Conventional explanations for the relative decline of agriculture in developing countries stress secular, demand-side phenomena, specifically Engel effects. This view has been challenged by quantitative analyses emphasizing supply-side effects such as differences in factor endowment growth rates. The innovation in this paper is to investigate the extent to which agricultural decline is in fact generated by policies rather than by fundamental preference or endowment shifts. Econometric results using Thai data indicate that policies are strongly influential, but that the direction and strength of influence varies over time. We explore implications for the interpretation of past development strategies and future policy formation.