July 1995 Time series estimates can provide an accurate picture of past behavioral relations, but they are not an adequate basis for forecasting the impact of policy reform. They typically generate a downward-biased estimate of the response to a credible reform. Schiff and Montenegro review several studies of the aggregate agricultural supply response. Using both economic and econometric reasons, they argue that time series estimation typically generates a downward-biased estimate of the response to a credible reform. Even though time series estimates can provide an accurate picture of past behavioral relations, they do not provide an adequate basis for forecasting the impact of policy reform. This is especially true in developing countries, where policy reforms involve large changes and have included agricultural price reform, industrial trade liberalization, financial sector reform, and macroeconomic stabilization. Under those circumstances, parameter values obtained under the former policy regime have little relevance in the new regime. Schiff and Montenegro also argue that investments in public goods should be viewed as complementary to, not competitive with, price policy. They claim that to select the policy with the biggest impact on output makes no sense. They provide what they consider to be better criteria for choosing the best from alternative policies. This paper--a product of the International Trade Division, International Economics Department--is part of a larger effort in the department to understand the impact of policy reforms.