While there is rich literature covering theoretical concepts of transaction costs very few empirical strategies have been provided to estimate them. The theoretical framework proposed in this paper is based on a unit value decomposition and defines transaction costs as the difference between a unit value and a frontier value realized in a situation without transaction costs. Estimates of transaction costs are obtained by means of stochastic frontier models with the data from Kenyan maize farmers. We find a magnitude of 12-18% for maize transactions in rural Kenya and identify drive time, market distance, education and counterparts in negotiations as main determinants.