Imputing Dairy Producers' Quota Discount Rate Using the Individual Export Milk Program in Quebec

Current Agriculture, Food & Resource Issues

M. Doyon, C. Brodeur and J-P. Gervais

Table 1 OLS estimates for the prediction equations

Note: The numbers between parentheses are the standard errors associated with each coefficient.

solids. The IEM program was administered by an independent agency, and export prices
for each contract settled between producers and processors are publicly available. The
independent variables of the forecast model were selected to balance the necessity for
producers to form their subjective distribution of prices using all information available
and the desire to keep the model parsimonious in terms of the number of parameters to
estimate. The Breusch-Pagan Lagrange multiplier test for the null hypothesis of a diagonal
variance-covariance matrix of residuals produces a statistic of 0.05 (p-value of 0.83).
Thus, there are no efficiency gains related to estimating the forecast equations jointly.
OLS estimates for each equation are reported in table 1, along with their standard errors
between parentheses. All independent variables are statistically different from zero at the
90 percent confidence level and have the expected algebraic sign. The statistical fit of
each equation is good, as the adjusted R² measures in (2) and (3) are 0.90 and 0.89
respectively.

Simulation

The final task involves specifying the objective function the producer optimizes. Suppose
that the utility function of the producer is of the constant relative risk aversion (CRRA)
type:

ln π if у = 1

where χ is the Arrow-Pratt relative risk aversion parameter. This is a convenient way to
approximate producers’ risk preferences, and it has the advantage of not assuming that
risk preferences are independent from initial wealth level. The CRRA assumption implies
that absolute risk aversion is decreasing in wealth. The empirical strategy is as follows.

123

<<   <   1   2   3   4   5   6   7   >   >>