摘要：Building on earlier work by Cramer, this paper tests for price-advertising interaction effects in the U.S. non-alcoholic beverage market. Full and restricted specifications of the Rotterdam model are tested along with compensated and first difference double-log models using annual time-series data for the period 1970-2000. Results are mixed in that the full Rotterdam and first difference double-log models fail to reject the null hypothesis of no rotation while the restricted Rotterdam and compensated double-log models indicate rejection. One reason for the ambiguous results may be multicollinearity problems associated with the unrestricted models. Still, our overall conclusion is that there is little solid evidence to suggest that generic advertising rotated the demand curves for non-alcoholic beverages in the United States.
