necessary information to obtain consistent estimates of σ, β0 , and the covariance matrix.
B. Data and Estimation Results
Producer cost and returns data come from 1997-2000 Agricultural Resource Man-
agement Study (ARMS) surveys conducted by the USDA’s National Agricultural Statistics
Service (NASS). The surveys are independent annual cross-sections in which it is not possi-
ble to identify individual producers across time. They collect producer-level data on input
expenditures, output quantities, and land. Disaggregated input and output price data come
from the Bureau of Labor Statistics for capital and labor, the Federal Reserve for interest
rates and NASS for other inputs and crop and livestock outputs.
I aggregate outputs into a single category and variable inputs into capital services,
energy, materials, labor using a multilateral Tornqvist index (see Caves et al., 1982).4 Since
ARMS surveys record capital assets as estimated market value at year end, I calculate
capital services adapting the methodology of Hall and Jorgenson (1969). Table 1 contains
the summary statistics for the data set.
Since the estimation procedure implicitly assumes all producers have the same gen-
eral production technology (up to the type parameter), I focus attention on one relatively
homogenous area, the “Heartland” Farm Resource Region.5 This region comprises the entire
states of Illinois, Indiana, and Iowa, as well as portions of Kentucky, Minnesota, Missouri,
Nebraska, Ohio, and South Dakota. It is the region with most farms, most cropland, and
greatest value of production (Economic Research Service, 2000).
To control for systemic production shocks such as weather and technological change
that affect the whole region in a given year, I replace the constant in Eq. (5) with an annual
dummy vector d ≡ (d1997, ..., d2000)0, and β0 with the parameter vector δ ≡ (δ1997, ..., δ2000)0.