In 1997-1998 the USDA failed to adjust subsidy rates to reflect changing market dynamics. The consequence was that farmers in neighboring counties received substantially different subsidy payments. The Government Accountability Office (GAO) reported differences as high as $0.08/bushel in neighboring counties. This policy distortion caused farmers to transport and store harvested commodities in adjoining counties where subsidies were higher. It also created a natural experiment to determine the effect of the subsidy rate on crop production.

Political pressure prevented the USDA from quickly fixing the discrepancies, so the differences were allowed to stand for years. For the sake of fairness, the USDA assigns county-level subsidy rates so the spatial distribution mimics the spatial distribution of commodity prices. The farm bill, however, specifies a national average subsidy rate. So if the USDA increases the subsidy rate in one county, they must reduce it somewhere else to maintain the required national average. The only way the USDA could have reassigned subsidy rates and maintained the national average would have been to substantially decrease subsidies for a significant number of farmers. Rather than draw farmers’ ire, the USDA chose to let the discrepancies remain until 2002 when a new farm bill was implemented and large changes in subsidy rates could be blamed on Congress.

Empirical economists’ dreams are made of such policy “quirks” because they create “natural experiments.” Farmers who face very similar growing conditions suddenly receive different subsidies simply because they are on opposite sides of an arbitrary line. It’s almost as if the subsidies were assigned randomly! The map illustrates the magnitude and location of the policy-induced subsidy changes.