Unraveling the Impact of Product Market Competition and Earnings Volatility on Zero-Leverage Policies

Abstract

This paper investigates the impact of product market competition (PMC) on firms’ adoption of zero-leverage (ZL) strategies and examines whether this relationship is influenced by earnings volatility. Using data from 1989 to 2019, we analyze the PMC and ZL behavior of firms using logistic regression. Our results indicate that as PMC intensifies, firms are more likely to adopt ZL policies. This effect is stronger in firms with higher earnings volatility, suggesting a significant interaction between PMC and earnings volatility in shaping capital structure decisions. This study extends existing research by highlighting the role of earnings volatility in strengthening the relationship between PMC and ZL behavior, offering new insights into the dynamics between market competition, financial decisions, and firm risk.

1. Introduction

Firms’ capital structure decisions have long been a central focus of corporate finance, with traditional theories (e.g., trade-off theory, pecking-order theory) proposing that firms balance the benefits of debt, such as tax shields, against the potential downsides, including financial distress and agency costs (Miller, 1977; Jensen, 1986). However, empirical evidence reveals a persistent divergence between theory and practice, as many firms maintain debt levels well below what these models would predict (Graham, 2000). Particularly, a puzzling phenomenon is the growing prevalence of zero-leverage (ZL) firms, which opt to entirely avoid debt despite the significant tax advantages they forgo.1 This raises an important question: why do some firms choose to adopt a ZL policy, seemingly contradicting the predictions of conventional capital structure theories? Understanding this deviation is critical, as it suggests that the factors driving ZL policies may extend beyond the traditional trade-offs considered in the literature. Despite the substantial body of work on capital structure, there is limited research exploring the influence of product market competition (PMC) on ZL policies. This leads us to the following key research question: how does product market competition influence firms’ adoption of zero-leverage policies? Our study fills this gap in the literature by examining the implications of PMC for ZL policies, with a specific focus on how earnings volatility influences the dynamics of this relation.

The rationale behind our research question is threefold. First, various factors have contributed to the intensification of competition in U.S. product markets, such as antitrust laws, deregulatory initiatives, import competition, and economic globalization (e.g., Bernard et al., 2006; Bloom et al., 2016; Irvine & Pontiff, 2009). These efforts toward creating freer markets aim to enhance economic efficiency and productivity. With the sharp increase in globalization and trade liberalization in recent years, it is crucial to examine the implications of rising competitive pressures. Second, because firms do not operate in isolation, their strategic decisions are influenced by interactions with their rivals in the product market (Frésard & Valta, 2016). Companies in competitive industries must continuously adapt strategies such as investing in innovation, improving product quality, differentiating their offerings, and expanding to strengthen their market position. These strategies often require significant financial resources, suggesting that competition may influence firms’ financing policies, including the choice of a ZL strategy. Third, the effects of competition on firms are complex and multifaceted. On one hand, classical political economists (Caves, 1980; Smith, 1776) argue that competition compels firms to perform at their best to survive. A growing body of research supports this view, showing that competition acts as a powerful disciplinary tool, pushing managers to focus on efficiency and curb self-serving behaviors, thereby promoting better resource allocation and improved performance (e.g., Ammann et al., 2013; Bharath & Hertzel, 2019; Chhaochharia et al., 2017; Hart, 1983; Sassi et al., 2019; Tang, 2018). On the other hand, firms in highly competitive markets must share profit opportunities with rivals, which can increase default risk and threaten long-term survival. In such conditions, firms may adopt more conservative financial policies, such as ZL, to minimize their exposure to financial distress (e.g., MacKay & Phillips, 2005; Hoberg et al., 2014; Xu, 2012). Given that debt can either provide agency benefits (Jensen, 1986) or amplify financial distress (Myers, 1977), the impact of competition on a firm’s decision to adopt a zero-debt policy is far from straightforward.

Following this rationale, we develop two competing views regarding how PMC affects the adoption of ZL policies. The relation could be driven by two opposing mechanisms. The first view, which we label the “disciplinary force” argument, suggests that competition serves as an external disciplinary force. As competition intensifies, entrenched managers will face greater pressure to optimize firm performance, which reduces their incentives to avoid debt and its associated monitoring benefits. As a result, we would expect to see fewer ZL policies in more competitive environments. Previous research, such as that of Strebulaev and Yang (2013), supports this perspective, showing that ZL policies often stem from weak corporate governance, where managers avoid debt to escape external scrutiny. In contrast, the second view, which we label the “risk-increasing” argument, posits that the heightened risks associated with increased competition may prompt managers to adopt more conservative financial policies, such as avoiding debt altogether. This is because taking on debt under competitive pressure could further exacerbate the risk of financial distress. Prior studies, such as Devos et al. (2012), indicate that firms with ZL are often financially constrained, suggesting that in highly competitive markets, ZL may be a strategic response to mitigate the risks of debt obligations. Consequently, as competition intensifies, we may observe more ZL policies under this second view.

To disentangle these two opposing views, we further advance our analysis by examining how earnings volatility interacts with PMC, allowing us to determine which mechanism dominates in shaping firms’ ZL policies. Specifically, if the relationship between PMC and ZL behavior is stronger in firms with high earnings volatility, it would suggest that firms view the heightened risk from competition as a reason to avoid debt, supporting the second view. Conversely, if this relationship weakens with greater volatility, the first view—that competition curbs managerial avoidance of debt—would be more plausible. Thus, earnings volatility serves as a key factor in clarifying the underlying drivers of ZL policies.

To empirically test our hypotheses, we investigate the relationship between firms’ ZL policies and the intensity of PMC using an unbalanced panel of U.S. firms over the 1989–2019 period. Our findings demonstrate that heightened PMC significantly increases the likelihood of firms adopting a ZL policy, lending support to the “risk-increasing” argument of competition’s effect on capital structure decisions. Furthermore, we provide evidence that the interaction between earnings volatility and PMC plays a critical role in shaping firms’ financial decisions. Specifically, our analysis reveals that the positive effect of PMC on ZL adoption is more pronounced among firms with elevated earnings volatility. These results remain consistent across various robustness checks, including the use of alternative measures for both dependent and independent variables, as well as addressing potential endogeneity concerns.

Our paper contributes to existing knowledge in several important ways. First, it advances the understanding of the ZL puzzle by identifying new determinants of this phenomenon. While previous studies have focused on country-level factors, such as creditor protection (Bessler et al., 2013), cultural influences (El Ghoul et al., 2018), and labor laws (Boustanifar & Verriest, 2023), as well as firm-level factors, including financial conservatism (Bigelli et al., 2014), dividend policy (Dang, 2013), and financial flexibility (Huang et al., 2017b), our research extends this body of work by examining the role of industry-level factors. Specifically, we investigate the intensity of PMC and introduce earnings volatility as a moderating factor, shedding light on how these forces shape firms’ ZL decisions. This dual focus on industry competition and earnings volatility provides a clearer understanding of the strategic choices firms make in response to competitive pressures and risk exposure.

Second, our study adds to the industrial organization literature by examining the impact of PMC on corporate decisions and firm-level outcomes, such as corporate disclosure (Li, 2010), earnings management (Lee & Liu, 2016), financial policies (Chi & Su, 2016; Xu, 2012), investment choices (Akdoğu & MacKay, 2012; Frésard & Valta, 2016), employment strategies (Boubaker et al., 2022), firm performance (Dasgupta et al., 2018), and the cost of capital (Sassi et al., 2019). Specifically, our research aligns with the strand of literature that investigates the interaction between product markets and capital structure decisions (e.g., Zheng et al., 2021; Hajda & Nikolov, 2022; Do et al., 2022; Cao & Chen, 2023), but extends beyond the conventional focus on the debt-versus-equity choice (e.g., Xu, 2012) or the selection of debt sources (Boubaker et al., 2018). Instead, it focuses on the ZL phenomenon, a puzzling and less explored aspect of financial policy, offering fresh insights into how competitive forces shape firms’ financing behavior.

More importantly, our findings contribute to the ongoing debate on globalization, particularly in light of the uncertainty surrounding recent trade frictions between the US, China, the EU, and several Asia–Pacific countries (Charoenwong et al., 2020). The study offers valuable insights into how firms strategically adjust their financial policies, such as adopting ZL policies, in response to heightened competitive pressures. For investors, understanding these strategic shifts is essential for assessing firms’ financial health and risk profiles. Firms that adopt ZL policies in response to competitive pressure may signal a more cautious approach to risk management, which could influence investor decisions regarding firm valuation and long-term stability. Policymakers, too, can benefit from these insights, as they seek to develop frameworks that ensure market efficiency while safeguarding firms from excessive financial distress. By recognizing how competitive dynamics shape corporate financial behavior, policymakers can craft regulations and trade policies that balance fostering competition with mitigating the risks firms face in global markets.

The remainder of the paper is organized as follows. In Section 2, we present background literature and develop the main hypotheses. In Section 3, we describe the sample and the main research design. In Section 4, we discuss the empirical results, and in Section 5, we perform additional robustness checks. Finally, in Section 6, we conclude the paper.

2. Literature Review

2.1. Product Market Competition

In recent decades, numerous global policy changes have aimed to ensure sufficient competition and prevent market abuses, giving rise to a significant body of research that highlights the “bright side” of competition. This perspective, rooted in agency theory, argues that product market competition enhances efficiency by mitigating agency conflicts. The key idea is that in competitive environments, managers face greater pressure to perform efficiently—if they fail to do so, they risk being driven out of the market (Hart, 1983; Schmidt, 1997). Additionally, competition facilitates relative performance evaluation (RPE), enabling owners to compare their managers’ performance against industry peers. With more rivals in a competitive industry, owners gain access to information on common industry shocks, allowing them to distinguish between firm-specific issues and broader market conditions, thus making more precise assessments of managerial effort (Holmstrom, 1982; Nalebuff & Stiglitz, 1983). Recent studies further show that competition serves as an external disciplinary mechanism, disciplining managerial behavior (Dasgupta et al., 2018), reducing financing costs (Sassi et al., 2019), lowering monitoring costs like audit fees (Chung & Kim, 2024), and acting as a substitute for other governance mechanisms (Giroud & Mueller, 2010, 2011; Chhaochharia et al., 2017; Boubaker et al., 2018; Bharath & Hertzel, 2019).

Opponents of the “bright side” view argue that PMC has a “dark side”, as it drives down profit margins and increases default risk (Tirole, 2006). This argument is rooted in predation theory, which has two key aspects. First, firms facing intense competition lack the pricing power to pass idiosyncratic shocks onto customers, unlike firms with strong market power. Second, predatory threats from financially stronger rivals in competitive markets can directly undermine weaker firms’ growth prospects. Bolton and Scharfstein (1990) explain that in highly competitive environments, stronger rivals may adopt aggressive pricing or production strategies to inflict losses on weaker firms, potentially driving them out of the market. Studies show that this competitive risk exacerbates idiosyncratic fluctuations (Gaspar & Massa, 2006; Irvine & Pontiff, 2009), raises financing costs (Valta, 2012), and discourages investment and productivity growth due to the uncertain returns on new investments (Aghion et al., 2005; Akdoğu & MacKay, 2012).

These two opposing effects of competition—the disciplining and risk-increasing mechanisms—have significant implications for managerial decision-making. On one hand, the disciplining effect of competition can reduce managerial inefficiency, prompting managers to cut costs, improve product quality (Baggs & de Bettignies, 2007), engage in corporate social responsibility and avoid unethical practices (Flammer, 2015; Arouri et al., 2021), and encourages innovation (Le et al., 2021) or the acceleration of investments to stay ahead of competitors in high-growth environments (Jiang et al., 2015). On the other hand, the risk-increasing mechanism suggests that managers may adopt precautionary measures to safeguard against heightened uncertainty and risk. These measures can include building up cash reserves (Chi & Su, 2016), adopting conservative payout policies (Hoberg et al., 2014), making cautious investment decisions (Frésard & Valta, 2016; Boubaker et al., 2022), and withholding competition-sensitive information to prevent rivals from using it strategically (e.g., Huang et al., 2017a; Li & Zhan, 2019; Ryou et al., 2022). More pertinent to our study are the capital structure decisions that competition influences, which we explore in the following section.

2.2. PMC and Capital Structure

Product and financial markets are closely intertwined. First, a firm’s financial structure can influence its behavior in product markets. Second, firms anticipate the impact of financial decisions on product market outcomes, meaning that market conditions also affect financial decisions.

Brander and Lewis (1986) show that a firm’s financial structure affects product market strategies. Increasing debt encourages firms to adopt aggressive product strategies, boosting returns in favorable conditions while lowering them in adverse situations. Since shareholders can ignore losses in bankruptcy (where bondholders bear the risk), firms may use debt strategically to force rivals into reducing their output, particularly in uncertain markets.

Building on this foundation, existing research has explored how competition, through its two mechanisms—those disciplinary and risk-increasing—shapes firms’ capital structure decisions, particularly their use of debt. This body of work can be divided into two key strands. The first strand suggests that, under the disciplinary effect of competition, firms are less likely to avoid external monitoring mechanisms such as debt. This view aligns with Flannery’s (1986) signaling theory, which posits that firms willingly subject themselves to strict financing mechanisms, like debt, as a signal to outsiders of their commitment to reducing agency costs, particularly in environments with high information asymmetry. The second strand of research argues that, in response to the risk-increasing effect of competition, firms may reduce their exposure to debt to avoid exacerbating bankruptcy risk. Xu (2012) finds that competition leads firms to lower their leverage ratios and issue more equity instead. Chu and Pham (2021) similarly demonstrate that PMC encourages firms to adopt debt-free capital structures, as competition increases debt costs (Valta, 2012; Platt, 2020) while reducing equity costs (Sassi et al., 2019). This makes equity financing more attractive in competitive markets. Moreover, Cumming et al. (2022) show that heightened competition increases business uncertainty, leading banks to become reluctant to lend, pushing entrepreneurial firms toward venture capital financing.

Building on these studies, with further details provided in

Table A1. Literature Review Matrix.

Study	Year	Theory/Concept	Methodology	Key Findings	Relevance
Bright side of product market competition
Hart	1983	Agency theory, competition and efficiency	Theoretical analysis	Product market competition enhances efficiency by disciplining managers, reducing agency conflicts.	Supports the “bright side” of competition, showing how competition can lead to managerial efficiency.
Schmidt	1997	Agency theory	Theoretical analysis	Managers in competitive markets face stronger incentives to perform efficiently due to the threat of losing market share.	Reinforces the disciplining effect of competition on managerial behavior.
Holmstrom	1982	Relative performance evaluation (RPE)	Theoretical analysis	RPE allows owners to compare managerial performance with peers, helping distinguish firm-specific issues from market-wide factors.	Highlights how competition aids in more accurate performance evaluation of managers.
Nalebuff & Stiglitz	1983	Relative performance evaluation	Theoretical analysis	Similar findings to Holmstrom (1982) emphasizing the role of competition in enabling better performance evaluation by comparing firms in the same industry.	Supports the use of competition as a tool for more precise managerial assessments.
Dasgupta et al.	2018	Disciplining mechanism of competition	Empirical study	Competition acts as an external governance mechanism, disciplining managerial behavior.	Provides empirical evidence on the disciplining role of competition, relevant to examining the impact of PMC on managerial decisions.
Sassi et al.	2019	Financing costs and competition	Empirical study	Product market competition reduces financing costs, indicating competition’s ability to substitute other governance mechanisms.	Relevant for understanding the relationship between competition and financial decisions, particularly debt policies.
Dark side of product market competition
Tirole	2006	Predation theory, “dark side” of competition	Theoretical analysis	Competition can lower profit margins and increase default risk, particularly through predatory behavior by stronger rivals.	Provides the counter-argument to the “bright side” of competition, relevant for discussing the risk-increasing effects of competition.
Bolton & Scharfstein	1990	Predation theory	Theoretical analysis	Stronger rivals in competitive markets can force weaker firms out through aggressive strategies, exacerbating risks.	Demonstrates the risk-increasing effect of competition, highlighting predatory risks from rivals.
Valta	2012	Financing costs and competition	Empirical study	Competition increases debt costs.	Links competition to firms’ financing decisions, suggesting how firms may tend to avoid costly debt in competitive markets.
Hoberg et al.	2014	Precautionary policies and competition	Empirical study	Firms in competitive markets tend to adopt conservative payout and investment policies due to heightened risk and uncertainty.	Relevant for examining how PMC leads firms to adopt conservative financial policies, such as maintaining zero-leverage.
Platt	2020	Competition’s impact on financing decisions	Empirical study	Competition increases the cost of public debt	Links competition to firms’ financing decisions, suggesting how firms may tend to avoid costly debt in competitive markets.
Product market competition and capital structure
Flannery	1986	Signaling theory	Theoretical analysis	Firms use debt as a signal of commitment to reducing agency costs, particularly in competitive environments with high information asymmetry.	Shows how competition can lead to higher leverage as a signal to outsiders, important for discussing PMC’s influence on capital structure.
Brander and Lewis	1986	Interaction between financial structure and product market strategy	Theoretical analysis	Firms use debt to adopt aggressive product strategies	Firms increase debt to force rivals into reducing output, especially in uncertain markets
Xu	2012	Impact of competition on capital structure decisions	Empirical study	Firms reduce leverage and issue more equity in response to competition	PMC leads to lower leverage ratios due to increased risk of bankruptcy
Cumming et al.	2022	PMC and entrepreneurial financing decisions	Empirical study	Banks reduce lending in competitive markets, encouraging VC financing	PMC increases uncertainty, limiting access to debt financing and pushing firms to VC
Zero-debt phenomenon
Strebulaev & Yang	2013	Zero-leverage puzzle	Empirical study	A significant proportion of large U.S. firms adopt zero-leverage policies, often driven by profitability, liquidity, and low distress risk.	Provides foundational insights into the ZL phenomenon, relevant to understanding how firm characteristics drive conservative capital structures.
Dang	2013	Zero-leverage and macroeconomic conditions	Empirical study	Firms avoid debt during unfavorable macroeconomic conditions, especially when GDP growth is low or interest rates are high.	Connects broader macroeconomic factors to zero-leverage behavior, providing context for the role of external factors on capital structure decisions.
Bessler et al.	2013	Zero-leverage and legal factors	Empirical study	ZL behavior is more common in countries with strong creditor rights, where firms face harsher bankruptcy consequences.	Demonstrates how legal and institutional environments influence firms’ debt policies, complementing the analysis of PMC’s effect.
El Ghoul et al.	2018	Cultural factors and zero-leverage	Empirical study	Cultural factors, such as conservatism and trust levels, shape the prevalence of ZL policies across countries.	Introduces the role of cultural factors in capital structure decisions, relevant for understanding broader external influences beyond competition.

in Appendix A, our research focuses on ZL policies, a crucial yet underexplored aspect of firms’ capital structures in the context of competitive pressure.

2.3. Zero-Debt Policy

Graham (2000) underscores a paradox in corporate finance, where large, liquid, and highly profitable firms—those with minimal distress risk—often choose to maintain conservative debt levels, contrary to what traditional capital structure theories would predict (i.e., trade-off theory, pecking-order theory). This occurs despite the well-established advantages of debt financing, such as lower cost of asymmetric information (Myers & Majluf, 1984), lower equity agency costs (Jensen, 1986), and the tax benefits of debt (Modigliani & Miller, 1963; Miller, 1977). Moreover, numerous studies have documented a global trend toward debt-free capital structures, contributing to what is now referred to as the “zero-leverage puzzle.” For instance, Strebulaev and Yang (2013) show that between 1962 and 2009, an average of 10.2% of large, publicly traded U.S. firms carried zero debt, and nearly 22% had leverage ratios below 5%. Persistence in ZL behavior is evident, with 30% refraining from debt even after five years. Similarly, Dang (2013) finds that 12.18% of UK non-financial firms maintained zero debt between 1980 and 2007. These findings are echoed in other contexts, with studies by Z. Huang et al. (2017b) in China, Bessler et al. (2013), and El Ghoul et al. (2018) analyzing broader international samples. This widespread ZL phenomenon has spurred significant academic interest in understanding why firms favor such conservative debt policies.2

Strebulaev and Yang (2013) highlight the role of firm-level characteristics in driving the adoption of ZL policies. Specifically, the authors document that firms that consistently maintain zero debt tend to have higher profitability, larger cash reserves, and greater tax payments compared to their peers, aligning with Graham’s (2000) observation that profitable firms often avoid debt. CEO characteristics also play a role—firms with higher CEO ownership and longer tenure are more likely to adopt ZL policies, as CEOs may avoid debt to protect personal wealth and maintain control over strategic decisions. Additionally, family-owned firms, which prioritize long-term survival, often opt for ZL to minimize financial risk. The study further highlights that weaker corporate governance, such as smaller or less independent boards, encourages conservative debt usage, as managers may avoid debt to reduce external scrutiny from debtholders.

Beyond firm-level characteristics, macroeconomic conditions also influence ZL policies. Dang (2013) finds that firms are more likely to avoid debt during unfavorable macroeconomic periods, such as periods of low or negative GDP growth or when there is a wider term structure of interest rates. Interestingly, the impact of these macro-level variables on ZL policies, as well as on firms’ decisions to issue debt, is more significant for less financially constrained firms, such as dividend payers, suggesting that firm-specific financial characteristics interact with broader economic conditions to shape capital structure decisions.

The prevalence of ZL policies can also be influenced by country-level factors, including both legal and extra-legal determinants. Bessler et al. (2013) show that ZL behavior is more common in countries with stronger creditor rights, where creditors can quickly liquidate distressed firms without pursuing reorganization. Similarly, in common law countries, enhanced investor protection improves access to external equity, making ZL policies more feasible. El Ghoul et al. (2018) extend this analysis by highlighting the impact of cultural factors, demonstrating that firms in countries with higher Conservatism or Mastery scores, or in those characterized by high levels of trust, are more inclined to adopt ZL strategies. This underscores the role of both legal systems and prevailing cultural norms in shaping firms’ financial policies.

To build on these existing insights, our paper shifts the focus to a less explored but equally important determinant of capital structure decisions: product market competition (PMC). Unlike firm-level characteristics such as profitability or macroeconomic and country-level factors like creditor rights and legal systems, PMC operates at the industry level, exerting competitive pressures that influence a firm’s strategic decisions, including its leverage policy. In this paper, we investigate how the intensity of competition within industries affects the adoption of ZL policies. Our study contributes to the broader understanding of how external pressures, beyond firm-specific or country-level variables, shape conservative capital structures.

2.4. Hypothesis Development

In this section, we aim to articulate how PMC influences the likelihood of firms adopting ZL policies, an extreme form of conservative capital structure. This forms the basis of our first hypothesis. Furthermore, we introduce earnings volatility as an interaction term in our second hypothesis to investigate how it moderates the relationship between competition and ZL behavior.

The relationship between competition and ZL policies is theoretically ambiguous. On the one hand, competition can act as a disciplinary force on firms, while on the other, it can increase the risk of financial distress, leading firms to adopt more conservative financial strategies.

The first perspective, which we label the “disciplinary force” argument, posits that competition serves as an external mechanism that pressures managers to perform more effectively. As competition intensifies, managers—particularly those who are entrenched or have the ability to avoid accountability—face heightened pressure to work diligently and make more efficient decisions to avoid the risk of bankruptcy. This perspective is grounded in agency theory, which suggests that competition reduces the scope for managerial slack by intensifying performance expectations. This increased discipline encourages managers to avoid behaviors that could raise concerns about their competence. In this context, debt plays a critical role in disciplining managers, as it subjects them to periodic scrutiny through its monitoring mechanisms (Flannery, 1986). By committing to debt financing, managers signal their willingness to reduce agency costs and improve governance. As a result, we would expect firms operating in highly competitive markets to be less inclined to adopt ZL policies, which would allow them to escape the disciplining effects of debt. In support of this view, Strebulaev and Yang (2013) argue that ZL policies are often associated with weak corporate governance, where managers avoid debt to escape external scrutiny and monitoring. Hence, in more competitive environments, firms may be more likely to use debt to signal their commitment to governance and performance, reducing the likelihood of ZL policies.

Conversely, the second perspective, which we label the “risk-increasing” argument, suggests that heightened competition increases the risk faced by firms. This perspective is rooted in predation theory, which suggests that financially stronger rivals may engage in aggressive pricing and production strategies to weaken their competitors, exacerbating the risk of failure for financially vulnerable firms (Bolton & Scharfstein, 1990). This particular situation amplifies default risk by eroding profit margins and increasing earnings volatility (Tirole, 2006), this leading to more conservative financial policies, such as avoiding debt altogether. In this context, managers may opt for ZL strategies as a precautionary measure to shield the firm from the risk of financial distress, particularly when competition is intense. Research by Devos et al. (2012) corroborates this argument, showing that firms with zero-leverage are often financially constrained, suggesting that in competitive environments, firms may strategically avoid debt to mitigate the risks associated with increased competition. Additionally, Valta (2012) and Platt (2020) find that competition raises the cost of both bank and public debt, making equity financing a more attractive alternative (Sassi et al., 2019). As a result, firms may adopt ZL policies as a strategic response to manage risk and preserve financial flexibility in the face of competition.

In light of these two opposing mechanisms, we formulate our first hypothesis as follows:

H1. 

The likelihood of zero-leverage behavior increases/decreases with an escalation of competition in the product market.

In the following, we elaborate on the role of earnings volatility in shaping the relationship between PMC and the likelihood of firms adopting ZL strategies. Earnings volatility, which measures the extent to which a company’s profits fluctuate over time, is a critical concern for businesses as it impacts their ability to plan for the future, make strategic decisions, and manage financial risks. From an operational perspective, firms with highly volatile earnings face challenges in securing external financing, as lenders and investors perceive these fluctuations as a significant financial risk (Subramanyam, 1996). Firms with unstable earnings may also attract increased scrutiny from regulatory bodies and market participants, potentially triggering interventions or negative reactions that could further exacerbate financial difficulties. From a regulatory and market perspective, managing earnings volatility becomes vital for maintaining a stable financial appearance. Companies with fluctuating earnings may engage in earnings management to smooth out these variations, ensuring that their financials appear more predictable and attractive to investors (Lai, 2011). This strategy is especially pertinent in competitive industries, where firms aim to maintain an edge by projecting stability (e.g., Datta et al., 2013; Harris, 2024). Firms with volatile earnings may opt for ZL policies to enhance creditworthiness and appeal to investors who favor stable returns (Deesomsak et al., 2004).

Given this context, examining how earnings volatility interacts with PMC can help disentangle two competing views on the PMC- ZL relationship. According to one perspective, competition acts as an external disciplinary mechanism, reducing entrenched managers’ incentives to avoid debt. If this view holds, we would expect less reliance on ZL strategies as competition intensifies, especially for firms with lower earnings volatility, suggesting that competition curbs managerial tendencies to avoid debt, even in volatile environments. The second view posits that heightened competitive pressure increases risk, prompting managers to avoid debt obligations that could further depress earnings in volatile environments. If this perspective is correct, we would observe more ZL policies in response to increased competition, particularly among firms with elevated earnings volatility, indicating that competition intensifies the perceived risks associated with volatile earnings, leading firms to avoid debt. Based on this reasoning, we formulate our second hypothesis as follows:

H2. 

Firms with higher levels of earnings volatility will exhibit a stronger/weaker relationship between PMC and the likelihood of adopting a zero-leverage approach.

3. Sample and Methodology

In this section, we describe the methodology used to empirically test hypotheses regarding the relationship between PMC, ZL policies, and earnings volatility.

We use FLUIDITY to measure PMC, a forward-looking, firm-level indicator of competitive pressure. Traditional measures of competition often rely on industry concentration proxies, such as the Herfindahl–Hirschman Index (HHI) or concentration ratios (e.g., the share of sales held by the largest firms in an industry), which have been widely used in studies linking competition to financial outcomes (e.g., DeFond & Park, 1999; Gaspar & Massa, 2006; Hou & Robinson, 2006; Giroud & Mueller, 2010, 2011). However, these measures have their several well-known limitations, as they are generally calculated based on SIC or NAICS classifications, which group firms by the relatedness of their production processes rather than by product similarity. This can result in firms producing distinct products being grouped together, creating a classification bias (Hoberg & Phillips, 2010). Also, traditional industry concentration measures, such as those derived from Census data, are often updated infrequently—typically every five years—limiting their ability to capture real-time changes in market dynamics (Akdoğu & Mackay, 2012). To overcome these issues, we employ FLUIDITY, a text-based measure developed by Hoberg et al. (2014), which analyzes product descriptions in firms’ 10-K filings to assess the degree of competitive threat. FLUIDITY captures the extent to which a firm’s product space evolves relative to rivals, offering a dynamic and ex ante perspective on competition. Unlike traditional measures, FLUIDITY is updated annually and reflects real-time product market shifts, providing a more accurate representation of the competitive pressures firms face. Moreover, it mitigates endogeneity concerns by focusing on external competitive threats rather than internal firm strategies, making it a valuable exogenous tool for examining the impact of competition on financial decisions such as capital structure.

To capture firms that adopt a zero-debt policy, we employ two classification variables: ZL and AZL. A firm is classified as zero-leveraged (ZL) if it has no long-term debt and no current liabilities at the fiscal year-end, consistent with the approach used by Strebulaev and Yang (2013). This strict criterion ensures that the firm operates entirely without debt. Additionally, we define almost-zero-leveraged (AZL) firms as those with a book–leverage ratio not exceeding 5% in any given year. This allows us to account for firms that maintain minimal leverage, providing a more nuanced view of firms that operate with very low debt levels. These classifications will enable us to examine the use of ZL policies in response to PMC and other influencing factors.

We also include earnings volatility as a key variable in our analysis. This measure is calculated as the standard deviation of a firm’s profitability over the past 10 years, with a minimum requirement of three years of available data. By incorporating earnings volatility, we capture the historical fluctuations in firm performance, which may influence how firms adjust their capital structure decisions in response to competitive pressures.

We further include a set of control variables following Strebulaev and Yang (2013) to account for factors commonly examined in zero-leverage studies. These variables are: dividend, firm size, market-to-book asset ratio, profitability, tangibility, firm age, R&D expenditures, capital expenditures, asset sales, and tax rates.

Table 1. Variables’ Definitions. The table below outlines the formula for each variable, utilizing the abbreviated names of Compustat primitive variables.

Variable	Definition
Dividend	The ratio of common dividends to book assets (dvc/at)
Size	Natural logarithm of book assets adjusted to 2000 dollars (log(ATt × CPI2000 × CPIt))
Market to Book Ratio	The ratio of market assets to book assets (Tobin’s q) ((lt + pstkl-txditc + csho*prcc_f)/at)
Profitability	The ratio of earnings before interests, taxes, and depreciation to book assets (oibdp/at)
Tangibility	The ratio of fixed assets to book assets (ppent/at)
Age	Natural logarithm of the number of years since the firm’s record first appears in Compustat (log(fyear–IPOyear))
R&D	The ratio of research and development expenses to sales (xrd/sale)
Capital Expenditure	The ratio of capital expenditure to book assets (capex/at)
Asset Sale	The ratio of asset sales to book assets ((sppe (Sale of Property, Plant, and Equipment) + siv (Investments − Decrease))/at)
Tax	The ratio of taxes paid to book assets (txt/at)
Earnings Volatility	The volatility of profitability calculated for the past 10 years (minimum of three years of data required)

provides a detailed definition of these variables.

To construct our sample, we merged PMC data from the Hoberg and Phillips Data Library with financial data from Compustat. The Compustat data are used to calculate our key variables, including ZL, AZL, earnings volatility, and a set of control variables necessary for our analysis. Due to the non-availability of certain financial information for some firms, the final sample consists of an unbalanced panel of 162,904 firm-year observations over the 1989–2019 period.

To test our hypotheses, we employ logistic regression models. Specifically, we use a multivariate logistic regression to examine the first hypothesis, capturing the relationship between PMC and the likelihood of firms adopting a ZL policy. We use one-year lagged independent and control variables in all models. This approach helps mitigate endogeneity concerns and avoids simultaneity bias, ensuring that changes in independent variables precede changes in the dependent variable. The model specification is as follows:

$$\mathrm{H}1(\mathrm{a}):ZL={\beta }_0+{\beta }_1\times Fluidity+\sum {\beta }_n\times Controls+IndustryFE+YearFE+\epsilon .$$

(1)

$$\mathrm{H}1(\mathrm{b}):AZL={\beta }_0+{\beta }_1\times Fluidity+\sum {\beta }_n\times Controls+IndustryFE+YearFE+\epsilon$$

(2)

For the second hypothesis, we extend the multivariate logistic regression by introducing an interaction term between PMC and earnings volatility, allowing us to explore how earnings volatility moderates the impact of competition on ZL policies. The model for the second hypothesis is as follows:

$$\mathrm{H}2(\mathrm{a}):ZL={\beta }_0+{\beta }_{1}Fluidity+{\beta }_{2}EV+{\beta }_{3}EV\times Fluidity+\sum {\beta }_n\times Controls+IndustryFE+YearFE+\epsilon .$$

(3)

$$\mathrm{H}2(\mathrm{b}):AZL={\beta }_0+{\beta }_{1}Fluidity+{\beta }_{2}EV+{\beta }_{3}EV\times Fluidity+\sum {\beta }_n\times Controls+IndustryFE+YearFE+\epsilon .$$

(4)

4. Empirical Analysis

4.1. Descriptive Statistics and Univariate Analysis

We begin with preliminary analyses by summarizing key descriptive statistics in

Table 2. Summary Statistics. The table presents an overview of summary statistics of our main variables, i.e., product market competition (Fluidity) from the Hoberg and Phillips Data Library, and other financial variables sourced from the Compustat (North America) database, covering the period from 1989 to 2019. For a comprehensive understanding of each variable, detailed descriptions are available in Table 1.

Variable	Obs	Mean	Std.Dev.	Min	Max
FLUIDITY	162,765	7.59	4.08	0.06	37.23
Earnings Volatility	103,492	0.08	0.1	0	0.83
Dividend	158,630	0.01	0.02	0	0.16
Size	157,730	5.6	2.13	0.91	11.01
Market to Book Ratio	132,026	2.01	1.65	0.53	13.09
Profitability	141,556	0.03	0.23	−1.45	0.42
Tangibility	154,913	0.23	0.24	0	0.9
Age	153,026	2.2	1.05	0	4.29
R&D	143,634	0.17	0.84	0	13.05
Capital Expenditure	140,924	0.05	0.06	0	0.37
Asset Sale	110,501	0.04	0.11	0	0.91
Tax	157,196	0.01	0.03	−0.08	0.13

, and then highlighting correlations among the ZL dummy variable, earnings volatility, and fluidity. The correlation coefficients presented in

Table 3. Pairwise Correlation Matrix. The table below displays the pairwise correlations among the dependent, independent, and control variables. This dataset comprises information on product market competition (Fluidity) from the Hoberg and Phillips Data Library and other financial variables sourced from the Compustat (North America) database, covering the period from 1989 to 2019. For a comprehensive understanding of each variable, detailed descriptions are available in Table 1. “***” and “**” denote statistical significance levels at 1% and 5%, respectively.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)
(1) ZL	1.000
(2) FLUIDITY	0.006 **	1.000
(3) Earnings Volatility	0.204 ***	0.137 ***	1.000
(4) Dividend	−0.002	−0.094 ***	−0.158 ***	1.000
(5) Size	−0.248 ***	0.073 ***	−0.451 ***	0.217 ***	1.000
(6) Market to Book Ratio	0.196 ***	0.136 ***	0.335 ***	−0.012 ***	−0.203 ***	1.000
(7) Profitability	−0.124 ***	−0.241 ***	−0.486 ***	0.218 ***	0.384 ***	−0.231 ***	1.000
(8) Tangibility	−0.180 ***	−0.091 ***	−0.137 ***	0.113 ***	0.090 ***	−0.176 ***	0.194 ***	1.000
(9) Age	−0.071 ***	−0.230 ***	−0.209 ***	0.191 ***	0.326 ***	−0.143 ***	0.220 ***	0.120 ***	1.000
(10) R&D	0.097 ***	0.186 ***	0.279 ***	−0.079 ***	−0.148 ***	0.210 ***	−0.414 ***	−0.106 ***	−0.109 ***	1.000
(11) Capital Expenditure	−0.084 ***	0.007 ***	−0.051 ***	−0.010 ***	0.011 ***	0.028 ***	0.131 ***	0.590 ***	−0.056 ***	−0.049 ***	1.000
(12) Asset Sale	0.066 ***	0.127 ***	−0.009 **	0.001	0.110 ***	−0.004	−0.030 ***	−0.116 ***	0.005	0.041 ***	−0.080 ***	1.000
(13) Tax	0.067 ***	−0.171 ***	−0.136 ***	0.184 ***	0.091 ***	0.132 ***	0.420 ***	0.054 ***	0.089 ***	−0.098 ***	0.117 ***	−0.036 ***	1.000

*** p < 0.01, ** p < 0.05.

do not indicate any multicollinearity issues among the ZL dummy variable, earnings volatility, and fluidity. To further verify this, we complemented the correlation matrix with Variance Inflation Factor (VIF) tests in

Table 4. Variance inflation factors. This table shows the variance inflation factors (VIF) for the regression models used to test Hypothesis 1 (models 1 and 2) and Hypothesis 2 (models 3 and 4).

For Hypothesis 1			For Hypothesis 2
Variable	VIF	1/VIF	Variable	VIF	1/VIF
Profitability	1.9	0.525888	FLUIDITY	1.89	0.529141
Tangibility	1.81	0.553873	Earnings Volatility (EV)	6.7	0.149356
Capex	1.69	0.592348	FLUIDITY × EV	7.85	0.127335
Tax	1.45	0.687737	Dividend	1.17	0.856487
Size	1.34	0.748798	Size	1.5	0.66531
R&D	1.3	0.767276	Market/Book Ratio	1.3	0.771331
Market/Book Ratio	1.22	0.821402	Profitability	2.15	0.464258
FLUIDITY	1.21	0.826446	Tangibility	1.84	0.544868
Age	1.2	0.83283	Age	1.26	0.791402
Dividend	1.15	0.867708	R&D	1.38	0.722641
Asset Sale	1.03	0.967202	Capex	1.7	0.587114
Mean VIF	1.41		Asset Sale	1.05	0.955681
			Tax	1.47	0.679172
			Mean VIF	2.4

. For the models testing Hypothesis 1, the average VIF is 1.41, with all individual VIFs below 2, indicating very low multicollinearity. For Hypothesis 2, while the interaction term FLUIDITY × EV has a VIF of 7.85 and Earnings volatility (EV) has a VIF of 6.70, these values remain below the commonly accepted threshold of 10. Overall, the VIF results confirm that multicollinearity is not a concern.

We then perform a univariate analysis to examine the differences in characteristics between firms adopting ZL policies and those that do not. The results are presented in

Table 5. Univariate analysis of ZL Policy. This table displays the results of the univariate analysis, employing t-tests, conducted on two distinct sets of companies: those classified as ZL firms and those categorized as Non-ZL firms. This dataset comprises information on product market competition (Fluidity) from the Hoberg and Phillips Data Library and other financial variables sourced from the Compustat (North America) database, covering the period from 1989 to 2019. For a thorough comprehension of each variable, comprehensive descriptions can be found in Table 1. Additionally, the table provides information on p-values and standard error values.

	Non-ZL Number of Firms	ZL Number of Firms	Non-ZL Mean	ZL Mean	Difference	St_Err	p_Value
Dividend	136,673	21,957	0.008	0.0075	0	0	0.4385
Size	135,321	22,409	5.8165	4.3015	1.515	0.015	<0.0004
Market to Book Ratio	111,566	20,460	1.875	2.766	−0.891	0.0125	<0.0004
Profitability	120,629	20,927	0.044	0.1245	−0.0805	0.0015	<0.0004
Tangibility	132,231	22,682	0.243	0.1235	0.1195	0.0015	<0.0004
Age	132,058	20,968	2.2275	2.01	0.2175	0.008	<0.0004
R&D	123,476	20,158	0.1325	0.3675	−0.235	0.0065	<0.0004
Capital Expenditure	119,460	21,464	0.0515	0.038	0.013	0.0005	<0.0004
Asset Sale	92,000	18,501	0.036	0.056	−0.02	0.001	<0.0004
Tax	135,372	21,824	0.0135	0.019	−0.005	0	<0.0004
Earnings Volatility	88,581	14,911	0.069	0.128	−0.059	0.001	<0.0004

and

Table 6. Univariate analysis of AZL Policy. This table displays the results of the univariate analysis, employing t-tests, conducted on two distinct sets of companies: those classified as AZL firms and those categorized as non-AZL firms. This dataset comprises information on product market competition (Fluidity) from the Hoberg and Phillips Data Library, and other financial variables sourced from the Compustat (North America) database, covering the period from 1989 to 2019. For a thorough comprehension of each variable, comprehensive descriptions can be found in Table 1. Additionally, the table provides information on p-values and standard error values.

	Non-AZL Number of Firms	AZL Number of Firms	Non-AZL Mean	AZL Mean	Difference	St_Err	p_Value
Dividend	114,068	44,562	0.0085	0.0065	0.002	0	<0.0004
Size	112,961	44,769	5.9635	4.6875	1.2765	0.0115	<0.0004
Market to Book Ratio	93,667	38,359	1.753	2.6475	−0.8945	0.0095	<0.0004
Profitability	100,657	40,899	0.0535	0.126	−0.0725	0.0015	<0.0004
Tangibility	109,914	44,999	0.2645	0.1305	0.134	0.0015	<0.0004
Age	111,526	41,500	2.273	1.9955	0.2775	0.006	<0.0004
R&D	103,761	39,873	0.1025	0.3295	−0.227	0.005	<0.0004
Capital Expenditure	99,966	40,958	0.053	0.041	0.0115	0.0005	<0.0004
Asset Sale	75,730	34,771	0.0345	0.0505	−0.0165	0.0005	<0.0004
Tax	113,170	44,026	0.013	0.018	−0.005	0	<0.0004
Earnings Volatility	75,609	27,883	0.065	0.112	−0.0475	0.0005	<0.0004

. Specifically, non-ZL and non-AZL firms tend to have higher average values for variables such as size, age, tangibility, and capital expenditure. In contrast, ZL and AZL firms exhibit higher averages for variables like market-to-book ratio, profitability, R&D expenditures, asset sales, tax, and earnings volatility, compared to their counterparts.

4.2. Multivariate Analysis

In this section, we perform multivariate logistic regression analysis by estimating regression models (1) and (2) using four specifications. The first two specifications (

Table 7. Logistic regression models for ZL Policies. This table reports results from logistic regression models, with the dependent variable being a dummy variable signifying the presence of a ZL policy. The primary independent variable in these models is “Fluidity”, which measures PMC. Model 1 includes control variables such as dividend, size, market to book ratio, profitability, and tangibility. Model 2 extends the control variables to include age, R&D, capex (Capital Expenditure), asset sale, and tax. This dataset comprises information on product market competition (Fluidity) from the Hoberg and Phillips Data Library and other financial variables sourced from the Compustat (North America) database, covering the period from 1989 to 2019. Table 1 provides descriptions of all variables in detail. The robust standard error adjusted for heteroskedasticity is used for all regressions and is reported in parentheses. “***” and “**” denote statistical significance levels at 1% and 5%, respectively.

	(1)	(2)
	ZL	ZL
FLUIDITY	0.0250 ***	0.0215 ***
	(0.0057)	(0.0069)
Dividend	11.7891 ***	12.2598 ***
	(1.0435)	(1.3413)
Size	−0.3792 ***	−0.3933 ***
	(0.0133)	(0.0154)
Market/Book Ratio	0.1325 ***	0.0866 ***
	(0.0095)	(0.0115)
Profitability	0.8102 ***	0.5697 ***
	(0.0769)	(0.0953)
Tangibility	−3.5276 ***	−4.0244 ***
	(0.1666)	(0.2288)
Age		−0.0175
		(0.0225)
R&D		0.0879 ***
		(0.0164)
Capex		3.3309 ***
		(0.4709)
Asset Sale		1.7270 ***
		(0.1480)
Tax		9.3141 ***
		(0.6736)
Constant	−1.0666 **	−1.0169 **
	(0.4230)	(0.5028)
Obs.	107432	71918
Pseudo R2	0.1812	0.1844
Industry Dummy	Yes	Yes
Year Dummy	Yes	Yes

Standard errors are in parentheses—*** p < 0.01, ** p < 0.05.

) focus on the ZL policy: one includes the main control variables, while the second incorporates additional control variables. The next two specifications (

Table 8. Logistic regression models for AZL Policies. This table reports results from logistic regression models, with the dependent variable being a dummy variable signifying the presence of the AZL (Almost Zero-leverage) policy. The primary independent variable in these models is “Fluidity”, which measures product market competition. Model 3 includes control variables such as dividend, size, market to book ratio, profitability, and tangibility. Model 4 extends the control variables to include age, R&D, capex (Capital Expenditure), asset sale, and tax. This dataset comprises information on product market competition (Fluidity) from the Hoberg and Phillips Data Library and other financial variables sourced from the Compustat (North America) database, covering the period from 1989 to 2019. Table 1 provides descriptions of all variables in detail. The robust standard error adjusted for heteroskedasticity is used for all regressions and is reported in parentheses. “***” denote statistical significance levels at 1%.

	(3)	(4)
	AZL	AZL
FLUIDITY	0.0276 ***	0.0265 ***
	(0.0047)	(0.0057)
Dividend	9.2264 ***	10.1871 ***
	(0.9548)	(1.1993)
Size	−0.3417 ***	−0.3481 ***
	(0.0112)	(0.0130)
Market/Book Ratio	0.1942 ***	0.1315 ***
	(0.0091)	(0.0103)
Profitability	0.8434 ***	0.5029 ***
	(0.0674)	(0.0850)
Tangibility	−3.1766 ***	−3.6577 ***
	(0.1229)	(0.1696)
Age		−0.0215 ***
		(0.0043)
R&D		0.0966 ***
		(0.0174)
Capex		3.5923 ***
		(0.3620)
Asset Sale		1.4405 ***
		(0.1362)
Tax		10.2051 ***
		(0.5616)
Constant	−0.1182	−0.0997
	(0.3143)	(0.3901)
Obs.	107484	72251
Pseudo R2	0.1953	0.1952
Industry Dummy	Yes	Yes
Year Dummy	Yes	Yes

Standard errors are in parentheses—*** p < 0.01.

) focus on the AZL policy.

Throughout all four specifications’ results, reported in Table 7 and Table 8, we observe a positive and significant coefficient on FLUIDITY, suggesting that the likelihood of the ZL policy increases with heightened PMC, aligning with the risk-increasing mechanism of PMC. Indeed, as competition increases, firms face greater risk and uncertainty in the marketplace, prompting them to avoid the additional financial risk associated with debt financing. By avoiding debt, firms can reduce their exposure to bankruptcy risk, thus maintaining greater financial flexibility to navigate volatile market conditions. This strategic response to competitive pressures highlights the importance of considering market dynamics when evaluating capital structure decisions.

The coefficients on control variables are consistent with Strebulaev & Yang (2013). Specifically, ZL firms are smaller, possess a higher market-to-book ratio, exhibit greater profitability, own fewer tangible assets, and pay higher dividends (Table 7). Smaller firms, contending with larger rivals, may strategically choose a ZL policy to maintain flexibility and mitigate debt-related financial risks. ZL policy-adopters often boast a higher market-to-book ratio, indicating that the market values their assets more than their book values, potentially anticipating future growth and profitability. ZL firms are more likely to be profitable, as avoiding debt enables more efficient resource allocation. These firms, focusing on intangible assets like intellectual property and technology, often thrive in competitive, rapidly evolving industries. ZL policy adopters pay higher dividends, benefiting from lower financial obligations. By allocating resources to R&D, capital expenditures, and asset sales, they invest in innovation and optimize their asset portfolio, potentially incurring higher tax liabilities due to improved profitability. Similar patterns emerge for AZL firms (Table 8).3

Finally, we estimate regression models (3) and (4) to assess the interaction between earnings volatility and PMC for the ZL policy (Column (5) of

Table 9. Logistic regression models for earnings volatility in Interaction Term. This table reports results from logistic regression models, with the dependent variable being a dummy variable signifying the presence of the AZL (Almost Zero-leverage) policy. The primary independent variables in these models are “Fluidity”, which measures product market competition, earnings volatility (EV) and the interaction term of Fluidity and EV. This dataset comprises information on product market competition (Fluidity) from the Hoberg and Phillips Data Library and other financial variables sourced from the Compustat (North America) database, covering the period from 1989 to 2019. Table 1 provides descriptions of all variables in detail. The robust standard error adjusted for heteroskedasticity is used for all regressions and is reported in parentheses. “***”, “**”, and “*”denote statistical significance levels at 1%, 5% and 10%, respectively.

	(5)	(6)
	ZL	AZL
FLUIDITY	0.0242 **	0.0291 ***
	(0.0098)	(0.0080)
Earnings Volatility (EV)	0.0510 **	0.2105 *
	(0.0210)	(0.1169)
FLUIDITY × EV	0.0160 **	0.0109 **
	(0.0064)	(0.0045)
Dividend	12.1150 ***	10.0194 ***
	(1.4232)	(1.2811)
Size	−0.4025 ***	−0.3694 ***
	(0.0174)	(0.0148)
Market/Book Ratio	0.0859 ***	0.1356 ***
	(0.0134)	(0.0120)
Profitability	0.5098 ***	0.4845 ***
	(0.1124)	(0.1017)
Tangibility	−3.9266 ***	−3.5977 ***
	(0.2411)	(0.1829)
Age	−0.0635 **	−0.0408
	(0.0320)	(0.0272)
R&D	0.1032 ***	0.1014 ***
	(0.0187)	(0.0203)
Capex	3.4054 ***	3.7155 ***
	(0.5299)	(0.4160)
Asset Sale	1.8430 ***	1.5107 ***
	(0.1644)	(0.1506)
Tax	9.3940 ***	9.9727 ***
	(0.7349)	(0.6174)
Constant	−1.1635 *	0.1787
	(0.6334)	(0.4117)
Obs.	59722	59966
R2	0.1887	0.1969
Industry Dummy	Yes	Yes
Year Dummy	Yes	Yes
Wald chi2 statistic	5.11	5.66
Prob > chi2	0.017	0.012

Standard errors are in parentheses—*** p < 0.01, ** p < 0.05, * p < 0.1.

), and the AZL policy (Column (6) of Table 9). We observe a positive and statistically significant coefficient on the interaction term FLUIDITY*EV, indicating that the risk-increasing effect of competition, which raises the likelihood of ZL adoption, is amplified when earnings volatility (EV) is higher.4

Following the methodology of Ai and Norton (2003) and Norton et al. (2004), the interaction between FLUIDITY and earnings volatility (EV) was examined using conditional marginal effects rather than raw coefficients. Specifically, we calculated the conditional marginal effects of FLUIDITY on the probability of ZL and AZL at EV levels of 0, 0.5, and 1. The results, presented in

Table 10. Marginal effects of FLUIDITY on ZL and AZL at different levels of earnings volatility. This table presents the marginal effects (dy/dx) of FLUIDITY on the probability of ZL and AZL across varying levels of earnings volatility (EV). Standard errors are reported in parentheses, and statistical significance is indicated by *** p < 0.01, * p < 0.1.

	dy/dx	dy/dx
	ZL	AZL
at: Earnings Volatility (EV) = 0	0.0027 *	0.0047 ***
	(0.0011)	(0.0013)
at: Earnings Volatility (EV) = 0.5	0.0029 *	0.0053 *
	(0.0014)	(0.0022)
at: Earnings Volatility (EV) = 1	0.0032 *	0.0058 *
	(0.0016)	(0.0029)
Obs.	59722	59966

Standard errors are in parentheses—*** p < 0.01, * p < 0.1.

, show that the marginal effects are statistically significant and increase as EV rises. For example, at EV = 0, the marginal effect of FLUIDITY on ZL is 0.0027 (p < 0.05), while for AZL, it is 0.0047 (p < 0.001). These effects further increase at EV = 1, with marginal effects of 0.0032 (p < 0.05) for ZL and 0.0058 (p < 0.05) for AZL. These findings confirm that the interaction is not only statistically significant but also meaningful in the context of the model.

In fiercely competitive markets, maintaining consistent earnings becomes increasingly difficult due to factors such as pricing pressures, shifting customer preferences, and heightened market uncertainty. Intense competition often results in price wars and aggressive pricing tactics that erode profit margins, while fluctuations in demand, sales, and earnings contribute to greater volatility. This increased earnings volatility amplifies the risks associated with debt financing, making it more challenging for firms to meet debt obligations in an unstable environment. Consequently, as PMC rises and market uncertainty intensifies, firms become more likely to adopt zero-leverage (ZL) policies. By avoiding debt, firms reduce their exposure to the risks of financial distress and bankruptcy, preserving financial flexibility to better navigate an unpredictable market landscape. This key finding highlights how earnings volatility moderates the relationship between competition and the strategic choice to adopt ZL policies, offering firms a risk-averse approach to managing financial stability.

5. Robustness Check

5.1. (HHI) as a Proxy for PMC

To validate our findings, we conducted a robustness check using the Herfindahl–Hirschman Index (HHI) as a proxy for PMC. Despite the limitations of the HHI discussed in Section 3, we use it as a robustness check because it is a widely accepted and common measure in the literature. This allows us to ensure that our results are not sensitive to the choice of competition measure and are robust across different proxies for PMC. The HHI, a measure of market concentration, is calculated as the sum of the squared market shares of all firms in the same industry. Higher HHI values indicate increased concentration and reduced competition. This relationship suggests that as industry concentration rises, dominant firms gain more control, limiting competition. Regulators use the HHI for antitrust evaluations (Laksmana & Yang, 2014; Le et al., 2021).

To align the normalized HHI with fluidity, we multiply its values by −1. This adjustment ensures consistency and facilitates comparison. Results in

Table 11. Robustness Check Models. This table presents findings derived from logistic regression models, wherein the dependent variable is represented as a dummy variable indicating either a ZL policy in Models 7 and 9 or an AZL policy in Models 8 and 10. To bolster the robustness of our analysis, we performed a validation check using the Herfindahl–Hirschman Index (HHI) as a proxy for product market competition. In these models, the key independent variable is “HHI” for Models 7 and 8. Meanwhile, Models 9 and 10, introduce earnings volatility (EV) and the interaction term of HHI and EV as additional independent variables. Across all models, we maintained a consistent set of core control variables, including dividend, size, market to book ratio, profitability, and tangibility. This dataset comprises information on product market competition (Fluidity) from the Hoberg and Phillips Data Library and other financial variables sourced from the Compustat (North America) database, covering the period from 1989 to 2019. Table 1 provides descriptions of all variables in detail. The robust standard error adjusted for heteroskedasticity is used for all regressions and is reported in parentheses “***” and “**” denote statistical significance levels at 1% and 5%, respectively.

	(7)	(8)	(9)	(10)
	ZL	AZL	ZL	AZL
HHI	0.6499 **	0.4041 **	0.4496 ***	0.1810 ***
	(0.2708)	(0.1616)	(0.0428)	(0.0157)
Earnings Volatility (EV)			0.2506 **	0.3611 **
			(0.1044)	(0.1389)
HHI × EV			4.9747 **	6.1069 **
			(2.0305)	(2.7449)
Dividend	11.5926 ***	8.9485 ***	12.6183 ***	9.8348 ***
	(1.0426)	(0.9533)	(1.2129)	(1.1129)
Size	−0.3699 ***	−0.3327 ***	−0.3804 ***	−0.3542 ***
	(0.0131)	(0.0111)	(0.0151)	(0.0130)
Market/Book Ratio	0.1368 ***	0.1993 ***	0.1467 ***	0.2030 ***
	(0.0095)	(0.0091)	(0.0124)	(0.0115)
Profitability	0.7323 ***	0.7534 ***	0.7216 ***	0.7953 ***
	(0.0764)	(0.0669)	(0.0945)	(0.0849)
Tangibility	−3.5654 ***	−3.2064 ***	−3.4966 ***	−3.0649 ***
	(0.1676)	(0.1236)	(0.1882)	(0.1412)
Constant	−1.2403 **	−0.1600	−1.3968 **	−0.0381
	(0.4992)	(0.3585)	(0.6379)	(0.4060)
Obs.	107432	107484	84210	84210
Pseudo R2	0.1806	0.1944	0.1888	0.1966
Industry Dummy	Yes	Yes	Yes	Yes
Year Dummy	Yes	Yes	Yes	Yes
Wald chi2 statistic			4.95	4.33
Prob > chi2			0.026	0.031

Standard errors are in parentheses—*** p < 0.01, ** p < 0.05.

show that an alternative proxy for PMC, the HHI, supports our primary finding: an increased likelihood of ZL/AZL policy implementation with rising competition, aligning with the risk-increasing argument and supporting our baseline inference. Coefficients for the HHI variable, earnings volatility, and their interaction term are positive and statistically significant, corroborating our primary finding.

5.2. Alternative Earnings Volatility Measures

In the primary analysis, we utilized a 10-year earnings volatility (EV) window to measure the impact of product market competition (PMC) on firms’ adoption of zero-leverage (ZL) policies. Using a 10-year window allows for the smoothing of cyclical variations and accounts for strategic financial decisions that often unfold over extended periods. This approach mitigates the risk of short-term fluctuations—such as temporary economic or industry-specific shocks—distorting the results, thus providing a more accurate representation of a firm’s long-term financial stability. This methodology is consistent with prior studies in the literature, ensuring comparability while preserving the robustness of the findings.

However, to further address concerns that our results might be sensitive to the choice of the time window for measuring earnings volatility, we conducted additional robustness checks using alternative time horizons. Specifically, we re-estimated models (5) and (6) of Table 11 by replacing the 10-year earnings volatility variable (EV) with a 3-year earnings volatility (EV3) (columns (11) and (12) of

Table 12. Robustness check: logistic regression models for EV3 and EV5 in Interaction Term. This table reports results from logistic regression models, with the dependent variable being represented as a dummy variable indicating either a ZL policy in Models 11 and 13 or an AZL policy in Models 12 and 14. The primary independent variables in these models are “Fluidity”, which measures product market competition, EV3 EV5, and the interaction terms of Fluidity and EV3 and EV5. This dataset comprises information on product market competition (Fluidity) from the Hoberg and Phillips Data Library and other financial variables sourced from the Compustat (North America) database, covering the period from 1989 to 2019. Table 1 provides descriptions of all variables in detail. The robust standard error adjusted for heteroskedasticity is used for all regressions and is reported in parentheses. “***”, “**”, and “*” denote statistical significance levels at 1%, 5% and 10%, respectively.

	(11)	(12)	(13)	(14)
	ZL	AZL	ZL	AZL
FLUIDITY	0.0224 ***	0.0277 ***	0.0173 *	0.0268 ***
	(0.0080)	(0.0066)	(0.0096)	(0.0079)
Earnings Volatility (EV3)	0.5480	0.6701
	(0.4445)	(0.3767)
FLUIDITY × EV3	0.0245 *	0.0253 *
	(0.0133)	(0.0119)
Earnings Volatility (EV5)			0.6631	0.2544
			(0.5333)	(0.4652)
FLUIDITY × EV5			0.0187 *	0.0066 *
			(0.0089)	(0.0080)
Dividend	12.3428 ***	10.4227 ***	12.4227 ***	10.5655 ***
	(1.4040)	(1.2483)	(1.4485)	(1.2994)
Size	−0.3948 ***	−0.3591 ***	−0.3990 ***	−0.3691 ***
	(0.0161)	(0.0136)	(0.0173)	(0.0147)
Market/Book Ratio	0.0912 ***	0.1344 ***	0.0882 ***	0.1322 ***
	(0.0121)	(0.0108)	(0.0134)	(0.0120)
Profitability	0.5376 ***	0.4967 ***	0.5182 ***	0.5417 ***
	(0.1028)	(0.0922)	(0.1118)	(0.1017)
Tangibility	−4.1059 ***	−3.6942 ***	−3.9989 ***	−3.6193 ***
	(0.2356)	(0.1749)	(0.2433)	(0.1843)
Age	−0.0260	−0.0298	−0.0690 **	−0.0451 *
	(0.0238)	(0.0205)	(0.0319)	(0.0272)
R&D	0.0889 ***	0.0918 ***	0.1009 ***	0.0993 ***
	(0.0172)	(0.0181)	(0.0186)	(0.0199)
Capex	3.5531 ***	3.7701 ***	3.4818 ***	3.7208 ***
	(0.4884)	(0.3785)	(0.5344)	(0.4183)
Asset Sale	1.7986 ***	1.5117 ***	1.8815 ***	1.5634 ***
	(0.1540)	(0.1427)	(0.1653)	(0.1526)
Tax	9.3155 ***	10.1001 ***	9.1044 ***	9.9772 ***
	(0.6966)	(0.5849)	(0.7411)	(0.6246)
Constant	−1.0923 **	−0.0871	−1.0745 *	0.0279
	(0.4977)	(0.4002)	(0.6312)	(0.4372)
Obs.	66056	66368	58696	58939
R2	0.1864	0.1974	0.1890	0.1984
Industry Dummy	Yes	Yes	Yes	Yes
Year Dummy	Yes	Yes	Yes	Yes

Standard errors are in parentheses—*** p < 0.01, ** p < 0.05, * p < 0.1.

), and a 5-year earnings volatility (EV5) variable (columns (13) and (14) of Table 12). The results across these alternative measures of earnings volatility remained consistent with those obtained using the 10-year window. This suggests that our key findings—particularly the impact of earnings volatility on firms’ adoption of ZL policies in the context of heightened competition—are not sensitive to the choice of the time window used to measure earnings volatility.

5.3. Addressing Endogeneity Issues

Investigating the link between competition and our dependent variables incurs the challenge of endogeneity, particularly regarding the potential for reverse causality in the relationship between PMC, ZL behavior, and earnings volatility. Firms with higher earnings volatility may adopt ZL to reduce financial risk, while firms adopting ZL may aim for competitiveness, leading to higher earnings volatility. To address this, we use an instrumental variable (IV) approach, where we instrument for competition using the number of firms in each industry (Industry_firms) based on Le et al. (2021). This choice is motivated by its strong correlation with competition, indicating a more competitive environment.

Our initial test in

Table 13. Correlation between Instrumental Variable and Fluidity. This table provides a validation test assessing the suitability of using the number of firms within each industry as a valid instrumental variable for Fluidity. We specifically report results of regression analyses in which the dependent variable is “Fluidity” and the primary independent variable under consideration is the count of firms operating within each industry. Model 15 includes control variables such as dividend, size, market to book ratio, profitability, and tangibility. This dataset comprises information on product market competition (Fluidity) from the Hoberg and Phillips Data Library and other financial variables sourced from the Compustat (North America) database, covering the period from 1989 to 2019. Table 1 provides descriptions of all variables in detail. The robust standard error adjusted for heteroskedasticity is used for all regressions and is reported in parentheses. “***” denote statistical significance levels at 1%.

	(15)
	FLUIDITY
Industry_firms	0.0005 ***
	(0.0001)
Dividend	−0.5021
	(0.9016)
Size	0.3318 ***
	(0.0151)
Market/Book Ratio	0.1106 ***
	(0.0082)
Profitability	−1.4534 ***
	(0.0693)
Tangibility	−0.0889
	(0.1275)
Constant	5.3860 ***
	(0.2968)
Obs.	107484
R2	0.3254
Industry Dummy	Yes
Year Dummy	Yes

Standard errors are in parentheses—*** p < 0.01.

validates the instrument’s relevance, showing a significant positive correlation between Industry_firms and FLUIDITY. Using two-stage least squares (2SLS) in

Table 14. Instrumental Variable Models. In this table, we report results of logistic regression models, where the dependent variable takes the form of a ZL policy dummy variable for Models 16 and 18 or an AZL policy dummy variable for Models 17 and 19. These models represent the second stage of 2SLS (two-stage least squares) logistic regressions, designed to address the endogeneity issue in our analysis. In these models, the primary independent variable is “Instrumented Fluidity”, a variable derived from the number of firms within each industry. In Models 18 and 19, we introduce earnings volatility (EV) and the interaction term of Instrumented fluidity and EV as additional independent variables. All models include control variables, such as dividend, size, market to book ratio, profitability, and tangibility. This dataset comprises information on product market competition (Fluidity) from the Hoberg and Phillips Data Library and other financial variables sourced from the Compustat (North America) database, covering the period from 1989 to 2019. Table 1 provides descriptions of all variables in detail. The robust standard error adjusted for heteroskedasticity is used for all regressions and is reported in parentheses. “***”, “**”, and “*” denote statistical significance levels at 1%, 5% and 10%, respectively.

	(16)	(17)	(18)	(19)
	ZL	AZL	ZL	AZL
FLUIDITY (Instrumented)	0.1632 ***	0.1920 ***	0.1615 ***	0.1618 ***
	(0.0451)	(0.0433)	(0.0540)	(0.0500)
Earnings Volatility (EV)			0.0360 **	0.0434 **
			(0.0161)	(0.0189)
FLUIDITY (Instrumented) × EV			0.0004 *	0.0024 *
			(0.0003)	(0.0016)
Dividend	12.9571 ***	10.4677 ***	12.9582 ***	10.4440 ***
	(1.2585)	(1.1250)	(1.3006)	(1.1744)
Size	−0.4265 ***	−0.4091 ***	−0.4304 ***	−0.4088 ***
	(0.0180)	(0.0168)	(0.0209)	(0.0192)
Market/Book Ratio	0.1450 ***	0.1988 ***	0.1472 ***	0.2018 ***
	(0.0114)	(0.0107)	(0.0128)	(0.0118)
Profitability	0.9016 ***	0.9774 ***	0.8817 ***	0.9907 ***
	(0.0875)	(0.0794)	(0.0988)	(0.0912)
Tangibility	−3.6124 ***	−3.1382 ***	−3.5472 ***	−3.1023 ***
	(0.1854)	(0.1358)	(0.1961)	(0.1460)
Constant	−1.8565 ***	−0.9444 **	−1.9271 ***	−0.7586 *
	(0.5167)	(0.4167)	(0.6008)	(0.4560)
Obs.	89861	89900	79750	79750
Pseudo R2	0.1857	0.1956	0.1880	0.1969
Industry Dummy	Yes	Yes	Yes	Yes
Year Dummy	Yes	Yes	Yes	Yes
Wald chi2 statistic			4.83	5.71
Prob > chi2			0.028	0.011
Weak identification test (Cragg-Donald Wald F statistic) (Model 12):			13.236
Stock–Yogo weak ID test critical values (Model 12):
	10% maximal IV size		16.38
	15% maximal IV size		8.96
	20% maximal IV size		6.66
	25% maximal IV size		5.53
Source: Stock and Yogo (2005). Reproduced by permission.

Standard errors are in parentheses—*** p < 0.01, ** p < 0.05, * p < 0.1.

, we find that, even after addressing endogeneity, the fluidity coefficient remains significant and positive for the ZL policy. The Cragg–Donald Wald F-statistic and Stock–Yogo test confirm the instrument’s strength, rejecting the null hypothesis of weak instruments. Different critical values support the robustness of the instrument, with the statistic closer to the 10% value.

Examining our second hypothesis in Table 13, we show that the instrumented fluidity variable confirms the persistence of positive and statistically significant coefficients for earnings volatility and the interaction term. These consistent results strengthen the validity of our findings, affirming that heightened PMC and increased earnings volatility significantly influence the adoption of ZL policies.

Additionally, we apply a one-year lag to independent variables (excluding dummies) to mitigate endogeneity issues, enhancing the robustness of our analysis.

6. Conclusions

In this study, we examine the relationship between product market competition (PMC), zero-leverage (ZL) policies, and earnings volatility, utilizing a comprehensive dataset spanning from 1989 to 2019. We employ fluidity data from the Hoberg and Phillips Data Library to measure PMC, while identifying zero-leveraged firms based on the framework established by Strebulaev and Yang (2013). Earnings volatility is captured through the variability in profitability over the past 10 years, with control variables derived from the Compustat (North America) database.

Our empirical analysis supports the first hypothesis, demonstrating that firms operating in highly competitive markets are more likely to adopt ZL policies. Moreover, the second hypothesis is confirmed, with heightened PMC leading to an increased probability of both ZL and almost-zero-leverage (AZL) adoption, particularly in firms exhibiting higher earnings volatility. The significant interaction between PMC and earnings volatility further strengthens this conclusion, highlighting the compounded effect of heightened competition on capital structure decisions.

To ensure the robustness of our findings, we conducted a Herfindahl–Hirschman Index (HHI) robustness check, which affirmed the stability of our results. Additionally, we addressed potential endogeneity concerns by using an instrumental variable, further validating the credibility of our conclusions.

This study contributes to the literature by providing new insights into how PMC influences the adoption of ZL policy, particularly through the lens of earnings volatility. Our analysis highlights the dynamic relationship between competition, risk, and financial policy, adding to the growing body of research on the role of competition in shaping corporate financing decisions. For example, Boubaker et al. (2018) demonstrate that competition reduces firms’ reliance on strict bank debt, while Do et al. (2022) find that competitive pressures accelerate leverage adjustments. Our study complements these findings by focusing on how heightened competition and earnings volatility interact to drive firms toward adopting ZL policies as a strategy to mitigate financial risk. Similarly, Hajda and Nikolov (2022) show how industry dynamics and product life cycles influence firms’ financing decisions, emphasizing the importance of understanding the broader competitive environment. In line with Cao and Chen (2023), who examine the role of CEO turnover and competition in leverage policy, our study underscores how competition influences risk management strategies in the absence of debt financing. Finally, studies such as Cumming et al. (2022) and Zheng et al. (2021) explore the broader implications of PMC on firm success and cost of capital, aligning with our findings that heightened competition increases risk and uncertainty, prompting firms to adopt ZL strategies as a safeguard against potential financial distress.

However, our study has several limitations that should be considered when interpreting the results. First, our sample excludes small and private firms that do not have publicly available 10-K reports, which limits the generalizability of our findings. This exclusion could bias the results, as these firms may operate under different competitive pressures or capital structure decisions than their publicly traded counterparts. Second, while our analysis focuses on U.S. firms, the results may not fully capture the competitive dynamics present in other geographical markets, especially considering differences in industry concentration and regulatory environments across countries. Future research could explore the cross-country variation in the relationship between PMC and capital structure decisions.

In conclusion, our study underscores the critical role of product market competition in shaping corporate capital structure decisions, particularly in environments characterized by high earnings volatility. These findings have important implications for understanding risk management strategies and the strategic choices firms make to navigate competitive pressures. Future research could expand the analysis to include a broader range of firms, industries, and geographical markets, as well as explore alternative measures of risk and competition, to further enhance our understanding of how firms adapt to competitive dynamics in their capital structure choices.