An integrated policy evaluation tool is proposed for assessing the effects of agricultural policy measures using all the information available at farm level. The tool combines the positive mathematical programming methodology with the cluster analysis technique by using the same panel of data. The PMP model proposed here allows to measure the effects of policy in term of agricultural supply responses including output market price variations. The novel procedure by which the PMP model is articulated permits to recover the set of farm level demand functions for agricultural products and the cost function characterizing the given sample of farms. Cluster analysis is useful for better appreciating the behaviour of farms before and after the policy scenario analysis by considering the transfers of farms among clusters. A decoupling scenario assessment presents the responses that the integrated tool can provide for evaluating agricultural policy instruments.