shift the demand curve, which is an illustration of equation (8). In other words, milk
advertising is more effective in shifting milk demand when milk prices are higher
(the dispersion between DR and D0 gets wider). In this instance, a satiation
phenomenon may be at work whereby the advertising elasticity increases as the
quantity consumed decreases. From a policy perspective, the positive term γ in the
milk equation would imply that milk producers might enhance profits by timing
advertising to coincide with high-price periods. Conversely, figure (2b) indicates
that the soft-drink advertising shifts its demand outward (from D0 to DS) but it is
more effective in doing so when soft-drink prices are lower (from DS to DR). The
point is that the duality relation permits a richer interpretation of the interaction
parameter than otherwise possible.
Figure 3 plots the retail prices of the four beverages in real terms. Milk has
the second highest prices in the group. Although milk has stable prices in the most
recent 15 years, its relative price to other non-alcoholic beverages rose gradually
from 1 in 1995 to 1.28 in 2005, a large increase of 28% (Kaiser 2006). The current
high prices of milk, coupled with our finding that milk advertising is more effective
when milk prices are high, warrant the continuous existence of the milk check-off
program, which funds the generic advertising for milk. On the other hand, since
soft-drink prices have been low and declining in the past 30 years, our finding that
soft-drink advertising makes its consumers more sensitive to the price decline
indicates that soft-drink producers enhanced their profits by the advertising-induced
rotation in the demand curve. Conversely, since the prices of coffee and tea have
been the lowest in the group and have shown downward trend in the past 10 years,
16