Statistical Modelling of Credit Riskiness of Bodaboda Credit Business in Kenya Using Survival Analysis

Dennis Mwenda Muchangi *

Department of Basic Sciences, Tharaka University, P.O. BOX 193-60215, Marimanti, Kenya.

Daniel Mwangi

Department of Basic Sciences, Tharaka University, P.O. BOX 193-60215, Marimanti, Kenya.

Dennis Muriithi

Department of Physical Sciences, Chuka University, P.O. BOX 109-60400, Chuka, Kenya.

*Author to whom correspondence should be addressed.


Abstract

Background: Motorcycle hire-purchase financing in Kenya exposes asset-financing providers to loan default risk that varies throughout repayment.

Aims: To develop a Cox Proportional Hazards (Cox-PH) model to estimate the time-dependent credit riskiness of bodaboda asset financing businesses in Kenya

Study Design:  Survival Analysis Research Design.

Place and Duration of Study: The study was carried out in Kenya. It involved secondary data from credit businesses dealing with Bodaboda. Data was used from the period between 2015 to 2023.

Methodology: We analysed 850 bodaboda credit accounts using Kaplan-Meier estimation, log-rank tests, and a multivariable Cox-PH model fitted via Breslow-approximated partial likelihood, with age, gender, income level, employment status, credit score, loan amount, down-payment ratio, and interest rate as covariates.

Results: Kaplan-Meier estimation revealed a median time-to-default of approximately 14.8 months, with the steepest decline in survival probability occurring between the fifth and twentieth month of the loan term. The multivariable Cox-PH model, fitted on nine covariates, achieved a statistically significant overall fit (LR χ² = 340.90, p < .001) and acceptable discriminatory power (C = 0.714), with credit score, down-payment ratio, loan amount, interest rate, income level, and age identified as statistically significant predictors of default hazard.

Conclusion: The Cox Proportional Hazards model offers a statistically robust and practically interpretable framework for modelling credit riskiness in the bodaboda sector. The recommendation is that lenders should place greater weight on credit history and down payment requirements in loan approval and pricing decisions. However, further research on effect of macroeconomic factors on survival rates would be essential.

Keywords: Survival analysis, cox proportional hazards, credit risk, bodaboda, hire-purchase


How to Cite

Muchangi, Dennis Mwenda, Daniel Mwangi, and Dennis Muriithi. 2026. “Statistical Modelling of Credit Riskiness of Bodaboda Credit Business in Kenya Using Survival Analysis”. Asian Journal of Probability and Statistics 28 (12):26-35. https://doi.org/10.9734/ajpas/2026/v28i12984.

Downloads

Download data is not yet available.