A COMPARISON OF CAPITAL MEASURES IN U.S. AGRICULTURE

This study compares two panel data sets that measure capital input at the state-level in U.S. agriculture. Despite a number of similarities between the data sets, such as the composition of assets, aggregation procedures, and time frame, an examination of the final estimates of capital service flows reveals that they are drastically different for all 48 contiguous states. We examine the methods used to construct the capital series for each data set, consider some important differences in data sources and the types of data used to construct the capital measures, and outline the main assumptions concerning depreciation, service lives, interest rates, aggregation, and the scope of goods included in each of the data sets. The analysis indicates that an important statistic in the index of capital services in U.S. agriculture is the stock of buildings on farms. We conclude that the primary difference between the measures of capital input in the data sets relates to differences in estimates of the stock of buildings on farms. Given the apparent importance of the measure of the stock of buildings in the aggregate index of capital services in U.S. agriculture, more research is needed to ensure that the measure of the stock of buildings is accurate and meaningful. Once this has been accomplished there should be more agreement on an accurate measure of capital services in U.S. agriculture.