastate banking deregulation are positively associated with SOA. These positive effects are stronger for firms that are financially constrained, are financially dependent on banks and have less access to the public debt market and by deregulated banks’ ability to geographically diversify credit risks. Using the boom-bust cycle from 1987 to 2014 in Japan as a bank loan supply shock, Shikimi (2020) shows that financially constrained firms adjust their capital structure towards the target level slower during credit-crunch periods than during other periods. Liu et al. (2020) use the Chinese Lending Guideline 2007 as a quasi-natural experiment to study the impact of short-term lending regulation shocks on SOA and reveal that SOA is slower after the rollover restrictions were introduced.

Credit ratings, which constitute a proxy for the probability of default and play an important role in financial markets, have been identified as a strong determinant of SOA heterogeneity. In particular, survey results in the study by Graham and Harvey (2001) indicate that CFOs focus on credit ratings to guide debt decisions. Credit ratings enable markets and investors to set the required rate of return in line with the level of default risk carried by the rated entity or financial security (Ferri et al., 1999), thereby affecting the access to and costs of borrowed funds (Gu et al., 2018). Consistent with this argument, Kisgen (2009) observes that firms adjust their capital structure towards the target faster after being downgraded, whereas rating upgrades have no association with SOA. Extending the study of Kisgen (2009), Huang and Shen (2015) demonstrate that firms adjust their capital structure towards the optimal level faster in countries with better financial development and strong legal and institutional environments than in weaker equivalents, regardless of the upgraded and downgraded rating changes. Wojewodzki et al. (2018) focus on credit rating levels and provide empirical evidence that firms with poorer credit ratings converge back to the target leverage significantly faster than firms with better credit ratings. In contrast, Wojewodzki et al. (2020) focus on financial institutions and find that credit ratings have relatively little economic effect on the speed at which banks’ capital structure is adjusted towards the optimal level.

Jiang et al. (2017) use Chinese data, where bank concentration varies across both years and provinces, and find that under-levered firms adjust back to the target leverage faster when bank competition is high. They additionally find that smaller firms and non-state-owned firms exhibit a faster SOA when bank competition is high. Im (2019) indicates that over-levered firms’ SOA and peer firm shocks have a U-shaped relationship, while under-levered firms’ SOA monotonically increases with peer firm shocks. An et al. (2015) show that firms with higher crash-risk exposure tend to adjust more slowly towards the target leverage position. Their results further indicate that the negative relation between crash-risk exposure and SOA is less pronounced in countries with a more transparent information environment.

The effects of mergers and acquisitions (M&A) have also been analysed as an influential factor for SOA. Uysal (2011) demonstrates that managers take deviations from their target capital structures into account when planning and structuring acquisitions. A higher likelihood of forgoing valuable acquisition opportunities could generate quicker leverage adjustments for over-levered firms. Khoo et al. (2017) argue that firms engaging in M&A activity deviate from their target leverage levels due to the financing transactions necessitated by the acquisition. Empirically, they find that acquirers engaging in a leverage-increasing (leverage-decreasing) transaction subsequently decrease (increase) their leverage ratios. Acquirers, regardless of the way in which they finance acquisitions, actively rebalance their leverage by incorporating 22% to 62% of the leverage deviation in the acquisition year.

3.6. Economy-wide attributes

According to Haas and Peeters (2006), financial development can increase the capital stock and productivity, enhancing economic development. To obtain a better understanding of the quantitative and qualitative development of financial systems, they analyse a sample of firms in 10 Central and Eastern European countries from 1993 to 2001 and show that the gradual development of financial systems enables firms to bring their actual capital structure closer to the target structure. Similar results are observed by Antoniou et al. (2008), who confirm that the speed at which firms adjust their capital structure crucially depends on the financial system and corporate governance traditions of each country. Öztekin and Flannery (2012) show that firms from countries with better functioning financial systems, a financial structure based on the effectiveness of capital markets instead of intermediaries and strong legal institutions adjust towards their target leverage as much as 50% more rapidly. Regarding financial institutions, Jonghe and Öztekin (2015) demonstrate that banks have a faster SOA in countries with more developed capital markets, more stringent capital requirements, better supervisory monitoring and higher inflation. In times of crises, banks adjust their capital structure more quickly.

Similarly, Drobetz et al. (2015) find that firms in market-based countries rebalance faster after leverage shocks than firms in bank-based countries. In addition, their results reveal that firms adjust towards the target leverage level more slowly during bad macroeconomic states and the business cycle effect is more pronounced for financially constrained firms in market-based countries. Also, Cook and Tang (2010) demonstrate that firms tend to adjust faster towards the target leverage level in good macroeconomic states embodying firms’ favourable access to capital markets, as defined by term spread, default spread, GDP growth rate and dividend yield. Differentiating between financially unconstrained and constrained firms, their results still exhibit a faster SOA in good states compared to bad states, regardless of firms’ access to external capital markets.

Meanwhile, Baum et al. (2017) document that over-levered firms converge back to the optimal capital structure more rapidly when macroeconomic risk is low. In contrast, under-levered firms adjust more quickly towards the target leverage in times of low firm-specific risk and high macroeconomic risk. Ahsan and Qureshi (2017) document a slower SOA towards the short-term target capital structure during a post-financial liberalization period as compared to a pre-financial liberalization period. Mai et al. (2017) investigate whether in China, a country large in size, geographically diverse and imbalanced in regional economic development, there is a significant difference in SOA of firms located in different regions during the macroeconomic recovery period. Their results indicate that during the process of economic recovery, there are apparent regional variations in SOA where the fastest adjustment in capital structure is in East China while that of West China followed and that of Mid China was the slowest.