Accurate assessment of farmers' credit constraint condition is important in order to understand the circumstances under which credit would have its greatest impact. In this study a switching regression model was used to determine the impact of credit on smallholder dairy farms in the East African highlands using farm level data from Ethiopia and Kenya. Farmers were classified as credit constrained or credit non-constrained based on their responses from the farm level surveys. No consistent relationship was found between farmers' credit constraint condition and their borrowing status. Most of the variation in milk output per farm was explained by the number of crossbred milking cows in the dairy herd. As credit is likely to facilitate investment in crossbred dairy cows it will have substantial impacts on smallholder dairy farms especially if it is targeted to credit constrained farms. © 1998 Elsevier Science B.V. All rights reserved.