SOCIAL SUSTAINABILITY PRACTICES AND FINANCIAL PERFORMANCE OF COMMERCIAL BANKS IN KENYA

This paper examined the effect of social sustainability practices on the financial performance of commercial banks in Kenya. Inclusive lending to women, youth and small and medium enterprises occupies an ambivalent position in banking: it advances a developmental mandate that Kenyan policy has actively promoted, yet it concentrates credit in borrower segments characterized by limited collateral, informal income streams and shorter credit histories, and a substantial share of it is originated under concessional or directed schemes. Whether it contributes to or detracts from profitability is therefore an empirical rather than a normative question. The study adopted a positivist philosophy and an explanatory longitudinal panel design. The target population comprised all 38 commercial banks licensed by the Central Bank of Kenya, a census was adopted, and secondary data were extracted from audited annual reports, sustainability disclosures and regulatory publications for 2015 to 2024, yielding a balanced panel of 380 bank-year observations with complete coverage. Social sustainability practices were measured by the inclusive lending ratio and financial performance by return on assets. Estimation used hierarchical panel fixed effects regression with HC1 robust standard errors, following a full diagnostic battery and a Hausman specification test. The social coefficient was negative and statistically insignificant in every specification: in the direct-effects model (β = -0.018, SE = 0.013, p = 0.159), with firm size controlled (β = -0.011, SE = 0.014, p = 0.419), and with the interaction terms present (β = -0.010, SE = 0.014, p = 0.445). The null hypothesis was not rejected. The result held under pooled ordinary least squares (β = -0.014, p = 0.352), under two-way fixed effects with year dummies (β = -0.017, p = 0.290) and in the post-2021 sub-period (β = 0.031, p = 0.111). The interaction between social practices and firm size was insignificant (β = 0.104, p = 0.293), so the null result is not an artefact of averaging across institutions of different scale. The evidence is consistent with a credit-risk and concessional-pricing cost that offsets the relationship and reputational benefits of inclusive lending within an annual accounting horizon, and with the limitation of a volume-based proxy that captures the quantity of directed credit but not the quality or depth of social engagement.