We used results from previous studies combined with information about commodity-specific CSEs to estimate the likely changes in prices of farm commodities given the removal of U.S. farm subsidies, including indirect subsidies provided by border measures. Then we used these estimates of farm commodity price impacts in a simulation model of the supply chain to calculate the implied changes in caloric consumption of ten categories of food and beverage products.

Farm commodities used as ingredients represent a small share of the total cost of retail food products, and this share has been shrinking for all farm commodities over the past three decades (USDA-ERS, 2008). On average the farm commodity cost share is approximately 20 percent, but it varies widely: for grains, sugar, and oilseeds, it is less than 10 percent; for soda, a food product that is often associated with obesity, the share is approximately 2 percent; for retail fruit and vegetable products—fresh and processed—it is approximately 18 percent; and it is closer to 35 percent for meat and eggs. In the U.S., food consumed away from home accounts for nearly 37 percent of food expenditures and 33 percent of average daily calories (Figure 2), but the cost of farm commodities as a share of the value of foods consumed in this category is tiny.

The small farm commodity shares of food costs mean that small commodity price impacts from removal of farm policy would lead to very small effects on consumer costs of food and beverages, especially for some of the categories most commonly associated with obesity. In addition, if such changes in the costs of food products were not fully passed on to consumers, they would see even smaller percentage changes in retail food prices.

We calculated changes in retail prices associated with the removal of farm subsidies for ten food product categories using simulated changes in commodity prices together with recent farm-retail marketing margins. The mechanisms are complex because some food items use multiple commodities and some commodities, such as feed grains, themselves are inputs into other farm products. For example, the retail cost of meat and dairy products would increase because the removal of subsidies would increase the cost of corn. The farm cost of livestock represents only about one-fifth of the retail cost of meat. Hence, assuming corn and other feedstuff represent about 30 percent of the farm cost of meat and dairy items, a 5 percent increase in the farm price of corn would imply a 1.5 percent increase in the farm cost and a 0.3 percent increase in the retail price of meat for consumers. Similar calculations apply for other retail foods, with the multipliers from farm commodity prices to retail prices varying with the farm-commodity share of the consumer food dollar.

Policy simulations

We used an equilibrium displacement model (Okrent, 2010) to simulate the effects of farm subsidies applied to eleven agricultural commodity categories (Table 1) on caloric intake patterns of 10 categories of food products.

The food products include eight exhaustive categories of food-at-home products—cereals and bakery products, red meat, poultry and eggs, fish and seafood, dairy products, fruits and vegetables, nonalcoholic beverages, other foods including fats and oils, and sugars and sweeteners: a composite food-away-from-home product; and alcoholic beverages. Four simulation experiments were performed to better understand how various policy changes would affect commodity prices, food prices, food consumption, and ultimately annual per capita caloric intake. We first simulated the effects of removing only grain subsidies, based on published measures of policy price impacts consistent with policies in the mid-2000s. The other three simulations show the effects of removing all U.S. farm subsidies, including indirect subsidies provided by trade barriers that have applied in different time periods—with price impacts based on the published measures plus CSEs for different time periods—along with the measures of grain subsidy impacts on prices used in the first simulation. One set of simulations is based on CSEs in 2006, a reasonably representative recent year just before the spike in commodity prices after which CSEs have been relatively low. The other two simulations are based on CSEs for the previous 10 years and the previous 21 years. The percentage changes in farm commodity prices for each of the four simulations are given in Table 1, along with key findings.

Our results indicate U.S. farm subsidy policies, for the most part, have not made food commodities significantly cheaper and have not had a significant effect on caloric consumption. Eliminating U.S. grain subsidies alone would lead to a small decrease in annual per capita caloric consumption—simulated to be 977 calories per adult per year, which would imply a 0.16% per year reduction in average body weight assuming 3,500 calories per pound. In contrast, removing all farm subsidies, including those provided indirectly by trade barriers, would lead to an increase in annual per capita consumption in the range of 200 to 1,900 calories—equivalent to an increase in body weight of 0.03% to 0.30%, depending on the size of the policy-induced price wedges to be removed, as represented by the CSEs. The CSEs were generally smaller in 2006 than over the decade 2000–2009 and more than over the longer period 1989–2009. Thus the smaller estimates, based on the 2006 CSEs, are probably the most relevant.

As our results show, the measured economic effects of a simulated policy reform depend on specific characteristics of the analysis:

a. what the reform includes—partial, applied to grains alone, versus more comprehensive;

b. when it applies—which determines the size of the distortions to be eliminated; and

c. the modeling details—in particular, how we allow for shifting incidence between farmers and consumers at home and abroad through different elasticities and different detailed representation of policies.

In estimating price impacts we assumed that the impact of policy change would be transmitted entirely to consumers. In this sense our estimates are at the high end of the feasible range. Nevertheless, our simulated results show fairly small positive or negative impacts on total caloric consumption and thus potentially on obesity. This result holds even when we allow for comparatively large policy impacts on buyer prices of farm commodities, with the sign and size of the effect contingent on whether it is assumed that import barriers that raise the buyer prices of dairy, sugar, and fruits and vegetables are to be eliminated along with subsidies on grains.