Board diversity and stock price crash risk: exacerbate or mitigate

Stock price crash risk (SPCR) is prevalent in global markets and can inflict severe consequences, especially in emerging economies with fragile institutions. Notable Chinese cases (Huishan Dairy 2017, Changsheng Biologicals 2018, Kangmei Pharmaceutical 2019) illustrate how fraud and poor governance precipitate sharp crashes, eroding investor confidence and destabilizing markets. The literature emphasizes that SPCR stems largely from information environment failures, including managerial manipulation and asymmetry, reflecting breakdowns in corporate governance. Boards are central to governance and monitoring; their composition and diversity may influence monitoring effectiveness and thus SPCR. While prior research has examined specific board attributes (e.g., gender representation, informal hierarchy), there is limited evidence on comprehensive board diversity (BD) and its effect on SPCR, particularly in emerging markets like China. This study addresses that gap by constructing a multidimensional index of BD tailored to the Chinese context and testing its effect on SPCR among A-share firms from 2010 to 2020. The research questions are whether overall BD mitigates or exacerbates SPCR and through which mechanisms. The authors posit that BD can influence SPCR via agency cost reduction, improved investment efficiency, enhanced disclosure, and more diverse executive appointments. They find BD significantly mitigates SPCR, with results robust to endogeneity and validity tests.

The study connects two literatures: board diversity (BD) and corporate risk, and the antecedents of stock price crash risk (SPCR). On BD, most prior work examines single dimensions such as gender, age, education, professional background, and experience. However, focusing on one dimension risks bias because board effectiveness depends on the combined influence of diverse attributes. Much of the evidence also comes from developed markets, with limited attention to emerging markets. Recent efforts propose comprehensive BD indicators spanning educational background, demographics, cultural attitudes, managerial traits, occupational background, and experience, providing a template for broader BD measurement in China. Studies link BD to varied corporate outcomes: some suggest BD increases risk-taking and innovation, while others find BD lowers firm, credit, and financial risk; specific skills and cultural diversity can reduce risk. BD can boost risk information disclosure and firm performance, with some evidence of threshold effects (e.g., gender diversity critical mass). On SPCR, research links it to information disclosure, managerial incentives, market factors, and institutions. A common mechanism is bad-news hoarding by managers due to information asymmetry and self-interest (e.g., to increase option values or protect careers), which raises SPCR. Earnings management reduces information transparency and heightens crash risk, including in China, underscoring the role of governance—and boards in particular—in mitigating SPCR. Prior work on boards and SPCR focuses on structure (size, outsider share) and limited internal characteristics (e.g., gender, experience) with generally negative associations between certain diversities and SPCR. Yet comprehensive BD measures, especially in emerging markets, remain scarce and mechanisms are unclear. The authors develop hypotheses for both mitigating and exacerbating effects of BD on SPCR. Mitigating view: BD enhances monitoring efficiency, lowers agency costs, reduces inefficient investment, improves disclosure, and increases executive team diversity, thereby reducing SPCR (H1a and sub-hypotheses). Exacerbating view: BD can increase decision costs, slow decisions, create subgroups and conflicts, reduce cohesion and monitoring effectiveness, potentially increasing agency costs, inefficient investment, and information opacity and thus raising SPCR (H1b and sub-hypotheses).

Data: The sample includes all Chinese A-share listed companies from 2010 to 2020 sourced from CSMAR, yielding 28,462 firm-year observations initially. Exclusions: financial/insurance firms; ST/*ST/delisted firms; firms listed less than one year; missing financials; fewer than 30 trading weeks in a year. Final sample: 17,412 firm-year observations. To reduce outlier impact, continuous variables are winsorized at the 1st and 99th percentiles. To address endogeneity, BD and controls are measured in 2010–2019 and SPCR in 2011–2020. Variables: Dependent variable (SPCR) measured by NCSKEW and DUVOL. Firm-specific weekly returns are obtained by regressing weekly stock returns on market returns (with leads and lags) to remove market effects; NCSKEW is the negative skewness of firm-specific weekly returns; DUVOL is the log ratio of down-week to up-week volatility. Independent variable (BD): A multidimensional index constructed from director attributes tailored to China, covering five first-tier attributes and eleven second-tier indicators: demographic (age, gender), educational (degree diversity), professional (finance/legal expertise, position diversity, academic background), director experience (foreign experience, Great Chinese Famine early-life experience), managerial traits (independent/outside director diversity, board size diversity). Diversity measures use coefficient of variation (age, outside/independent directors) or Herfindahl-Hirschman-based diversity index (1−Σp_i^2) for categorical attributes; board size uses count of directors. Notably, the famine experience is measured as a birth year at or before 1961. Each second-tier indicator is ranked annually into deciles (1–10), summed to form first-tier indicators, and the total BD index is the sum of first-tier indicators divided by 100 to scale. Controls: RET (mean firm-specific weekly return), SIGMA (std. dev. of firm-specific weekly return), DTURN (change in average monthly turnover), SIZE (log assets), LEV (liabilities/assets), ROA (net income/assets), BM (book-to-market), ACCM (three-year sum of absolute discretionary accruals as transparency proxy). Model: Panel regressions with year and industry fixed effects: SPCR_it+1 = α0 + α1 BD_it + Controls_it + Year FE + Industry FE + ε_it. Mechanism tests use mediation models: (1) M_it = β0 + β1 BD_it + Controls + FE + ε; (2) SPCR_it+1 = γ0 + γ1 BD_it + γ2 M_it + Controls + FE + ε, where mediators are agency costs (selling & admin expenses / operating income), inefficient investment (per Richardson 2006; higher values indicate less efficient investment), risk information disclosure (amount of risk-related disclosure), and executive team diversity (excluding executives who are also directors).

- Descriptive statistics: NCSKEW_t+1 mean -0.273 (SD 0.670), DUVOL_t+1 mean -0.180 (SD 0.461). BD has mean 0.605 and SD 0.107, indicating greater BD variation in China than reported for the U.S. in prior work (SD ~0.060). Sample size 17,412 firm-years. - Univariate analysis: Firms with high BD (above-median) have significantly lower SPCR than low-BD firms. Means: NCSKEW -0.299 vs -0.247; DUVOL -0.198 vs -0.162; both differences significant at 1%. - Baseline regressions (year and industry FE): BD is negatively and significantly associated with SPCR. With controls, coefficients on BD are -0.188 (t = -4.007) for NCSKEW_t+1 and -0.128 (t = -3.953) for DUVOL_t+1. Without full controls, coefficients are -0.254 (t = -5.413) and -0.176 (t = -5.450), respectively. R^2 ranges ~0.044–0.060. - Economic significance: A one standard deviation increase in BD (0.107) reduces NCSKEW by about 7.368% of its mean (-0.188*0.107/-0.273) and reduces DUVOL by about 7.609% of its mean (-0.128*0.107/-0.180). - Mechanism tests (mediation): • Agency costs (M1): BD lowers agency costs (β = -0.039, t = -4.927). Agency costs raise SPCR (coefficients 0.134 and 0.120 on NCSKEW and DUVOL, both significant). Including M1, BD remains significantly negative on SPCR (≈ -0.183, -0.123), supporting H1a-1. • Inefficient investment (M2): BD reduces inefficient investment (β = -0.024, t = -7.520). Inefficient investment increases SPCR (0.350 for NCSKEW; 0.204 for DUVOL; both significant). BD remains negative on SPCR, supporting H1a-2. • Information disclosure (M3): BD increases risk information disclosure (β = 0.074, t = 4.768). Greater disclosure reduces SPCR (-0.096 for NCSKEW; -0.060 for DUVOL; both significant). BD remains negative, supporting H1a-3. • Executive diversity (M4): BD increases executive team diversity (β = 0.249, t = 44.013). Executive diversity is associated with lower SPCR (-0.132 for NCSKEW; -0.076 for DUVOL). BD remains negative, supporting H1a-4. - Robustness and endogeneity: The negative BD–SPCR relationship is robust after addressing endogeneity and validity checks (details in Appendix). Overall, evidence supports that broader board diversity mitigates stock price crash risk in Chinese listed firms via enhanced monitoring and improved information environments.

The study addresses whether comprehensive board diversity (BD) alleviates or exacerbates stock price crash risk (SPCR) in an emerging market context. The empirical evidence demonstrates that BD significantly reduces SPCR, aligning with governance and resource-dependence perspectives that diverse boards enhance monitoring quality, reduce agency problems, improve investment efficiency, and increase transparency. The findings help reconcile mixed prior evidence by moving beyond single-dimension diversity to a multidimensional BD construct tailored to China. Mediation analyses open the black box by identifying four channels—lower agency costs, fewer inefficient investments, more risk disclosure, and more diverse executive appointments—through which BD reduces crash risk. The results are particularly relevant for emerging markets where external governance is weaker and information asymmetry is acute: BD can substitute or complement external mechanisms by strengthening internal oversight. The managerial and policy implications underscore the value of cultivating varied boards while balancing potential coordination costs; regulators can encourage BD to bolster market stability, and investors can use BD as a governance quality signal to predict crash risk.

The paper constructs a multidimensional index of board diversity for Chinese listed firms and shows that greater BD mitigates stock price crash risk. The core results are robust to endogeneity checks and alternative specifications. By expanding beyond single dimensions of diversity, the study contributes new evidence that BD enhances board monitoring and reduces SPCR, and it reveals four operative mechanisms: lowering agency costs, improving investment efficiency, increasing information disclosure, and promoting executive team diversity. These insights inform corporate governance practice, regulatory policy, and investment decision-making in emerging markets. Future research can refine BD measurement (e.g., cognitive and cultural dimensions), examine interactions among mechanisms through chain mediation, and test cross-country heterogeneity in institutional settings.

- Mechanisms are tested individually; interactions among mechanisms are not modeled. Future work could use chain mediation to explore interdependencies. - BD is measured from observable personal attributes due to data constraints; future research could incorporate cognitive and cultural diversity for more precise measurement. - Findings are based on Chinese listed firms; institutional, cultural, and legal differences may yield different BD–SPCR relationships elsewhere, motivating cross-country comparative studies.