The models were estimated using ordinary least squares. The parameter estimates for
research expenditures and expected price are used in the calculation of the annual shift in the
supply curve, which is then used in the calculations of the consumer and producer surplus
measurements. The parameter estimates for all the exogenous variables resulting from the
empirical estimation procedure are shown in Table 2. P-values are displayed so readers can
evaluate statistical significance of the coefficients. In general, the estimated coefficients are
highly significant statistically and have signs that conform to economic theory and magnitudes
that pass a common sense test.
The estimated coefficient on acreage in the peanut regression had a negative sign, which
could at first glance indicate that as acreage harvested increased, the quantity produced
decreased. This result appears to conflict with prior theoretical expectations. However, a little
thought suggests that the coefficient is the result of the federal quota program for peanuts and the
aggregate time series nature of these data. Because the peanut quota is in pounds, not acres, as
average yields rise over time the acreage (planted and) harvested will tend to decline unless
farmers choose to grow a greater amount of additional peanuts outside the quota program. Thus,
we are picking up this long-run correlation in our regression results, not any causative effect of
acreage that reduces production.
12