Second, the non-exact identification of the long-run demand system for the group
comprising the behaviours of highly educated individuals does not allow us to dis-
cuss the net substitution effects of changes in healthy over unhealthy food prices.
However, those who achieved, at maximum, the level of secondary school are sig-
nificant and of the expected (higher) dimensional effect with respect to the average
population. Individuals in this subgroup respond to changes in the rise of relative
prices of healthy foods by consuming more unhealthy foods, 10 basis points more
sensitive with respect to the whole sample estimated at sample average.
Thus, in line with the results of Ball and Crawford’s review (2005), less educated
Italians tend to prefer eating unhealthy foods, with higher calorie intake. The
estimated elasticity of substitution also supports the remarkable rate of growth of
the share of overweight and obesity prevalence for less educated individuals (2.8)
with respect to those with higher levels of education (0.8)14.
Third, among individuals grouped according to age, the insignificant cross-price
responses of the youngest adults were in line with data shown in Figure 3, in which
individual responses of quantities are very little sensitive to changes in relative food
prices. As an approximate explanation, this age group spends least on healthy foods
such as consumers tend to substitute this category of food less.
Bearing in mind that healthy food prices rise more than those of unhealthy ones,
the other age groups were found to make a key contribution to the (net) elasticity
of substitution by favouring the consumption of unhealthy foods. Higher elasticities
of substitution for middle and advanced ages with respect to those estimated for
the general population, leads to the fact that this group contributes more to body
weight increases. This is particularly interesting from a policy perspective, because it
suggests that the food consumption bundle may partly change with age, if consumers
believe that their substitution costs less though in line with Baum (2007) age changes
are not projected to increase obesity substantially.
Fourth, the substitution effects towards unhealthy foods which interest people
living below the relative poverty threshold are, remarkably, 6 basis points higher
than that of the mean of the population. As found in other countries by Komlos
and Baur (2004), the relative price mechanism for adults of lower economic status
passes to specific allocation of their disposable income on food that covers calorie
requirements more easily by substituting low-calorie products with cheaper high-
14 It is worth noting that these results obtained for the sample of less educated individuals, were computed by
averaging the growth of each group from 1999/2000 to 2005.
21