PurposeThe purpose of this study is to investigate the gender and ethnic diversity–performance relationship in Malaysia from two angles: (1) the impact of political regimes; and (2) a possible nonlinear relationship – at the boardroom and employee level.Design/methodology/approachThis study uses a sample of firms listed in Bursa Malaysia during a sample period that spans two political regimes. Two-stage least squares controlling for firm-specific effects, corporate governance and lagged variables to account for endogeneity issues is used to test the relationship.FindingsFindings show that the political alignment of the ruling government affects the significance of the gender/ethnic diversity–performance relationship. The relationship between board gender/ethnic diversity and firm performance is curvilinear while the relationship between employee gender/ethnic diversity is linear and positive.Research limitations/implicationsFirst, promoting gender/ethnic diversity not only requires strong policy but also political will to lead by example. Political regimes that provide lip-service without effective implementation threaten to derail any efforts in furthering the diversity agenda. Second, the presumption of a linear diversity–performance relationship is fallacious. Further studies, especially in pluralistic societies, must not discount the subtleties of intergroup conflicts. Third, in light of allegations of prejudicial hiring policies, Malaysian firms should embrace diversity, not only in the boardroom, but also among its workforce as employee diversity improves firm performance.Originality/valuePrior studies on gender/ethnic diversity in Malaysia have returned mixed results but thus far, there has been no satisfactory explanation for this phenomenon. This study attributes it to lack of political will and cultural subgroup conflicts – two pertinent issues that were never considered in the literature. Prior studies have also exclusively focused on boardroom diversity. This study goes further by examining employee diversity – particularly important since most empowerment and diversity initiatives are targeted at lower level employees. This study is also the first to provide an objective benchmark for gender diversity (30–35% female directors) and ethnic diversity (less than 40% from one ethnicity) to achieve optimal performance.