Impact of boardroom diversity on corporate financial performance

The paper addresses how boardroom diversity (BD) influences firms’ financial performance (FP), motivated by recent corporate failures and the drive to improve board effectiveness. Drawing on Upper Echelons Theory, Resource-Based Theory, and Managerial Network Theory, the authors argue that demographic and cognitive diversity in boards can enhance decision-making, monitoring, creativity, and competitiveness. However, prior findings on the BD–FP link are mixed and largely focused on developed economies. The study also emphasizes the role of strategic change (SC) as a critical contextual factor potentially moderating the BD–FP relationship, given that boards direct resource reallocation and strategy under changing environments. The research aims to provide evidence from four emerging markets (China, India, Pakistan, Russia), testing whether BD positively affects FP (H1) and whether SC moderates this relationship (H2).

Theoretical and empirical literature on BD–FP is inconclusive, with studies reporting positive, negative, or insignificant associations. Theoretical grounding includes: Upper Echelons Theory (board/top team characteristics shape organizational outcomes), Resource-Based Theory (human capital and diverse skills as valuable resources), and Managerial Network Theory (social/political ties reduce transaction costs and facilitate resource exchange). Empirical studies on gender, nationality, and broader diversity dimensions show mixed results across contexts (e.g., Carter et al., 2010; Green & Homroy, 2018; Terjesen et al., 2016; Rose, 2007; Ciavarella, 2017). Composite indices of diversity have been used to capture multidimensional BD (e.g., Hsu et al., 2019), with some findings that SC can negatively correlate with the BD–FP link. Additional literature highlights potential mechanisms and boundary conditions: social identity theory and demographic differences (gender risk preferences, age and tenure effects on openness to change and strategic capability). Hypotheses: H1—BD positively relates to FP; H2—SC moderates the BD–FP relationship.

Design: Quantitative panel analysis of 240 non-financial firms (60 each from Moscow Exchange, Shanghai Stock Exchange, Bombay Stock Exchange, and Pakistan Stock Exchange), 2008–2020, balanced panel. Sampling prioritized top market-cap firms per exchange (following Shehata et al., 2017) subject to data availability. Data sources: Accounting data from Bloomberg; director characteristics from company websites and DataStream; SC components from company websites and CSMAR (China). Variables winsorized at 5th and 95th percentiles. To mitigate reverse causality and endogeneity, lagged dependent and independent variables were used. Variables: Dependent—FP measured by ROA (main: net income/total assets); ROE used in robustness. Key independent—BD composite index (six dimensions) constructed via Blau index: gender (male/female), age (five brackets), expertise (financial expertise vs. not), nationality (domestic vs. foreign), education (five levels: technical to PhD), tenure (six brackets). Moderator—SC composite index computed as mean of six resource allocation profiles: advertising intensity (AIN), R&D intensity (R&DI), plant & machinery and innovation (PMI), non-production unit ratio (NPU), inventory level (IL), and financial leverage (FL). Controls—firm size (log revenues), organizational growth (sales growth), financial leverage (debt/equity), board size (number of directors), and state ownership (dummy). Econometric strategy: Diagnostic tests included pairwise correlations, panel unit root tests (LLC, IPS, ADF-Fisher, PP-Fisher), multicollinearity (VIF), and heteroskedasticity (White; Cameron & Trivedi IM-test). Baseline regressions estimated fixed effects with robust standard errors. Moderation tested via BD×SC interaction. Endogeneity addressed using two-step system GMM (Arellano–Bond) dynamic panel with lagged ROA, lagged regressors (BD at t−2; SC and controls at t−1), and country-wise estimations. Model fit and instrument validity assessed via Hausman tests (supporting FE), Sargan/Hansen tests (overidentification), and AR(1)/AR(2) serial correlation tests.

- Descriptive and diagnostics: Variables stationary at level and first difference per panel unit root tests. VIF values indicate no multicollinearity. White and Cameron–Trivedi tests indicated heteroskedasticity; FE models used robust errors. - Direct effects (FE models): BD positively and significantly associated with ROA across all four countries. Illustrative coefficients: China ~0.091 (p<0.01), Russia ~0.039 (p<0.01), India ~0.069 (p<0.01), Pakistan ~0.804 (p<0.01). Financial leverage tends to negatively affect ROA; firm size, growth, and board size generally positive; state ownership often negative. Model fit (R-squared adjusted) ranged roughly 0.10–0.17 with significant F-tests (p=0.000). - Moderation (FE with interaction): SC has a positive main effect on ROA in all countries. The BD×SC interaction is negative and significant in China, Russia, and India, indicating SC weakens the positive BD–FP relationship; in Pakistan, the interaction is statistically insignificant. Example interaction coefficients: China about −0.019 (p<0.01); Russia about −0.036 (p<0.01); India about −0.099 (p<0.05); Pakistan not significant. - Dynamic panel (two-step system GMM): Results corroborate FE findings. BD and SC remain positive and significant; BD×SC interaction negative and significant for China, Russia, and India, insignificant for Pakistan. Sargan/Hansen tests support instrument validity; AR(1) present and AR(2) absent as expected, indicating appropriate dynamic specification. Control variables broadly align with FE results (FL and SO negative; FS, OG, BS positive in most cases). - Robustness checks: Using ROE instead of ROA and Tobit models yields qualitatively similar results—BD and SC positive, BD×SC negative (except Pakistan). Inclusion of board independence as an additional control does not materially change conclusions.

Findings support H1 that broader boardroom diversity enhances financial performance, consistent with Upper Echelons and Resource-Based perspectives: diverse boards likely improve oversight, access to heterogeneous skills and networks, and decision quality. The consistent positive main effects across four emerging markets extend evidence beyond developed contexts. The moderating results support H2 that strategic change conditions the BD–FP link: although SC itself is associated with higher performance, its interaction with BD is negative in China, Russia, and India, implying that periods of strategic resource reallocation may dilute or delay the performance gains typically associated with diverse boards—possibly due to coordination costs, integration challenges, or lagged realization of strategic benefits. The absence of a significant interaction in Pakistan suggests contextual heterogeneity in how SC reshapes the diversity–performance pathway. Overall, the results indicate that promoting board diversity can improve performance, but managers should recognize that intensive strategic change may temporarily weaken the translation of diversity advantages into financial outcomes.

Using a multidimensional BD index and a composite SC index across 240 non-financial firms in four emerging markets (2008–2020), the study shows that boardroom diversity positively relates to financial performance, while strategic change weakens this relationship in three of the four countries examined. Results are robust to alternative estimators (robust FE, two-step system GMM), outcome measures (ROE), and model specifications (Tobit, additional governance controls). Contributions include: evidence from emerging markets, explicit modeling of SC as a moderator, a multidimensional diversity construct beyond gender, and rigorous handling of endogeneity. Policy implications emphasize promoting diversity (gender, age, expertise, nationality, education, tenure), setting targets and governance processes to support diverse boards, and carefully planning SC initiatives to avoid undermining diversity-driven performance gains. Future research could explore additional emerging markets, sectoral differences, alternative performance and SC measures, and mechanisms through which SC moderates the diversity–performance linkage.

The authors note that the study has limitations but do not enumerate them explicitly. They indicate that these limitations suggest avenues for future research.