Effect of Fee-Based Investment Strategy on the Financial Performance of Commercial Banks in Kenya

Abstract

This study examined the effect of fee-based investment strategy on the financial performance of commercial banks in Kenya. The persistence of interest-rate volatility, the 2016–2019 statutory interest rate cap and the accelerated digitisation of financial services have progressively weakened the reliability of interest income, yet the empirical evidence on whether fee-based income translates into superior shareholder returns remains fragmented and inconclusive. Anchored in the Resource-Based View, Portfolio Theory and Agency Theory, the study adopted an explanatory longitudinal research design implemented through a panel data methodology. The target population comprised all 43 commercial banks licensed by the Central Bank of Kenya over the period 2015 to 2024. Applying inclusion and exclusion criteria relating to continuity of licensed operations, completeness of audited disclosures and availability of board composition data yielded 38 qualifying banks and an unbalanced panel of 368 bank-year observations. Secondary data were extracted through a structured data collection checklist from audited annual financial statements, Central Bank of Kenya Bank Supervision Annual Reports and Nairobi Securities Exchange investor disclosures. Fee-based investment strategy was operationalised as income from fees and commissions expressed as a proportion of total operating income, while financial performance was proxied by Return on Equity. Employee Share Ownership Plan, officer directors, non-officer directors and affiliated directors were incorporated as governance control variables. Data were analysed using descriptive statistics, pairwise correlation and hierarchical panel regression in Stata. Diagnostic testing confirmed normality of residuals (χ²(2) = 4.60, p = .125; Jarque-Bera = 4.213, p = .121), absence of first-order serial correlation (Wooldridge F(1, 38) = 1.535, p = .203) and absence of multicollinearity (all VIF < 3). The Breusch-Pagan test detected heteroscedasticity, which was remedied through bank-clustered robust standard errors, and the Fisher-type Phillips-Perron test confirmed stationarity of all series at level. The Hausman test supported the fixed-effects specification, which explained 51.18 per cent of the within-bank variation in Return on Equity (F(4, 326) = 85.45, p < .001). Fee-based investment strategy was positively and significantly correlated with Return on Equity at the bivariate level (r = .394), but the association did not survive multivariate estimation: the coefficient was positive yet insignificant in the hierarchical specification (β = 0.590, p = .117) and negative and insignificant in the full specification (β = −0.087, p = .447), so H₁ was not supported. The governance variables, by contrast, exerted strong and significant effects non-officer directors (β = .546, p < .001), officer directors (β = .247, p < .001) and Employee Share Ownership Plan (β = .233, p < .001) while affiliated directors were insignificant (β = −0.058, p = .404). The study concludes that the apparent profitability of fee-based strategies in Kenyan banking is attributable principally to the institutional capabilities and governance structures of the banks that pursue them rather than to fee income as such, and that board independence is the more consequential determinant of shareholder returns. It recommends that banks treat investment in scalable service-delivery capability, rather than the expansion of charges in itself, as the operative lever; that they strengthen non-executive board representation; and that regulators weigh consumer-protection oversight of bank charges against the sector's need for sustainable non-interest revenue.