A Logit Model of Laundry Detergent Brand Choice in Melbourne

2017-06-06T01:35:48Z (GMT) by Fry, Tim R. L. Longmire, Ritchard
The Consumer Panel of Australia data collected by The Roy Morgan Research Centre is used to investigate the determinants of brand choice behavior in the laundry detergent market. To reduce potential problems of heterogeneity in the market all purchases of the eleven leading brands in the Melbourne metropolitan area over the period July 1992 to June 1993 are considered. The purchases over this period are divided into an estimation sample and a holdout sample. A mixed Logit model is then fitted using the estimation sample. Price, brand loyalty, household size and household income are found to be significant in explaining brand choice behavior. A test for the, potentially restrictive, property of Independence of Irrelevant Alternatives (IIA) is carried out and it is found that IIA is acceptable for this data. The model appears to fit well and estimated elasticities are presented. The estimated model is then validated using the holdout sample. It is found that the model is very good at predicting the brand choices made by consumers in the holdout sample.