Because of the complex survey design of CEAP-ARMS, variances of estimated parameters
are calculated based on standards established by the National Agricultural Statistical Service -
USDA, using the delete-a-group jackknife variance estimator (Kott, 1997; Dubman, 2000; El-Osta,
Mishra, and Ahearn, 2004). The delete-a-group jackknife procedure was used to estimate the
variances of the censored regressions.
For Model I, acreage allocation decisions for 2005 corn fields (without structural practices,
with only infield structures, with only perimeter-field structures, or with both infield and perimeter-
field structures) were modeled as a function of normalized per-unit input prices for nitrogen,
agricultural wages, and diesel fuel, as well as three technology choice variables and three structural-
installation time-period variables. These acreage-supply equations were estimated jointly for
conservation program participants and non-participants. Input prices were normalized using
average corn price (per bushel) by State.8 These normalized prices are expected to reflect the effect
of the primary economic factors affecting a conservation program participant/non-participant’s
perception of field production profitability for the alternative acreage allocation choices for the
field. Conservation technology class variables (for infield, perimeter-field, or both structures) and
installation time-period variables (installed in 2005, within the last 10 years, or prior to 1990) were
defined as (1,0) variables, where 1 defined participation for that variable.
For Model II, field acreage supply equations were modeled similar to Model I, but with
additional covariates to control for influences of farm structure, field crop management, and several
environmental attributes. Farm structure was proxied using total cropland acres operated for the
farm and a variable measuring land tenure (proportion of acres owned to total farm acres operated).
Total cropland acres are hypothesized to measure the influence of farm size on operator decisions to
install working-land conservation structures. Land tenure is hypothesized to reflect differences in
13