Pi = ∑ WP ,where Wj = ( PQ )∕(∑ PQu )
jj
and j denotes the commodities in the same sub-category. Because each fruit or vegetable is a
potential substitute for, or complement with, other fruit and vegetables, all sub-categories are
included in each equation.
Since price elasticities of demand are a primary focus of this research, each equation is
specified in its double logarithmic functional form to give direct demand elasticities. Given that
economic theory suggests a link between income and demand responsiveness, it is hypothesized
that lower-income consumers will show higher price elasticities than higher-income consumers.
Stated differently, the variable Pmkt in equation 2 is expected to be negative and statistically
significant.
Empirical Results for Fruit
Table 3 provides the estimated results for all sub-categories of fruit and vegetables. Store
variables and other independent variables are included in this table, but this discussion will focus
on own-price elasticities for the eight sub-categories of fruit. However, it should be noted that
statistically significant differences are found among the stores for all sub-categories of produce.
For example, store 1, a higher-income store, is used as the reference store and the results for
citrus show that purchases of citrus for stores 2 and 3 are statistically insignificant from
purchases in store 1. Lower-income stores, 4, 5 and 6, are shown to have citrus purchases that
are different from those in store 1. That is, relative to store 1, the intercepts shift downward for
stores 4, 5 and 6. Further, the variable produce sales, a proxy for income, shows that most fruit
and vegetables have positive and statistically significant elasticities that are less than one,
classifying these sub-categories as necessary goods, not luxury ones. Based on this classification
criterion, berries, soft fruit, melons and grapes are shown to be luxury goods. Lagged quantities