This study examines the performance of logistic regression in predicting probability of default using data from a microfinance company. A logistic regression analysis was conducted to predict default status of loan beneficiaries using 90 sampled beneficiaries for model building and 30 out of sample beneficiaries for prediction. Age, marital status, gender number of years of education, number of years in business and base capital were used as predictors. The predictors that were significant in the model were marital status, number of years in business and base capital. The explained variability in the response variable in the logistic regression was very weak.