This study constructs regional crop production models of 8 cities in Korea using the PMP Leontief-quadratic and the generalized-CES models suggested by Mérel and Bucaram (2010) and Mérel et al. (2011). A generalized maximum entropy estimation method was used to derive the set of supply elasticity estimates allowing a complete calibration of the models. We demonstrate that those two PMP models fit the Korean regional data quite well. The impacts of two policy changes, raising the rice producer price under the rice direct payment system and raising the minimum wage for hired rural labor were simulated. Our simulation results show that the policy impacts are varying over region and crop. Moreover, it is shown that the choice of the functional specification substantially affects the direction and the magnitudes of the policy impacts. More specifically, production responses are more flexible in the CES specification rather than in the Loentief specification, and the former specification suggests that a substantial increase in rural minimum wage may result in a decline in capital and labor employment in rural areas while the latter specification suggests that the overall resource use in rural areas will not be affected by the policy.