The farm sector in Japan is subject to significant economic distortions. These distortions, consequences of legislative activity in the 1950's and 1960's, have resulted in a flagging agricultural economy. Total agricultural output has declined significantly over the past three decades and the amount of farmland being abandoned is at record levels. Japan's Prime Minister, Shinzo Abe, has targeted the agricultural sector for structural reform as part of a plan for overall economic growth. The success of reform efforts will ultimately depend on the state of productivity in Japanese agriculture and the factors causing farmers to exit farming. Internal restructuring is critical to intrasectoral growth, but the dynamics of these processes are not well understood. Standard economic models based on the behavior of farm households motivate the use of reduced form statistical modeling. A Bayesian hierarchical model is specified to determine linkages between net farm exits over a period of time and relevant economic and demographic variables. Hierarchical models make it possible to utilize data at both the prefecture and town levels. Because several measures of productivity and market distortion are only available at the prefecture level, this modeling flexibility is especially important. The results of this approach make it possible to assess the factors influencing aggregate entry and exit from the farm sector. Additionally, this study demonstrates the importance of productivity dynamics in understanding the changing compositions of agricultural economies.