Unequal Grounds and Unstable Markets: Income Inequality and Stock Price Crash Risk

Abstract

This study analyzes the relationship between country-level income inequality and stock price crash risk using a comprehensive cross-country panel of 117,017 firm-year observations from 45 countries spanning 2000–2022. We document that firms headquartered in countries with higher income inequality face a significantly greater likelihood of experiencing stock price crashes. Building on behavioral finance theory, we argue that income inequality exacerbates managerial incentives to withhold negative information, thereby increasing crash risk. We further show that this relationship is moderated by both country-level and firm-level mechanisms that influence information transparency. Specifically, stronger national transparency, greater institutional ownership, and lower excess cash weaken the positive association between income inequality and crash risk. Our results remain robust across alternative crash risk measures and endogeneity tests, including instrumental variable and propensity score matching approaches. These findings highlight income inequality as an important macro-level determinant of financial market instability and underscore the role of transparency and monitoring in mitigating its adverse effects on capital markets.

1. Introduction

Investors are naturally worried about sudden drops in stock prices because it directly affects their financial well-being. This has sparked considerable academic interest in unraveling the complex reasons behind stocks’ vulnerability to major downturns. Given the fact that the tendency of managers to hoard bad news is one key factor contributing to the heightened stock price crash risk (Al Mamun et al., 2020; Y. Kim et al., 2014), factors affecting this tendency have a direct effect on stock price crash risk. While most research in this area examines the influence of firm-specific or CEO-specific elements on the likelihood of stock price crashes, there is a growing body of work looking into how country-level factors come into play (An et al., 2018; Yung & Askarzadeh, 2025). Despite the expanding literature on country-level determinants of stock price crash risk, existing studies have largely focused on legal, cultural, or governance-related factors, paying limited attention to income inequality as a fundamental macroeconomic characteristic that shapes information environments. This emerging interest along with documented sources about the effects of disparity in income on individuals’ behavior (e.g., Barone & Mocetti, 2016; Desai et al., 2009) lay the groundwork for our study, which aims to address the gap between income inequality and stock price crash risk by explaining how its effect operates through information asymmetry–related mechanisms.

We argue that income inequality, defined as the gap between the rich and the poor due to uneven income distribution or income growth across a population, heightens the risk of stock price crashes in two main ways. Initially, drawing upon Desai et al. (2009), we propose that in countries marked by pronounced income inequality, managers are likely to perceive themselves as more powerful compared to others. This self-perception emboldens them to adopt exploitative behaviors for personal gain. At the corporate level, Bebchuk et al. (2011) noted that a gap in CEO compensation signals increased managerial authority and power to control the flow of information. CEOs wielding more power have the leeway to focus on personal goals, which may not necessarily serve the shareholders’ interests. Such powerful CEOs might pressure CFOs into presenting skewed financial performance metrics (Feng et al., 2011; Friedman, 2014). Moreover, these CEOs often escape rigorous scrutiny from board members, enabling them to delay the disclosure of adverse information.

Secondly, we posit that income inequality is intertwined with societal factors that influence corporate outcomes. We argue that income inequality influences social dynamics such as trust among people and levels of corruption. The significant negative impact of income inequality on social trust is well-documented (Barone & Mocetti, 2016). Social trust plays a crucial role in shaping individual behaviors (Cialdini & Goldstein, 2004). In countries where social trust prevails, informal networks flourish over time, influencing executives’ actions and the flow of information. Consequently, in environments rich in social trust, executives are propelled to act more ethically in their business dealings, even if they may not be inherently trustworthy, due to the expectations and influences of their reputable peers and business associates. In such environments, adverse news is disseminated more swiftly to the market, thereby lowering the risk of stock price crashes.

Following Jong-Sung and Khagram (2005) and Policardo and Carrera (2018), we propose that income inequality has a direct relationship with incidents of fraud and corruption. Specifically, in areas with substantial income disparity, individuals with low incomes often lack the financial resources required for legal challenges against the more affluent, thus creating an environment where fraud among the wealthy becomes increasingly prevalent over time (Frank, 2019). In such settings, those with lower incomes might also turn to fraudulent means as a method to enhance their social status. Dal Bó and Rossi (2007) indicate that, under these conditions, operational efficiency tends to be significantly lower compared to firms in regions with lower levels of corruption. Moreover, in countries plagued by high levels of corruption, the generation of negative news becomes more frequent (Djankov et al., 2003). P. Cao et al. (2019) further argue that in such corrupt environments, managers are more inclined to withhold negative news, thereby increasing the likelihood of stock price crashes.

We contend that income inequality amplifies the incentive for CEOs to suppress unfavorable news. This dynamic is further shaped by elements contributing to information asymmetry. Because stock price crash risk arises from the accumulation of undisclosed negative information, variables that influence information asymmetry are expected to influence the strength of the relationship between income inequality and crash risk. We therefore focus on variables that affect monitoring effectiveness as a key link to information asymmetry. Specifically, information transparency at the country level strengthens external monitoring and reduces information asymmetry, institutional investor ownership enhances firm-level monitoring and constrains managerial discretion, while excess cash weakens monitoring effectiveness by reducing firms’ reliance on external capital markets. Together, these variables determine whether income inequality translates into greater managerial information hoarding and higher stock price crash risk.

Utilizing a dataset comprising 117,017 firm-year observations across 45 different countries, spanning from 2000 to 2022, our empirical analysis underpins our hypotheses, demonstrating a significant positive correlation between country-level income inequality, measured by the Gini index, and the stock price crash risk. Moreover, our study reveals that greater country-level transparency, a higher proportion of institutional ownership, and lesser excess cash negatively moderate the relationship between income inequality and the likelihood of stock price crashes. To counter potential concerns of endogeneity, we applied a two-stage least squares regression and a propensity score matching method, confirming the stability and reliability of our initial findings.

Beyond identifying this relationship, our contribution lies in clarifying the mechanisms through which income inequality affects crash risk. We argue that income inequality influences managerial incentives and disclosure behavior by shaping information asymmetry at both the societal and firm levels. Furthermore, by examining country-level transparency and firm-level monitoring mechanisms, specifically institutional ownership and excess cash, as moderating variables, we provide a more nuanced understanding of when and under what conditions income inequality is more likely to translate into heightened crash risk. In doing so, this study integrates macro-level income distribution with micro-level governance forces and offers new insights into the role of social inequality in financial market stability.

The structure of the rest of the article is as follows: Section 2 reviews related literature and formulates hypotheses, Section 3 details the data sources and introduces the main variables used, Section 4 describes our methodological approach and presents the findings, and Section 5 concludes.

2. Literature Review and Hypothesis Development

Stock price crash risk has attracted significant scholarly attention, with research identifying contributing factors operating at the individual (managerial), firm, and country levels. From a behavioral finance perspective, stock price crashes are commonly attributed to managers’ incentives and behavioral biases, such as overconfidence, self-interest, and reputational concerns, which lead to the strategic hoarding of bad news followed by its sudden release. In this context, managers utilize their privileged access to detailed company operations to exercise power over the timing and nature of information disclosure (Askarzadeh et al., 2023).

At the individual level, prior studies emphasize managerial traits and power. J.-B. Kim et al. (2016) show that firms led by overconfident CEOs face a higher likelihood of stock price crashes, consistent with behavioral finance theories suggesting that overconfidence distorts risk perception and disclosure decisions. Similarly, Al Mamun et al. (2020) document that powerful CEOs are more prone to behaviors that elevate crash risk, as increased discretion weakens internal constraints and amplifies incentives to conceal unfavorable information.

At the firm level, the literature highlights organizational characteristics that influence information transparency and monitoring. Studies examine the role of corporate social responsibility (Y. Kim et al., 2014), corporate tax avoidance (J.-B. Kim et al., 2011), and accounting conservatism (J.-B. Kim & Zhang, 2016), all of which affect the firm’s information environment. These firm-level attributes shape managers’ ability and willingness to delay the disclosure of bad news, reinforcing the behavioral mechanism underlying crash risk.

Beyond individual and firm-level factors, recent studies increasingly focus on country-level environments, which shape managerial behavior through formal and informal institutions. Return-anomaly research suggests that institutional settings influence how behavioral biases manifest in markets. For example, Mojtahedi et al. (2025) show that behavioral biases can sustain mispricing in developed markets, while Mashhadi et al. (2025) demonstrate that regulatory constraints in emerging markets can limit such effects.

Cultural and institutional characteristics further affect crash risk by shaping social norms and incentives. An et al. (2018) find that individualistic cultures are associated with higher crash risk, whereas Apergis (2017) shows that firms in more democratic countries are less prone to crashes. Within the United States, Callen and Fang (2015) document that higher local religiosity is linked to lower crash risk. Collectively, these findings indicate that country-level institutional and cultural environments influence managerial incentives to hoard bad news, a core behavioral mechanism driving stock price crashes.

2.1. Income Inequality and Stock Price Crash Risk

Building on this multilevel framework, our study examines how country-level income inequality affects stock price crash risk. Drawing on behavioral finance, we argue that income inequality amplifies crash risk through two interconnected channels: (i) an individual-level channel, by increasing CEO power and discretion; and (ii) a societal-level channel, by weakening social trust and strengthening fraud and corruption. This dual-channel approach offers a more nuanced understanding of how income inequality influences stock market stability.

2.1.1. Individual-Level Channel: Income Inequality and CEO Power

The literature on income inequality has expanded considerably over the past two decades, particularly regarding its implications for managerial behavior. Bebchuk et al. (2011) show that rising disparities in CEO compensation increase the centralization of managerial power. From a behavioral perspective, greater power allows CEOs to act on self-serving motives that may diverge from shareholder interests (Daily & Johnson, 1997).

At the country level, Desai et al. (2009) document that large pay gaps between managers and lower-income workers strengthen managerial authority, potentially increasing social distance between executives and employees. Tsui et al. (2018) similarly argue that income inequality heightens social distance, which can weaken moral constraints on managerial behavior.

Powerful CEOs face weaker monitoring and fewer behavioral constraints. Friedman (2014) finds that such CEOs pressure CFOs to manipulate performance metrics, while Feng et al. (2011) show that firms led by influential CEOs are more likely to engage in accounting manipulation. Al Mamun et al. (2020) further document that powerful CEOs face less stringent oversight. In behavioral finance terms, reduced monitoring strengthens incentives for bad-news hoarding, increasing the likelihood of sudden negative price corrections.

2.1.2. Societal-Level Channel: Income Inequality, Social Trust, and Corruption

Beyond individual effects, income inequality influences broader societal conditions that shape corporate behavior. Prior research documents a strong link between income inequality and lower social trust (Graafland & Lous, 2019) as well as higher fraud and corruption (Jong-Sung & Khagram, 2005). Elgar (2010) reports a negative relationship between income inequality and societal trust across 33 countries.

Income inequality alters social interactions by limiting cross-group engagement and information sharing. Personal networks tend to be homophilous (McPherson et al., 2001), and greater inequality intensifies what Bjørnskov (2008) describes as social fractionalization. From a behavioral standpoint, weaker social ties reduce norms of trustworthiness (Coleman, 1988), lowering the psychological cost of opportunistic behavior.

Following C. Cao et al. (2016) and Li et al. (2017), we argue that social trust affects stock price crash risk by shaping informal institutions and disclosure norms. In high-trust societies, executives are more likely to disclose bad news in a timely manner, reducing the accumulation of hidden negative information and, consequently, crash risk.

Income inequality also increases fraud and corruption (Jong-Sung & Khagram, 2005). As inequality rises, wealthy individuals possess greater incentives and resources to influence political and legal systems (Glaeser et al., 2003). Corruption weakens investor protection (La Porta et al., 2000), heightens information asymmetry, and facilitates insider expropriation. These conditions encourage managers to conceal adverse information, reinforcing the behavioral mechanism of bad-news hoarding.

Taken together, income inequality affects crash risk by shaping CEO behavior, social norms, and institutional quality, leading to greater accumulation of negative information. Accordingly, we propose:

H1. 

Firms located in countries with higher income inequality are more prone to stock price crash risk.

2.2. Moderating Effect of Country-Level Transparency

We propose that the increased risk of stock price crashes due to income inequality stems from the information asymmetry between managers and investors. Specifically, in environments where income inequality is stark, managers often see themselves in a position that allows them to withhold negative news from investors. This information asymmetry can be lessened by increasing CEO monitoring. While much of the literature focuses on firm-level factors influencing CEO monitoring, some studies have examined how country-level factors shape information asymmetry between managers and shareholders. For instance, An et al. (2020) identify higher media coverage within a country as a key factor in reducing information asymmetry, as it decreases managers’ propensity to withhold negative information. Similarly, Duong et al. (2022) explore the impact of democracy on IPO underpricing, arguing that firms located in countries with greater democratic governance experience lower levels of information asymmetry.

By examining the effect of country-level transparency on stock price crash risk using data from 37 countries, Abedifar et al. (2019) found that firms in countries with higher accounting transparency and stringent enforcement regulations tend to have a lower crash risk. This suggests that transparency and regulatory enforcement play vital roles in curtailing managers’ incentives to withhold adverse news. Supporting this view, An et al. (2015) state that in more transparent information environments, firm-level information is shared with the market more accurately and promptly. Porta et al. (1998) also affirm that financial market efficiency is enhanced in settings characterized by increased information transparency. This body of research collectively underscores the importance of transparency in influencing market dynamics and mitigating the risk of stock price crashes. Therefore, in our next hypothesis, we posit that a transparent environment can reduce information asymmetry, thereby weakening the positive relationship between income inequality and stock price crash risk.

H2. 

The positive relationship between income inequality and stock price crash risk is attenuated when country-level transparency is higher.

2.3. Firm-Level Moderating Variables

In the context of firm-level variables, the literature suggests that specific factors can alleviate the information asymmetry and subsequently the risk of stock price crashes. Chen et al. (2013) highlight that IPO firms typically display high levels of information asymmetry, which can be mitigated through the involvement of institutional investors. Shleifer and Vishny (1997) underscore the critical role that institutional investors play in curbing agency problems within firms.

Institutional investors are usually better informed and incorporate their information into the transparency level of the firms (Sias et al., 2006). Therefore, with their superior firm knowledge and significant influence over management, institutional investors are naturally more inclined to engage in monitoring activities than other investors. This is particularly true for stable institutional investors who demonstrate both the motivation and capacity for effective monitoring (Elyasiani & Jia, 2008). Their vital role in reducing agency conflicts and information risk is supported by the findings of Elyasiani et al. (2010), who report a significant negative relationship between institutional ownership stability and a firm’s cost of debt.

Moreover, the level of cash holdings within a firm can impact information asymmetry. Jensen (1986) argue that higher free cash flows lead to greater agency issues as managers have more capital at their disposal, potentially leading to inefficient investment decisions. Chung et al. (2015) show a negative relationship between information asymmetry and corporate cash holdings. In line with these studies, Choi et al. (2020) use cash holdings as an indicator of information asymmetry and find a positive link between higher cash holdings and increased investment inefficiency. Holding larger cash reduces the need to dependency on the market to raise funds, resulting in less pressure for transparency of financial statements. The literature demonstrates that cash holders have less transparent financial statements and experience higher level of information asymmetry (Gryko et al., 2024).

Based on these insights, we suggest that reduced information asymmetry, facilitated through increased monitoring by institutional investors, could lower managers’ incentives to withhold negative news, thereby weakening the link between income inequality and stock price crash risk. Additionally, we posit that higher excess cash can exacerbate monitoring ineffectiveness and information asymmetry, potentially intensifying the relationship between income inequality and the risk of stock price crashes. Based on the impact of these variables on the underlying mechanism role of information asymmetry on the link between income inequality and stock crash risk, our subsequent hypotheses are as follows:

H3a. 

The positive relationship between income inequality and stock price crash risk is attenuated when institutional ownership is higher.

H3b. 

The positive relationship between income inequality and stock price crash risk is enhanced when firm excess cash is higher.

3. Model Specification and Data

3.1. Crash Risk Measures

Following An et al. (2018), we use Equation (1) to calculate the firm-specific weekly return. The following regression is estimated for each firm-year:

$$\begin{array}{cc}\hfill r_{i,t}=& {\alpha }_0+{\alpha }_1{r}_{m,j,t-2}+{\alpha }_2{r}_{m,j,t-1}+{\alpha }_3{r}_{m,j,t}+{\alpha }_4{r}_{m,j,t+1}+{\alpha }_5{r}_{m,j,t+2}\hfill \\ \hfill & +{\alpha }_6\left(r_{\mathrm{U}.\mathrm{S}.,t-2}+E{X}_{j,t-2}\right)+{\alpha }_7\left(r_{\mathrm{U}.\mathrm{S}.,t-1}+E{X}_{j,t-1}\right)+{\alpha }_8\left(r_{\mathrm{U}.\mathrm{S}.,t}+E{X}_{j,t}\right)\hfill \\ \hfill & +{\alpha }_9\left(r_{\mathrm{U}.\mathrm{S}.,t+1}+E{X}_{j,t+1}\right)+{\alpha }_{10}\left(r_{\mathrm{U}.\mathrm{S}.,t+2}+E{X}_{j,t+2}\right)+{\epsilon }_{i,t}.\hfill \end{array}$$

(1)

where $r_{i,t}$ is the stock return of firm i in week t, $r_{m,j,t}$ is the local market return for country j in week t, $r_{\mathrm{U}.\mathrm{S}.,t}$ is the return of the U.S. market in week t, and $E{X}_{j,t}$ is the change in country j’s exchange rate versus the U.S. dollar in week t. We include two lags and two leads to account for non-synchronous trading.

In the next stage, we calculate the residual from Equation (1), ${\epsilon }_{i,t}$, which represents the portion of the weekly stock return not explained by either the local or global market and is therefore related to firm-specific idiosyncratic factors. We then log-transform this residual to compute the firm-specific weekly return as the natural logarithm of one plus the residual return, i.e., $W_{i,t}=log\left(1+{\epsilon }_{i,t}\right)$.

We use three measures for stock price crash risk. These crash-risk measures have been widely used in recent international studies, including Askarzadeh et al. (2024), who apply NCSKEW, DUVOL, and CRASH to examine how information opacity in globally diversified firms contributes to extreme negative return outcomes. Our first measure is NCSKEW, which captures the negative skewness of the firm-specific weekly returns and is calculated as follows:

$${\mathrm{NCSKEW}}_{j,T}={\displaystyle \frac{-n{(n-1)}^{3/2}\sum W_{j,T}^3}{(n-1)(n-2){\left(\sum W_{j,T}^2\right)}^{3/2}}},$$

(2)

where n is the number of firm-specific weekly returns during year T for firm i.

The next measure of crash risk is the “down-to-up volatility” (DUVOL), calculated as follows:

$$\mathrm{DUVOL}=log\left({\displaystyle \frac{(n_u-1){\sum }_{\mathrm{Down}}{W}_{j,T}^2}{(n_d-1){\sum }_{\mathrm{Up}}{W}_{j,T}^2}}\right),$$

(3)

where $n_u$ and $n_d$ are the numbers of up and down weeks over year T, respectively. ${\sum }_{\mathrm{Down}}{W}_{j,t}^2$ and ${\sum }_{\mathrm{Up}}{W}_{j,t}^2$ denote the sums of squared firm-specific weekly returns for down and up weeks. A firm-week is defined as a down (up) week if the firm-specific weekly return is below (above) its annual mean.

Following Hutton et al. (2009), our last measure is CRASH, which is based on the number of firm weekly returns exceeding 3.09 standard deviations above or below the mean firm weekly return over the calendar year. The value of 3.09 is chosen to generate frequencies of 0.1% in the normal distribution.

CRASH is calculated as the number of weeks in which the firm-specific weekly return is more than 3.09 standard deviations below the annual mean (down weeks) minus the number of weeks in which the return is more than 3.09 standard deviations above the annual mean (up weeks). For all three crash measures, a higher value represents a higher level of crash risk.

3.2. Baseline Model

To examine the effect of income inequality, we use the following regression model:

$${\mathrm{Crashrisk}}_{i,j,t+1}={\alpha }_0+{\alpha }_1{\mathrm{Gini}}_{j,t}+{\alpha }_2{X}_{i,j,t}+{\alpha }_{3}X{Y}_{i,j,t}+\mathrm{Industry}\mathrm{FE}+\mathrm{Year}\mathrm{FE}+e_{i,j,t+1}.$$

(4)

where Crashrisk is one of the crash risk variables NCSKEW, DUVOL, and CRASH. We use the Gini index as our explanatory variable to measure income inequality. The Gini index measures the area between the Lorenz curve and a hypothetical line of absolute equality, expressed as a percentage of the maximum area under the line. Thus, a Gini index of 0 represents perfect equality, while an index of 100 implies perfect inequality (Frees et al., 2011). To account for dynamic persistence and ensure proper temporal ordering, all explanatory variables are lagged by one period, allowing past conditions to influence current outcomes while mitigating simultaneity concerns.

X denotes a set of firm-level variables. We include the lagged NCSKEW as our first control variable. Additionally, we have Return, which represents the cumulative firm-specific weekly returns over the fiscal year. We also include SD, which denotes the standard deviation of firm-specific weekly returns over the fiscal year. We also incorporate Kurtosis, indicating the kurtosis of firm-specific weekly returns over the fiscal year. Furthermore, our set of firm-level variables includes MB, which represents the ratio of the market value of equity to the book value of equity measured at the end of the fiscal year. Size denotes the log value of market capitalization at the end of the fiscal year. Leverage represents the book value of all liabilities divided by total assets at the end of the fiscal year. Moreover, we include ROE, which denotes income before extraordinary items divided by the book value of equity at the end of the fiscal year. Additionally, our set of firm-level variables includes Dturn, which represents the average monthly share turnover over the fiscal year minus the average monthly share turnover over the previous year. The monthly share turnover is calculated as the monthly share trading volume divided by the number of shares outstanding over the month.

Y denotes a set of country-level variables. We have GDPgrowth, which is defined as the GDP growth for a firm’s headquarter country in year t. Additionally, GDP Capita is defined as the lagged GDP per capita for a firm’s headquarter country in year t divided by 100. Furthermore, we include CR, which is defined as the country’s creditor rights index. GOVERNANCE is defined as the summation of percentile ranks of six World Bank measures: voice and accountability, political stability, government effectiveness, regulatory quality, rule of law, and control of corruption.

Y denotes a set of country-level variables. Following (An et al., 2018; Yung & Askarzadeh, 2025), we have GDPgrowth, which is defined as the GDP growth for a firm’s headquarter country in year t. Additionally, GDP Capita is defined as the lagged GDP per capita for a firm’s headquarter country in year t divided by 100. Furthermore, we include CR, which is defined as the country’s creditor rights index. GOVERNANCE is defined as the summation of percentile ranks of six World Bank measures: voice and accountability, political stability, government effectiveness, regulatory quality, rule of law, and control of corruption. Moreover, we include GDS, which is defined as the gross domestic saving (percentage of GDP) for a firm’s headquarter country in year t. MARKET is defined as the market value of all stocks in the firm’s headquarter country divided by its GDP in year t. Additionally, we incorporate DCP, which is defined as the Domestic credit to the private sector (percentage of GDP) for a firm’s headquarter country in year t. Industry and year fixed effects are included in all models. By using time-invariant country-level control variable (CR), we control for country fixed effects.

3.3. Sample

To construct a sample that allows for reliable and comparable estimation of stock price crash risk across countries, we apply several exclusion criteria. First, we exclude American Depository Receipts (ADRs) and Global Depository Receipts (GDRs) because these securities are traded outside firms’ home countries and are subject to host-country disclosure rules and market conditions, which may weaken the link between home-country income inequality and firm-level crash risk. Second, we exclude utility firms due to their highly regulated nature, as regulatory oversight can substantially affect pricing behavior, disclosure incentives, and return dynamics. Financial firms are also excluded because their capital structure, leverage, and regulatory environment differ markedly from those of non-financial firms, making crash-risk measures less comparable across industries.

We further exclude firm-year observations with fewer than 26 weekly return observations to ensure reliable estimation of firm-specific weekly returns and crash-risk measures and to reduce the influence of thin trading. Firms with market capitalization below USD 1 million are excluded because extreme illiquidity and price discreteness among micro-cap firms can mechanically inflate volatility-based crash proxies. Finally, we exclude countries with fewer than 100 firm-year observations to avoid drawing inferences from sparse country-level coverage and to ensure sufficient variation for identifying the effect of income inequality. Collectively, these exclusion criteria are consistent with prior international crash-risk studies and are intended to improve measurement precision, reduce noise, and enhance the comparability of crash-risk estimates across countries.

We utilize the World Bank database to obtain our explanatory variable (Gini index) and country-level control variables (GDP growth, GDP per capita, Governance, Market, Gross Domestic Saving (GDS), and Domestic Credit to Private Sector (DCP)). Following Djankov et al. (2007), we acquire the Creditor Rights Index (CR). Additionally, we employ Thomson One Eikon to gather the firm-level control variables (RETURN, SD, KURTOSIS, MB, SIZE, LEVERAGE, ROE, and DTURN). Panel A of Table 1 presents the sample distribution. The average Gini index ranges from 24.86 for Slovenia to 63.06 for South Africa. We define a firm as “crashed” in year t if CRASH is higher than zero. Panel B of Table 1 exhibits the univariate tests of the crash risk variables in countries with Gini index above the cross-country sample median versus countries with Gini index below the cross-country sample median. As indicated in this panel, the mean differences are positive and statistically significant for all three crash measures, denoting that the stock price crash risk is higher in countries with a higher income inequality level.

Table 1. Sample and univariate analysis.

Panel A: Sample
Country	Gini Index	Number of Firms	Crashed Firms (%)	Country		Gini Index	Number of Firms	Crashed Firms (%)
ARGENTINA	44.54677	47	0.0836	KOREA		31.65539	1573	0.0928
AUSTRALIA	34.41243	702	0.1586	MALAYSIA		42.62457	512	0.1085
AUSTRIA	30.54782	44	0.1392	MEXICO		48.72584	101	0.1426
BELGIUM	28.23563	82	0.1410	NETHERLANDS		28.39669	58	0.1781
BRAZIL	53.20802	158	0.1120	NIGERIA		35.82162	48	0.1891
BULGARIA	35.91143	48	0.1028	NORWAY		26.67003	111	0.1403
CANADA	33.41358	731	0.1411	PAKISTAN		29.81973	179	0.0978
CHILE	46.94507	95	0.1267	PERU		46.82079	44	0.1137
CHINA	39.83597	3407	0.1197	PHILIPPINES		44.69363	152	0.0943
COLOMBIA	52.13036	19	0.2023	POLAND		31.76595	255	0.1113
CROATIA	31.19815	41	0.0432	PORTUGAL		36.10734	26	0.1129
EGYPT	31.37240	152	0.0800	RUSSIA		38.66623	93	0.0811
FRANCE	32.50476	402	0.1418	SLOVENIA		24.86289	13	0.1443
GERMANY	31.26421	390	0.1637	SOUTH AFRICA		63.06128	134	0.1197
GREECE	34.51150	123	0.1160	SPAIN		35.19540	109	0.1104
HUNGARY	29.74191	18	0.1176	SRI LANKA		38.16270	176	0.0684
INDIA	35.22140	2025	0.0694	SWITZERLAND		32.61931	144	0.1378
INDONESIA	38.86916	479	0.1213	THAILAND		37.08945	597	0.1411
IRELAND	32.13515	62	0.1838	TURKEY		40.85633	276	0.1063
ISRAEL	40.56778	309	0.1472	UNITED KINGDOM		34.63783	532	0.1569
ITALY	34.40717	114	0.1043	UNITED STATES OF AMERICA		40.89086	2154	0.2235
JAPAN	33.22555	2176	0.1338	VIETNAM		36.20923	759	0.0911
JORDAN	33.17619	84	0.1428	Total		37.47793	19754	0.1380
Panel B: Univariate analysis
	High Income Inequality			Low Income Inequality			Mean diff.	T-stat.
	N	Mean	SD	N	Mean	SD
NCSKEW	58992	−0.108	0.833	59709	−0.230	0.790	0.122	25.88 ***
DUVOL	58992	−0.066	0.544	59709	−0.147	0.525	0.081	26.09 ***
CRASH	58992	−0.104	0.658	59709	−0.178	0.678	0.074	19.08 ***

Panel A reports the distribution of sample firms across 45 countries during the period spanning 2000–2022. A “crashed firm” is defined as a firm with a positive CRASH value. Panel B provides the univariate test results. *** indicates significance at the 1% level. All variables are defined in the Appendix A.

Panel A of Table 2 provides descriptive statistics for the main variables used in our models. The mean values of the stock price crash risk measures NCSKEW, DUVOL, and CRASH are $-0.174$, $-0.110$, and $-0.145$, respectively. The mean value of the Gini index is $37.371$. Panel B of Table 2 displays a Pearson correlation matrix for the main variables used in the models. Our measures of future stock price crash risk are significantly correlated with each other at the 1 percent significance level.

Table 2. Summary statistics and pairwise correlations.

Panel A: Summary statistics
Variable	N			Mean			SD			p25			Median			p75
NCSKEW	117,017			−0.174			0.815			−0.615			−0.159			0.266
DUVOL	117,017			−0.110			0.536			−0.461			−0.115			0.227
CRASH	117,017			−0.145			0.671			−1			0			0
Gini	117,017			37.37			4.679			33.70			38.20			40.80
Return	117,017			−0.134			0.132			−0.175			−0.0945			−0.0498
SD	117,017			0.0508			0.0258			0.0324			0.0448			0.0627
Kurtosis	117,017			1.831			2.721			0.155			0.968			2.437
M/B	117,017			2.385			2.500			0.885			1.600			2.892
Size	117,017			19.37			2.165			17.74			19.49			20.87
Leverage	117,017			0.244			0.177			0.0954			0.223			0.364
ROE	117,017			0.0486			0.237			0.0137			0.0756			0.145
Dturn	117,017			0			0.0163			−0.0013			0			0.0009
GDPgrowth	117,017			3.869			3.096			2.005			3.202			6.690
GDPcapita	117,017			259.9			212.6			65.92			238.8			436.0
CR	117,017			2.017			0.919			1			2			2.840
Governance	117,017			338.5			114.3			215.9			375.5			447.1
GDS	117,017			28.92			10.97			18.62			26.72			34.84
Market	117,017			81.09			39.78			46.70			74.02			108.7
DCP	117,017			78.71			55.03			31.63			70.17			114.5
Variable	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)	(18)	(19)
(1) NCSKEW	1.000
(2) DUVOL	0.891	1.000
(3) CRASH	0.759	0.632	1.000
(4) Gini	0.037	0.036	0.030	1.000
(5) Return	0.053	0.047	0.061	0.024	1.000
(6) SD	−0.055	−0.049	−0.061	−0.010	−0.854	1.000
(7) Kurtosis	−0.044	−0.036	−0.031	0.011	−0.184	0.176	1.000
(8) M/B	0.060	0.060	0.047	0.088	0.004	0.041	−0.011	1.000
(9) Size	0.162	0.151	0.144	0.198	0.364	−0.390	−0.106	0.268	1.000
(10) Leverage	−0.002	−0.001	−0.003	−0.002	0.002	−0.004	−0.004	−0.002	−0.003	1.000
(11) ROE	0.024	0.014	0.029	0.031	0.220	−0.211	−0.039	0.029	0.181	−0.008	1.000
(12) Dturn	0.001	0.005	0.000	−0.002	−0.041	0.048	0.014	0.017	0.000	0.000	−0.010	1.000
(13) GDPgrowth	−0.035	−0.038	−0.031	0.196	0.037	0.011	−0.026	0.072	−0.027	−0.010	0.046	−0.006	1.000
(14) GDPcapita	0.149	0.151	0.118	−0.151	0.009	−0.071	0.006	0.046	0.257	0.004	−0.034	0.003	−0.534	1.000
(15) CR	−0.088	−0.086	−0.072	−0.291	−0.006	0.009	0.029	−0.067	−0.218	0.006	−0.001	−0.002	0.124	−0.249	1.000
(16) Governance	0.119	0.120	0.096	−0.299	−0.015	−0.056	0.010	−0.015	0.128	0.006	−0.040	0.006	−0.612	0.893	−0.083	1.000
(17) GDS	−0.078	−0.079	−0.062	0.057	0.036	0.022	−0.039	0.056	0.034	−0.007	0.010	−0.011	0.664	−0.531	0.236	−0.657	1.000
(18) Market	0.109	0.110	0.083	0.117	0.008	−0.046	0.022	0.098	0.206	−0.004	−0.001	0.005	−0.247	0.620	−0.132	0.564	−0.364	1.000
(19) DCP	−0.027	−0.031	−0.017	−0.023	−0.027	−0.001	−0.012	−0.056	−0.061	0.006	0.005	−0.009	−0.137	0.060	0.032	0.097	−0.093	−0.012	1.000

Table 2 presents the summary statistics and correlation coefficients of the variables used in the primary analysis. Panel A reports summary statistics and Panel B provides Pearson correlation coefficients.

4. Results

4.1. Effect of Income Inequality on Stock Price Crash Risk

To examine the effect of income inequality on future stock price crash risk, we employ Equation (4) for all three measures of stock price crash risk. As previously discussed, we anticipate a positive and statistically significant relationship between income inequality and stock crash risk. Table 3 presents the results for Equation (4). In Column 1, we report the results for $NCSKEW$, where the coefficient of the $GINI$ index is 0.0035 and significant at the 1 percent level. Moving to Column 2, we analyze the effect of income inequality on $DUVOL$. Here, the coefficient of the $GINI$ index is 0.0029 and also significant at the 1 percent level. Finally, in Column 3, we present the results for $CRASH$. The coefficient of the $GINI$ index is 0.0017 and significant at the 1 percent level. Interpreting the results, a one standard deviation increase in the $GINI$ index leads to a respective increase in stock price crash risk measured by $NCSKEW$, $DUVOL$, and $CRASH$ of 9.46 percent, 12.54 percent, and 5.74 percent compared to the sample mean.

Table 3. Income inequality and stock price crash risk.

VARIABLES	${\mathrm{NCSKEW}}_{\mathit{t}+1}$	${\mathrm{DUVOL}}_{\mathit{t}+1}$	${\mathrm{CRASH}}_{\mathit{t}+1}$
${\mathrm{Gini}}_t$	0.0035 ***	0.0029 ***	0.0017 ***
	(0.0006)	(0.0004)	(0.0005)
${\mathrm{Ncskew}}_t$	0.0595 ***	0.0390 ***	0.0349 ***
	(0.0035)	(0.0021)	(0.0026)
${\mathrm{Return}}_t$	0.296 ***	0.189 ***	0.246 ***
	(0.0780)	(0.0340)	(0.0590)
${\mathrm{SD}}_t$	1.977 ***	1.232 ***	1.249 ***
	(0.377)	(0.175)	(0.286)
${\mathrm{Kurtosis}}_t$	−0.008 ***	−0.0045 ***	−0.0039 ***
	(0.0011)	(0.0006)	(0.0008)
${\mathrm{M/B}}_t$	0.0067 ***	0.0048 ***	0.0037 ***
	(0.0011)	(0.0007)	(0.0009)
${\mathrm{Size}}_t$	0.0471 ***	0.0279 ***	0.0342 ***
	(0.0015)	(0.0010)	(0.0012)
${\mathrm{Leverage}}_t$	−0.0163	−0.0043	−0.0148
	(0.0144)	(0.0095)	(0.0118)
${\mathrm{ROE}}_t$	0.0133	−0.0164 **	0.0282 ***
	(0.0118)	(0.0074)	(0.0094)
${\mathrm{Dturn}}_t$	−0.0195	0.0565	−0.0081
	(0.144)	(0.0925)	(0.120)
${\mathrm{GDPgrowth}}_t$	0.0021	−0.0010	0.0024 **
	(0.0014)	(0.0009)	(0.0011)
GDP ${\mathrm{capita}}_t$	0.0003 ***	0.0002 ***	0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{CR}}_t$	−0.0195 ***	−0.0122 ***	−0.0132 ***
	(0.0033)	(0.0022)	(0.0026)
${\mathrm{Governance}}_t$	0.0001 *	0.0001	0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{GDS}}_t$	−0.0021 ***	−0.0009 ***	−0.0014 ***
	(0.0004)	(0.0002)	(0.0003)
${\mathrm{Market}}_t$	−0.0001 *	−0.0001 **	−0.0001 **
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{DCP}}_t$	−0.0003 ***	−0.0002 ***	−0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
Industry FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Observations	117,017	117,017	117,017
R-squared	0.062	0.064	0.038

Notes: Table 3 reports the results for the impact of income inequality on stock price crash risk. Robust standard errors corrected for firm clustering are reported in parentheses. Intercepts are included but suppressed for brevity. Year- and industry-fixed effects are included. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. All variables are defined in the Appendix A.

4.2. Moderating Effect of Country-Level Transparency

To assess the validity of Hypothesis 2, which posits that country-level information and accountability transparency can negatively moderate the positive relationship between income inequality and stock price crash risk, we adopt the methodology outlined by Williams (2015) to measure these variables. For information transparency, we utilize 13 distinct indicators that capture the quantity of information available, the processes through which information is generated, and the infrastructure supporting its dissemination. Similarly, our measure of accountability transparency is constructed from 16 indicators reflecting dimensions such as media freedom, fiscal transparency, and political constraints. By employing these indices, we aim to elucidate the moderating effects of transparency on the relationship between income inequality and stock price crash risk, as hypothesized.

Equation (5) is used to check the moderating effect of different types of transparency.

$$\begin{array}{cc}\hfill {\mathrm{Crash}}_{i,j,t}=& {\alpha }_0+{\alpha }_1{\mathrm{Gini}}_{j,t-1}+{\alpha }_2{\mathrm{Transparency}}_{j,t-1}+{\alpha }_3\left({\mathrm{Gini}}_{j,t-1}\times {\mathrm{Transparency}}_{j,t-1}\right)+{\mathrm{Controls}}_{i,j,t-1}\hfill \\ \hfill & +\mathrm{Year}\mathrm{FE}+\mathrm{Industry}\mathrm{FE}.\hfill \end{array}$$

(5)

Panel A of Table 4 presents the findings regarding the moderating effect of information transparency on the relationship between income inequality and stock price crash risk. As hypothesized, we anticipated a negative coefficient for the interaction term (${\alpha }_3$), indicating that the relationship between income inequality and stock price crash risk weakens as information transparency increases. Consistent with our hypothesis, the coefficient of income inequality (${\alpha }_1$) is positive and statistically significant at the 1 percent level across all three measures of stock price crash risk. However, the coefficient of the interaction between income inequality and information transparency (${\alpha }_3$) is negative and statistically significant at the 1 percent level for all three measures as well. A one standard deviation increase in transparency reduces the negative effect of income inequality on stock price crash risk by 13.7%. These results confirm that the negative association between income inequality and stock price crash risk is indeed attenuated in countries with higher levels of information transparency.

Table 4. Moderating role of transparency on the relationship between income inequality and stock price crash risk.

	Panel A: Information Transparency			Panel B: Accountability Transparency			Panel C: Total Transparency
Variable	NCSKEW	DUVOL	CRASH	NCSKEW	DUVOL	CRASH	NCSKEW	DUVOL	CRASH
${\mathrm{Gini}}_t$	0.0772 ***	0.0547 ***	0.0486 ***	0.0158 ***	0.0098 **	0.0125 **	0.0352 ***	0.0241 ***	0.0243 ***
	(0.0117)	(0.0078)	(0.0090)	(0.0061)	(0.0041)	(0.0048)	(0.0086)	(0.0059)	(0.0068)
${\mathrm{Transparency}}_t$	0.0367 ***	0.0285 ***	0.0206 ***	−0.0018	−0.0010	0.0008	0.0048	0.0045	0.0034
	(0.0060)	(0.0040)	(0.0047)	(0.0040)	(0.0027)	(0.0032)	(0.0052)	(0.0035)	(0.0041)
${\mathrm{Gini}}_t\times {\mathrm{Transparency}}_t$	−0.0010 ***	−0.0007 ***	−0.0006 ***	−0.0002 **	−0.0001 **	−0.0002 **	−0.0004 ***	−0.0003 ***	−0.0003 ***
	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(<0.0001)	(<0.0001)	(0.0001)	(<0.0001)	(0.0001)
${\mathrm{Ncskew}}_t$	0.0528 ***	0.0345 ***	0.0338 ***	0.0500 ***	0.0326 ***	0.0322 ***	0.0513 ***	0.0335 ***	0.0328 ***
	(0.0061)	(0.0039)	(0.0046)	(0.0061)	(0.0039)	(0.0047)	(0.0061)	(0.0039)	(0.0046)
${\mathrm{Return}}_t$	0.277 ***	0.185 **	0.179 **	0.347 ***	0.236 ***	0.215 ***	0.339 ***	0.229 ***	0.216 ***
	(0.104)	(0.0722)	(0.0787)	(0.104)	(0.0720)	(0.0784)	(0.105)	(0.0723)	(0.0787)
${\mathrm{SD}}_t$	0.568	0.221	0.199	1.038 *	0.551	0.451	0.967 *	0.491	0.448
	(0.565)	(0.391)	(0.424)	(0.564)	(0.390)	(0.423)	(0.566)	(0.392)	(0.424)
${\mathrm{Kurtosis}}_t$	−0.0051 ***	−0.0029 **	−0.0027 *	−0.0054 ***	−0.0030 **	−0.0030 **	−0.0049 **	−0.0026 **	−0.0027 *
	(0.0019)	(0.0012)	(0.0014)	(0.0019)	(0.0012)	(0.0014)	(0.0019)	(0.0012)	(0.0014)
${\mathrm{M/B}}_t$	0.0028	0.0021	0.0011	0.0032	0.0025 *	0.0012	0.0032	0.0025 *	0.0012
	(0.0022)	(0.0014)	(0.0017)	(0.0022)	(0.0014)	(0.0017)	(0.0022)	(0.0014)	(0.0017)
${\mathrm{Size}}_t$	0.0488 ***	0.0281 ***	0.0372 ***	0.0489 ***	0.0284 ***	0.0371 ***	0.0491 ***	0.0284 ***	0.0372 ***
	(0.0025)	(0.0017)	(0.0021)	(0.0025)	(0.0017)	(0.0021)	(0.0025)	(0.0017)	(0.0021)
${\mathrm{Leverage}}_t$	0.0437 *	0.0484 ***	−0.0075	0.0414 *	0.0453 ***	−0.0067	0.0396	0.0444 ***	−0.0080
	(0.0246)	(0.0166)	(0.0206)	(0.0245)	(0.0165)	(0.0205)	(0.0245)	(0.0165)	(0.0205)
${\mathrm{ROE}}_t$	−0.0104	−0.0277 **	0.0123	0.0002	−0.0216 *	0.0201	−0.0059	−0.0258 **	0.0168
	(0.0189)	(0.0125)	(0.0155)	(0.0189)	(0.0125)	(0.0155)	(0.0189)	(0.0125)	(0.0155)
${\mathrm{Dturn}}_t$	0.653 ***	0.598 ***	0.294	0.743 ***	0.655 ***	0.349 *	0.720 ***	0.637 ***	0.345 *
	(0.227)	(0.149)	(0.200)	(0.229)	(0.150)	(0.201)	(0.229)	(0.150)	(0.201)
${\mathrm{GDPgrowth}}_t$	0.0239 ***	0.0163 ***	0.0120 ***	0.0198 ***	0.0128 ***	0.0106 ***	0.0193 ***	0.0127 ***	0.00991 ***
	(0.0019)	(0.0012)	(0.0016)	(0.0019)	(0.0012)	(0.0015)	(0.0019)	(0.0012)	(0.0015)
GDP ${\mathrm{capita}}_t$	0.0006 ***	0.0004 ***	0.0004 ***	0.0004 ***	0.0003 ***	0.0002 ***	0.0004 ***	0.0003 ***	0.0003 ***
	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{CR}}_t$	−0.0428 ***	−0.0271 ***	−0.0278 ***	−0.0482 ***	−0.0307 ***	−0.0308 ***	−0.0452 ***	−0.0285 ***	−0.0294 ***
	(0.0052)	(0.0035)	(0.0040)	(0.0052)	(0.0036)	(0.0041)	(0.0052)	(0.0036)	(0.0041)
${\mathrm{Governance}}_t$	−0.0002	−0.0003 ***	−0.0001	0.0011 ***	0.0006 ***	0.0006 ***	0.0010 ***	0.0006 ***	0.0007 ***
	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)
${\mathrm{GDS}}_t$	−0.0083 ***	−0.0060 ***	−0.0049 ***	−0.0091 ***	−0.0065 ***	−0.0054 ***	−0.0092 ***	−0.0067 ***	−0.0056 ***
	(0.0009)	(0.0006)	(0.0007)	(0.0009)	(0.0006)	(0.0007)	(0.0009)	(0.0006)	(0.0007)
${\mathrm{Market}}_t$	−0.0004 ***	−0.0001	−0.0003 ***	−0.0003 **	−0.0001	−0.0002 **	−0.0004 ***	−0.0002 **	−0.0003 **
	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)
${\mathrm{DCP}}_t$	−0.0005 ***	−0.0004 ***	−0.0002 ***	−0.0005 ***	−0.0004 ***	−0.0003 ***	−0.0005 ***	−0.0004 ***	−0.0003 ***
	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)
Industry FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Observations	35,500	35,500	35,500	35,500	35,500	35,500	35,500	35,500	35,500
R-squared	0.069	0.066	0.047	0.073	0.070	0.048	0.071	0.068	0.048

Notes: This table reports the moderating role of transparency (information, accountability, and total transparency) on the relationship between income inequality and stock price crash risk. Robust standard errors corrected for firm clustering are reported in parentheses. Intercepts are included but suppressed for brevity. Year- and industry-fixed effects are included. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. All variables are defined in the Appendix A.

Panel B of Table 4 displays the outcomes concerning the moderating influence of accountability transparency on the association between income inequality and stock price crash risk. Our hypothesis suggested that this relationship would be negatively moderated by accountability transparency. As expected, the coefficient of the Gini index (${\alpha }_1$) is positive and statistically significant at the 1 percent level across all three measures of stock price crash risk, indicating a positive relationship between income inequality and stock price crash risk. Crucially, the coefficient of the interaction between income inequality and accountability transparency (${\alpha }_3$) is negative and statistically significant at the 1 percent level for all three measures as well. A one standard deviation increase in transparency reduces the negative effect of income inequality on stock price crash risk by 14.3%. This result provides support for our hypothesis that accountability transparency negatively moderates the positive relationship between income inequality and stock price crash risk. Thus, it confirms the validity of hypothesis 2 regarding accountability transparency.

Panel C of Table 4 presents the results regarding the moderating effect of a combined transparency index, which incorporates both information and accountability transparency. Our hypothesis posited that this transparency index would negatively moderate the positive relationship between income inequality and stock price crash risk. Consistent with our expectations, the coefficient of the Gini index (${\alpha }_1$) is positive and statistically significant at the 1 percent level across all three measures of stock price crash risk, indicating a positive relationship between income inequality and stock price crash risk. Furthermore, the coefficient of the interaction between income inequality and the transparency index (${\alpha }_3$) is negative and statistically significant at the 1 percent level for all three measures. A one standard deviation increase in transparency reduces the negative effect of income inequality on stock price crash risk by 15.1%. This finding supports our hypothesis that total transparency negatively moderates the positive association between income inequality and stock price crash risk. Therefore, it confirms the validity of our hypothesis regarding the moderating effect of total transparency.

4.3. Moderating Role of Firm-Level Monitoring

Hypothesis 3a posits that institutional ownership, serving as a proxy for effective monitoring, can mitigate the positive association between income inequality and stock price crash risk. To test this hypothesis, we employ Equation (6) and anticipate a negative coefficient for the interaction between income inequality and institutional ownership ($a_3$).

$$\begin{array}{cc}\hfill {\mathrm{Crash}}_{i,j,T}=& {\alpha }_0+{\alpha }_{1}Gin{i}_{j,T-1}+{\alpha }_{2}Institutio{n}_{i,j,T-1}+{\alpha }_3\left(Gin{i}_{j,T-1}\times Institutio{n}_{i,j,T-1}\right)+Control{s}_{i,j,T-1}\hfill \\ \hfill & +\mathrm{Year}\mathrm{FE}+\mathrm{Industry}\mathrm{FE}.\hfill \end{array}$$

(6)

where Institution represents the percentage of ownership held by institutional and strategic investors.

Table 5 presents the results regarding the moderating effect of corporate governance, as measured by institutional ownership, on the association between income inequality and stock price crash risk. Consistent with expectations, the coefficient of the Gini index is positive and statistically significant at the 1 percent level for NCSKEW and DUVOL, and at the 5 percent level for CRASH. Furthermore, the coefficient of the interaction between income inequality and institutional ownership is negative and statistically significant at the 5 percent level for DUVOL, at the 10 percent level for NCSKEW, and narrowly misses significance at the 10 percent level for CRASH. These results confirm that institutional ownership negatively moderates the positive association between income inequality and stock price crash risk.

Table 5. Income inequality and stock price crash risk: moderating role of institutional investors.

VARIABLES	${\mathrm{NCSKEW}}_{\mathit{t}+1}$	${\mathrm{DUVOL}}_{\mathit{t}+1}$	${\mathrm{CRASH}}_{\mathit{t}+1}$
${\mathrm{Gini}}_t$	0.0046 ***	0.0041 ***	0.0024 **
	(0.0012)	(0.0008)	(0.0009)
${\mathrm{Institution}}_t$	0.0575	0.0553	0.0355
	(0.0887)	(0.0594)	(0.0654)
${\mathrm{Gini}}_t\times {\mathrm{Institution}}_t$	−0.0046 *	−0.0038 **	−0.0023
	(0.0023)	(0.0015)	(0.0017)
${\mathrm{Ncskew}}_t$	0.0612 ***	0.0402 ***	0.0330 ***
	(0.0050)	(0.0030)	(0.0029)
${\mathrm{Return}}_t$	0.199	0.154 **	0.161 *
	(0.153)	(0.0734)	(0.0959)
${\mathrm{SD}}_t$	1.312 *	0.982 ***	0.587
	(0.765)	(0.372)	(0.477)
${\mathrm{Kurtosis}}_t$	−0.0077 ***	−0.0036 ***	−0.0029 ***
	(0.0018)	(0.0011)	(0.0008)
${\mathrm{M/B}}_t$	0.0042 ***	0.0031 ***	0.0020 ***
	(0.0008)	(0.0005)	(0.0006)
${\mathrm{Size}}_t$	0.0448 ***	0.0255 ***	0.0319 ***
	(0.0017)	(0.0011)	(0.0013)
${\mathrm{Leverage}}_t$	−0.0005 ***	−0.0002 *	−0.0008 **
	(0.0002)	(0.0001)	(0.0004)
${\mathrm{ROE}}_t$	0.0206 ***	0.00607	0.0156 ***
	(0.0076)	(0.0048)	(0.0059)
${\mathrm{Dturn}}_t$	0.125 ***	0.112 ***	0.121 ***
	(0.0342)	(0.0279)	(0.0340)
${\mathrm{GDPgrowth}}_t$	0.0045 ***	−0.0007	0.0036 ***
	(0.0017)	(0.0010)	(0.0013)
${\mathrm{GDPcapita}}_t$	0.0003 ***	0.0002 ***	0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{CR}}_t$	−0.0140 ***	−0.0075 ***	−0.0094 ***
	(0.0039)	(0.0026)	(0.0028)
${\mathrm{Governance}}_t$	−0.0001	−0.0001	0.00010
	(0.0001)	(<0.0001)	(<0.0001)
${\mathrm{GDS}}_t$	−0.0033 ***	−0.0015 ***	−0.0020 ***
	(0.0005)	(0.0003)	(0.0003)
${\mathrm{Market}}_t$	−0.0003 ***	−0.0002 ***	−0.0002 ***
	(0.0001)	(<0.0001)	(<0.0001)
${\mathrm{DCP}}_t$	−0.0001 **	−0.0001 ***	−0.0001
	(<0.0001)	(<0.0001)	(<0.0001)
Industry FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Observations	94,116	94,116	94,116
R-squared	0.059	0.061	0.038

Notes: Table 5 reports the results for the moderating role of institutional investors on the relationship between income inequality and stock price crash risk. Robust standard errors corrected for firm clustering are reported in parentheses. Intercepts are included but suppressed for brevity. Year- and industry-fixed effects are included. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. All variables are defined in the Appendix A.

To further explore the moderating role of monitoring, we consider the argument by Jensen (1986) regarding the direct impact of cash holdings on agency problems. We propose that the effect of income inequality on stock price crash risk is more pronounced for firms with higher excess cash. To test Hypothesis 3b, we follow Opler et al. (1999) and estimate excess cash by regressing cash holdings on firm characteristics such as size, market-to-book ratio, and leverage, and computing the residuals from this regression. We then employ Equation (7) and expect a positive and statistically significant coefficient for the interaction between income inequality and excess cash.

$$\begin{array}{cc}\hfill {\mathrm{Crash}}_{i,j,t}=& {\alpha }_0+{\alpha }_1{\mathrm{Gini}}_{j,t-1}+{\alpha }_2{\mathrm{Cash}}_{i,j,t-1}+{\alpha }_3\left({\mathrm{Gini}}_{j,t-1}\times {\mathrm{Cash}}_{i,j,t-1}\right)+{\mathrm{Controls}}_{i,j,t-1}\hfill \\ \hfill & +\mathrm{Year}\mathrm{FE}+\mathrm{Industry}\mathrm{FE}.\hfill \end{array}$$

(7)

where Cash denotes the excess cash in year t. Table 6 presents the moderating role of excess cash on the association between income inequality and stock price crash risk. The coefficient of the Gini index is positive and statistically significant at the 1 percent level for both DUVOL and CRASH. Additionally, the coefficient of the moderating variable (${\alpha }_3$) is positive and statistically significant at the 1 percent level for all three measures of crash risk.

Table 6. Income inequality and stock price crash risk: moderating role of monitoring effectiveness.

VARIABLES	${\mathrm{NCSKEW}}_{\mathit{t}+1}$	${\mathrm{DUVOL}}_{\mathit{t}+1}$	${\mathrm{CRASH}}_{\mathit{t}+1}$
${\mathrm{Gini}}_t$	0.0039 ***	0.0032 ***	0.0020 ***
	(0.0006)	(0.0004)	(0.0005)
${\mathrm{Cash}}_t$	0.802 ***	0.480 ***	0.437 ***
	(0.158)	(0.101)	(0.128)
${\mathrm{Gini}}_t\times {\mathrm{Cash}}_t$	0.0214 ***	0.0131 ***	0.0116 ***
	(0.0041)	(0.0026)	(0.0033)
${\mathrm{Ncskew}}_t$	0.0610 ***	0.0395 ***	0.0358 ***
	(0.0036)	(0.0022)	(0.0027)
${\mathrm{Return}}_t$	0.295 ***	0.186 ***	0.246 ***
	(0.0807)	(0.0346)	(0.0612)
${\mathrm{SD}}_t$	1.955 ***	1.202 ***	1.261 ***
	(0.387)	(0.177)	(0.294)
${\mathrm{Kurtosis}}_t$	−0.0079 ***	−0.0046 ***	−0.0039 ***
	(0.0011)	(0.0006)	(0.0008)
${\mathrm{M/B}}_t$	0.0069 ***	0.0050 ***	0.0036 ***
	(0.0012)	(0.0007)	(0.0009)
${\mathrm{Size}}_t$	0.0469 ***	0.0278 ***	0.0341 ***
	(0.0015)	(0.0010)	(0.0012)
${\mathrm{Leverage}}_t$	−0.0204	−0.0057	−0.0167
	(0.0155)	(0.0103)	(0.0128)
${\mathrm{ROE}}_t$	0.0132	−0.0174 **	0.0304 ***
	(0.0121)	(0.0077)	(0.0097)
${\mathrm{Dturn}}_t$	0.0078	0.0759	−0.0018
	(0.146)	(0.0939)	(0.123)
${\mathrm{GDPgrowth}}_t$	0.0018	−0.0012	0.0023 *
	(0.0015)	(0.0009)	(0.0012)
${\mathrm{GDPcapita}}_t$	0.0003 ***	0.0002 ***	0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{CR}}_t$	−0.0181 ***	−0.0110 ***	−0.0125 ***
	(0.0033)	(0.0022)	(0.0026)
${\mathrm{Governance}}_t$	0.0001 **	0.0001 ***	0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{GDS}}_t$	−0.0021 ***	−0.0010 ***	−0.0015 ***
	(0.0004)	(0.0002)	(0.0003)
${\mathrm{Market}}_t$	−0.0001 **	−0.0001 **	−0.0001 **
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{DCP}}_t$	−0.0003 ***	−0.0002 ***	−0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
Industry FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Observations	113,045	113,045	113,045
R-squared	0.063	0.065	0.039

Notes: Table 6 reports the results for the moderating role of monitoring effectiveness on the relationship between income inequality and stock price crash risk. Excess cash is used as the proxy for monitoring effectiveness. Robust standard errors corrected for firm clustering are reported in parentheses. Intercepts are included but suppressed for brevity. Year- and industry-fixed effects are included. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. All variables are defined in the Appendix A.

4.4. Endogeneity

Our findings could be susceptible to endogeneity issues in regression analysis if stock price crash risk and income inequality are determined endogenously. To tackle this concern, we employ two approaches: a two-stage instrumental variable method and a propensity score matching technique. These methods offer robust strategies for handling endogeneity concerns and enable us to better understand the causal relationship between income inequality and stock price crash risk.

4.4.1. Two-Stage Least Square (IV_2SLS)

To mitigate potential endogeneity concerns, we employ a two-stage instrumental variable approach. Following Sylwester (2002), we select EDUCATION (the ratio of government educational expenditure divided by GDP) as our instrumental variable.

Following Sylwester (2002), we select EDUCATION (the ratio of government educational expenditure divided by GDP) as our instrumental variable for income inequality. The validity of this instrument rests on two conditions. First, education inequality is strongly related to income inequality, as disparities in educational attainment directly translate into differences in earning potential and income distribution. Consistent with this argument, the first-stage results show a strong and statistically significant relationship between education inequality and the Gini index, indicating that the relevance condition is satisfied. Second, education inequality is plausibly exogenous to firm-level stock price crash risk, as it reflects long-run structural characteristics of a country’s education system rather than short-term corporate disclosure behavior or stock market dynamics. While education inequality may influence economic development broadly, its effect on crash risk is expected to operate primarily through income inequality rather than directly through firm-level information hoarding.

In the first stage, we observe a negative and statistically significant association between government expenditure on education and future income inequality, as presented in the first column of Panel A in Table 7. Subsequently, in the second stage, we utilize the fitted values from the first stage as our explanatory variable. As demonstrated in columns 2 through 4 of Panel A in Table 7, similar to income inequality, this explanatory variable exhibits a positive and statistically significant association with future stock price crash risk across all three measures at the 1 percent significance level.

Table 7. Endogeneity concern.

Panel A: Two-Stage Least Squares (2SLS)
VARIABLES	${\mathbf{Gini}}_{\mathbf{t}}$	${\mathbf{NCSKEW}}_{\mathbf{t}+\mathbf{1}}$		${\mathbf{DUVOL}}_{\mathbf{t}+\mathbf{1}}$	${\mathbf{CRASH}}_{\mathbf{t}+\mathbf{1}}$
${\mathrm{Education}}_{t-1}$	−0.529 ***
	(0.0119)
Fitted ${\mathrm{Gini}}_t$		0.0078 ***		0.0062 ***	0.0046 ***
		(0.0010)		(0.0007)	(0.0009)
${\mathrm{Ncskew}}_t$	−0.0121 ***	0.0478 ***		0.0314 ***	0.0269 ***
	(0.0043)	(0.0047)		(0.0028)	(0.0034)
${\mathrm{Return}}_t$	−0.614 ***	0.272 **		0.194 ***	0.258 **
	(0.0641)	(0.139)		(0.0568)	(0.112)
${\mathrm{SD}}_t$	−4.118 ***	2.165 ***		1.445 ***	1.456 ***
	(0.337)	(0.646)		(0.280)	(0.520)
${\mathrm{Kurtosis}}_t$	0.0014	−0.0078 ***		−0.0043 ***	−0.0039 ***
	(0.0013)	(0.0014)		(0.0008)	(0.0010)
${\mathrm{M/B}}_t$	−0.0010	0.0114 ***		0.0078 ***	0.0068 ***
	(0.0016)	(0.0015)		(0.0009)	(0.0012)
${\mathrm{Size}}_t$	−0.0198 ***	0.0518 ***		0.0321 ***	0.0357 ***
	(0.0024)	(0.0020)		(0.0014)	(0.0016)
${\mathrm{Leverage}}_t$	0.0862 ***	−0.0301		−0.0194	−0.0214
	(0.0217)	(0.0195)		(0.0129)	(0.0158)
${\mathrm{ROE}}_t$	0.0258	0.0020		−0.0302 ***	0.0186
	(0.0174)	(0.0165)		(0.0103)	(0.0128)
${\mathrm{Dturn}}_t$	−2.432 ***	−0.805 ***		−0.474 ***	−0.302
	(0.258)	(0.222)		(0.146)	(0.184)
${\mathrm{GDPgrowth}}_t$	0.0511 ***	−0.0048 **		−0.0051 ***	−0.0025
	(0.0025)	(0.0021)		(0.0013)	(0.0017)
${\mathrm{GDPcapita}}_t$	−0.0312 ***	−0.0001		−0.0001	−0.0001 *
	(0.0002)	(<0.0001)		(<0.0001)	(<0.0001)
${\mathrm{CR}}_t$	1.676 ***	−0.0031		−0.0031	−0.0013
	(0.0448)	(0.0044)		(0.0029)	(0.0035)
${\mathrm{Governance}}_t$	−0.0149 ***	0.0009 ***		0.0006 ***	0.0007 ***
	(0.0003)	(0.0001)		(<0.0001)	(<0.0001)
${\mathrm{GDS}}_t$	0.148 ***	0.0003		0.0009 **	0.0004
	(0.0029)	(0.0006)		(0.0004)	(0.0005)
${\mathrm{Market}}_t$	−0.0050 ***	0.0003 ***		0.0002 ***	0.0001 *
	(0.0002)	(0.0001)		(<0.0001)	(<0.0001)
${\mathrm{DCP}}_t$	0.0021 ***	−0.0001		−0.0001	0.0001
	(<0.0001)	(<0.0001)		(<0.0001)	(<0.0001)
Industry FE	Yes	Yes		Yes	Yes
Year FE	Yes	Yes		Yes	Yes
Observations	69,087	69,087		69,087	69,087
R-squared	0.951	0.060		0.062	0.038
Panel B: Propensity score matching
Method	Variable	Treated mean	Control mean	Mean diff.	T-statistic
Nearest neighbor matching	NCSKEW	−0.0819	−0.1086	0.0266	2.99 ***
	DUVOL	−0.0549	−0.0683	0.0134	2.34 **
	CRASH	−0.0913	−0.1008	0.0095	1.42
Radius matching	NCSKEW	−0.0819	−0.1302	0.0483	7.43 ***
	DUVOL	−0.0549	−0.0837	0.0288	6.88 ***
	CRASH	−0.0913	−0.1150	0.0237	4.84 ***
Kernel matching	NCSKEW	−0.0819	−0.1170	0.0350	5.27 ***
	DUVOL	−0.0549	−0.0758	0.0209	4.89 ***
	CRASH	−0.0913	−0.1059	0.0146	2.92 ***

Notes: Table 7 presents the results for the endogeneity concern. Panel A reports the Two-Stage Least Squares (2SLS) regression results for the impact of income inequality on stock price crash risk. The instrumental variable is educational expenditure (percentage of GDP). Panel B reports the results for the propensity score matching method using three approaches: nearest neighbor, radius, and kernel matching. Year- and industry-fixed effects are included. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. All variables are defined in the Appendix A.

4.4.2. Propensity Score Matching

To address potential endogeneity concerns through a different approach, we employ the propensity score matching (PSM) procedure. This involves matching observations of firms based on their probability of experiencing an increase in their Gini index. Specifically, we designate firms located in countries with a Gini index above the 75th percentile of the cross-country Gini index as the treated group, while other firms constitute the control group. To implement PSM, we construct a dummy variable labeled HIGH, which equals 1 if the firm’s country has a Gini index above the 75th percentile and 0 otherwise. Subsequently, we estimate a probit model where we regress the HIGH dummy on all control variables. The estimated scores from this model are then used to match each observation with a HIGH dummy value of 1 to an observation with a HIGH dummy value of 0. We utilize three different matching techniques for this purpose: nearest neighbor matching, radius matching, and kernel density match models. These techniques help ensure a robust matching process, allowing us to effectively compare firms with similar propensity scores across the treatment and control groups.

As depicted in Panel B of Table 7, the results for the propensity score matching (PSM) technique are presented, with the first three rows corresponding to the nearest neighbor matching technique. For NCSKEW and DUVOL, the mean difference between the treated and control groups is positive and statistically significant at the 1 percent level. However, for CRASH, while the mean difference is positive, it falls slightly short of statistical significance at the 10 percent level. Moving to the second set of three rows, which represent the results for radius matching, we observe that the mean differences are positive and statistically significant at the 1 percent level across all three crash measures. Lastly, the last three rows pertain to the results obtained from kernel density matching. In this case, the mean differences for all three crash measures are positive and statistically significant, indicating a consistent pattern across the different matching techniques.

4.5. Robustness Analysis

4.5.1. Alternative Measure of Income Inequality

Some scholars, such as Atkinson (1970), have raised concerns about the limitations of the Gini index as a sole measure of income inequality. They argue that relying solely on the Gini index may overlook variations in income distribution among countries with similar Gini index values. To address this concern and enhance the robustness of our analysis, we incorporate an alternative measure of income inequality proposed by Sitthiyot and Holasut (2020). This alternative measure, depicted in Equation (8), combines the Gini index with the ratio of income held by the top 10% to that held by the bottom 10%. By integrating these additional factors, this alternative measure provides a more nuanced understanding of income distribution dynamics, offering insights beyond those captured by the Gini index alone.

$$\mathrm{S}\&\mathrm{H}\mathrm{Index}={\displaystyle \frac{\sqrt{{\left(\mathrm{Gini}\right)}^2+{\left(T_{10}/B_{10}\right)}^2}}{100}},$$

(8)

where $T_{10}$ is income share held by the top 10% and $B_{10}$ is income share held by the bottom 10%. This alternative index takes the values between 0.01 and ∞. When everyone has the same share of income, it is equal to 0.01 (Gini = 0 and $(T_{10}/B_{10})=1$).

Table 8 presents the outcomes of examining the impact of income inequality measured by the S&H index on subsequent stock price crash risk. The findings reveal a positive and statistically significant association between income inequality, as gauged by the S&H index, and the likelihood of future stock price crashes.

Table 8. Alternative measure of income inequality.

VARIABLES	${\mathrm{NCSKEW}}_{\mathit{t}+1}$	${\mathrm{DUVOL}}_{\mathit{t}+1}$	${\mathrm{CRASH}}_{\mathit{t}+1}$
S&H ${\mathrm{index}}_t$	0.233 ***	0.195 ***	0.127 ***
	(0.0461)	(0.0311)	(0.0377)
${\mathrm{Ncskew}}_t$	0.0595 ***	0.0390 ***	0.0349 ***
	(0.0035)	(0.0021)	(0.0026)
${\mathrm{Return}}_t$	0.296 ***	0.189 ***	0.246 ***
	(0.0780)	(0.0340)	(0.0590)
${\mathrm{SD}}_t$	1.973 ***	1.229 ***	1.247 ***
	(0.377)	(0.175)	(0.286)
${\mathrm{Kurtosis}}_t$	−0.0079 ***	−0.0045 ***	−0.0039 ***
	(0.0011)	(0.0006)	(0.0008)
${\mathrm{M/B}}_t$	0.00681 ***	0.0049 ***	0.0037 ***
	(0.0011)	(0.0007)	(0.0009)
${\mathrm{Size}}_t$	0.0472 ***	0.0280 ***	0.0342 ***
	(0.0015)	(0.0010)	(0.0012)
${\mathrm{Leverage}}_t$	−0.0165	−0.0044	−0.0149
	(0.0144)	(0.0095)	(0.0118)
${\mathrm{ROE}}_t$	0.0129	−0.0168 **	0.0281 ***
	(0.0118)	(0.0074)	(0.0094)
${\mathrm{Dturn}}_t$	−0.0212	0.0551	−0.00877
	(0.144)	(0.0925)	(0.120)
${\mathrm{GDPgrowth}}_t$	0.0023	−0.0008	0.0026 **
	(0.0014)	(0.0009)	(0.0011)
${\mathrm{GDPcapita}}_t$	0.0003 ***	0.0002 ***	0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{CR}}_t$	−0.0195 ***	−0.0122 ***	−0.0131 ***
	(0.0033)	(0.0022)	(0.0026)
${\mathrm{Governance}}_t$	0.0001	0.0001	0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{GDS}}_t$	−0.0021 ***	−0.0009 ***	−0.0014 ***
	(0.0004)	(0.0002)	(0.0003)
${\mathrm{Market}}_t$	−0.0001	−0.0001 *	−0.0001 **
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{DCP}}_t$	−0.0003 ***	−0.0002 ***	−0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
Industry FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Observations	117,017	117,017	117,017
R-squared	0.062	0.064	0.038

Notes: Table 8 reports the results for the impact of income inequality on stock price crash risk. As an alternative measure of income inequality, this table uses the index proposed by Sitthiyot and Holasut (2020). Robust standard errors clustered at the firm level are reported in parentheses. Intercepts are included but suppressed for brevity. Year- and industry-fixed effects are included. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. All variables are defined in the Appendix A.

4.5.2. Wealth Inequality and Stock Price Crash Risk

We posited that income inequality exacerbates the risk of stock price crashes. If income inequality indeed correlates positively with stock price crash risk, other inequality metrics that heighten CEOs’ incentives to withhold adverse news should yield similar associations with stock crash risk. To investigate this, we turn to wealth inequality, represented by Wealth Gini sourced from the World Inequality Database. Table 9 presents the impact of wealth inequality on stock price crash risk. The results demonstrate a robust positive relationship, with the coefficient of Wealth Gini being statistically significant at the 1 percent level across all three measures of crash risk. Specifically, a one standard deviation increase in Wealth Gini corresponds to substantial increments in stock price crash risk: 14.4 percent for NCSKEW, 16.32 percent for DUVOL, and 9.68 percent for CRASH relative to their respective means.

Table 9. Wealth inequality and stock price crash risk.

VARIABLES	${\mathrm{NCSKEW}}_{\mathit{t}+1}$	${\mathrm{DUVOL}}_{\mathit{t}+1}$	${\mathrm{CRASH}}_{\mathit{t}+1}$
Wealth ${\mathrm{Gini}}_t$	0.429 ***	0.307 ***	0.240 ***
	(0.0626)	(0.0411)	(0.0491)
${\mathrm{Ncskew}}_t$	0.0581 ***	0.0379 ***	0.0341 ***
	(0.0036)	(0.0022)	(0.0026)
${\mathrm{Return}}_t$	0.302 ***	0.192 ***	0.249 ***
	(0.0799)	(0.0345)	(0.0603)
${\mathrm{SD}}_t$	2.033 ***	1.258 ***	1.283 ***
	(0.387)	(0.178)	(0.293)
${\mathrm{Kurtosis}}_t$	−0.0078 ***	−0.0044 ***	−0.0037 ***
	(0.0011)	(0.0006)	(0.0008)
${\mathrm{M/B}}_t$	0.0066 ***	0.0047 ***	0.0036 ***
	(0.0011)	(0.0007)	(0.0009)
${\mathrm{Size}}_t$	0.0522 ***	0.0313 ***	0.0372 ***
	(0.0015)	(0.0010)	(0.0012)
${\mathrm{Leverage}}_t$	−0.0216	−0.0087	−0.0154
	(0.0146)	(0.0096)	(0.0120)
${\mathrm{ROE}}_t$	−0.0015	−0.0267 ***	0.0193 **
	(0.0119)	(0.0075)	(0.0095)
${\mathrm{Dturn}}_t$	−0.0987	0.0068	−0.0571
	(0.145)	(0.0933)	(0.121)
${\mathrm{GDPgrowth}}_t$	0.0015	−0.0010	0.0022 *
	(0.0015)	(0.0009)	(0.0012)
${\mathrm{GDPcapita}}_t$	0.0002 ***	0.0002 ***	0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{CR}}_t$	−0.0096 ***	−0.0061 ***	−0.0073 ***
	(0.0034)	(0.0023)	(0.0027)
${\mathrm{Governance}}_t$	0.0001 ***	0.0001	0.0002 ***
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{GDS}}_t$	−0.0008 *	−0.0001	−0.0007 **
	(0.0004)	(0.0002)	(0.0003)
${\mathrm{Market}}_t$	−0.0001	−0.0001	−0.0001
	(<0.0001)	(<0.0001)	(<0.0001)
${\mathrm{DCP}}_t$	−0.0003 ***	−0.0002 ***	−0.0001 ***
	(<0.0001)	(<0.0001)	(<0.0001)
Industry FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Observations	114,513	114,513	114,513
R-squared	0.063	0.065	0.039

Notes: Table 9 reports the results for the impact of wealth inequality on stock price crash risk. Robust standard errors clustered at the firm level are reported in parentheses. Intercepts are included but suppressed for brevity. Year- and industry-fixed effects are included. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. All variables are defined in the Appendix A.

5. Conclusions and Discussion

This study examines the relationship between country-level income inequality and stock price crash risk using a large cross-country sample of publicly listed firms. The empirical results consistently show that higher income inequality is associated with a greater likelihood of stock price crashes, supporting the view that macro-level income distribution shapes firms’ information environments and managerial disclosure behavior. We further find that this relationship is moderated by country-level transparency and firm-level monitoring mechanisms, suggesting that the effect of income inequality on crash risk is conditional on the strength of information asymmetry and monitoring.

Our findings have several important implications for policymakers, regulators, and market participants. From a policy perspective, the results suggest that income inequality is not only a social or economic issue but also a factor with implications for financial market stability. In countries characterized by high income inequality, weaker information environments may allow negative information to accumulate within firms, increasing the likelihood of abrupt price corrections. Strengthening transparency, accountability, and information dissemination mechanisms may therefore help mitigate the adverse market consequences associated with inequality.

From a practical standpoint, our results highlight the importance of monitoring mechanisms in shaping crash risk. Firms operating in high-inequality environments may benefit from stronger internal and external monitoring, particularly through institutional investor oversight and disciplined cash management. For investors, the findings suggest that country-level income inequality and institutional conditions should be considered when assessing downside risk and portfolio exposure, especially in international investment contexts.

Despite the robustness of our findings, several limitations should be acknowledged: First, although we employ instrumental variable and propensity score matching techniques to address endogeneity concerns, establishing strict causality remains challenging in a cross-country setting. Income inequality is deeply intertwined with historical, political, and institutional factors that evolve slowly over time and may not be fully captured by observable controls.

Second, the measurement of income inequality and institutional variables relies on country-level indices, which may mask within-country heterogeneity. Regional disparities, informal institutions, and differences in enforcement quality within countries could influence firms’ information environments and disclosure behavior in ways not captured by aggregate measures.

Third, our sample focuses on publicly listed firms, and the results may not generalize to private firms or to countries with limited stock market development. These factors may affect the external validity of the findings.

Future studies could explore firm-level channels in greater depth, including executive compensation structures, board characteristics, or internal control quality, to better understand how income inequality shapes managerial disclosure incentives. In addition, further research could investigate cross-country heterogeneity more explicitly by examining regional variation within countries or by focusing on specific institutional settings. Finally, extending the analysis to periods of economic stress or financial crises may provide additional insights into how income inequality interacts with market conditions to influence stock price crash risk.