received much less attention in the literature. Recent research by Kinnucan et al.
(2001) firmly rejected the hypothesis that non-alcoholic beverage advertising has no
effect on the level of demand for the individual beverages. Specifically, they found
advertising redistributed demand within the non-alcoholic beverage group (juices
benefited the most from advertising), but had no effect on the overall group demand.
What is not known is whether the advertising affects the slopes of the demand
curves. Given the firm rejection of no shift effect, this would appear to be an
especially promising group in which to test whether there is a rotation effect.
Prior to model specification we explain how advertising affects slopes of the
demand curve based on Johnson and Myatt’s work (2006) and distinguish between
curve rotation and elasticity change based on the theoretical paper by Kinnucan and
Zheng (2004); thereafter we introduce price-advertising interaction terms into five
different demand models and develop methods to measure curve rotation caused by
the interaction, and derive some propositions about price-advertising interaction
using Frisch’s duality relationship.3 The results of the hypothesis tests are then
presented employing time-series data. The article concludes with a brief summary
of the key findings. Overall, this research is a full empirical extension of the work
of Kinnucan and Zheng (2004).
How Advertising Rotates the Demand Curve
This section illustrates advertising’s shift and pivotal effects on demand using a
numerical example. According to Johnson and Myatt (2006), advertising can shift a
demand curve by shifting the location of consumers’ willingness-to-pay (WTP); it