Household- and Market-Level Perspectives on the Peter Pan Peanut Butter Recall Using Nielsen Homescan Panel Data

Using household level scanner data for 2006, 2007, and 2008, this dissertation consists of four studies, which present household- and market-level analyses of food safety issues concerning the 2007 Peter Pan recall on the demand for peanut butter at the category level and at the brand level. Findings of the first study suggested that the recall had a statistically significant positive effect on the demand for peanut butter at the category level. At the brand level, spillover effects were evident in that the demand for Jif was positively affected, while the demand for Skippy was negatively affected.

The second study examined structural change in the demand for peanut butter using demand system models corresponding to the pre-recall and the post-recall periods. Matrices of own-price, cross-price, and expenditure elasticities were calculated for both recall periods, and upon comparison, there were statistical differences in the corresponding estimated elasticities. In general, most price elasticities in the post-recall period were larger in absolute value than the comparable elasticities in the pre-recall period.

The third study investigated the impact of household socio-economic characteristics associated with choices to purchase peanut butter across the pre- and the post-recall periods. Four choice scenarios were no buy-no buy, buy-no buy, no buy-buy, and buy-buy. Socio-economic characteristics considered included age, employment, education, race, ethnicity, presence of male and/or female household head, region, age and presence of children in the household, household size, and income. While the results varied by brand, region was the socio-demographic characteristic that was consistently significant among the choice scenarios for the respective peanut butter brands.

Conditional on households purchasing peanut butter in both the pre- and the post-recall periods, the final analysis examined the influence of the same aforementioned socio-economic variables as well as the change in the own price on the change in the quantity purchased. The results varied across brands, but the principal drivers of the conditional change in the quantity purchased were the change in the own price and the age and presence of children in the household.