We analyse the impacts of the CAP reforms on technical efficiency of Greek olive farms. We use a production frontier function and a non-monotonic inefficiency effects model which incorporates the influences of exogenous variables on the mean and the variances of farm efficiency. We formulate policy variables (e.g. the direct subsidies) and farm characteristics as explanatory variables in the inefficiency effects model. We use the 1995-2004 FADN data to estimate the production frontier, to derive technical efficiency, and to determine the effects of the explanatory variables. The study shows that the 10-year average technical efficiency of olive farms is 69%. Direct transfers have a negative and monotonic effect on technical efficiency, while the degree of specialization has a non-monotonic effect on technical efficiency.