Forecasting Meat Prices Using the Food Demand Survey (Foods)

Abstract

Food price changes have important implications for agribusinesses, consumers, and policy makers. Better predictions of food prices should allow for more rapid and efficient adjustment to changing market conditions. This research seeks to determine whether a new source of data from a monthly, nationwide survey of food consumers, the Food Demand Survey (FooDS), is predictive of meat prices included in the food component of the Bureau of Labor Statistics Consumer Price Index (CPI). Unlike many previous efforts to forecast components of the CPI, this study relies on a direct measure of consumer preferences and their stated expectations about future prices and consumption. We compare the predictive performance of simple autoregressive models (where previous prices are used to predict future prices) to models that include data from FooDS. We find that, in most cases, the best fitting models are those that include consumer survey data from FooDS, suggesting that direct measures of consumer preferences and expectations can be used to better anticipate future price changes.