Milk Demand Heterogeneity and Cost-Pass Through

A Micro Data Approach toward Competition Structure

Abstract

Nowadays, some studies have pointed out that retailers have strong buyer power in purchasing milk from processors, but there is not enough knowledge about the differences in market competitiveness by brands or by retailers. To consider the concrete policy implication of milk markets, I use microlevel data, and estimate demand function using the random coefficient logit model of Berry et al. (1995). It can control the endogeneity of price due to unobserved product characteristics. Then, I estimate demand elasticities by brand-retail combination at every prefecture-monthly level. Markup is recovered from demand estimates. The results show that NB milk has more markup than COOP or PB milk on average. By counterfactual simulation, I find that the 13％ NB milk in total has potentially negative cost-pass through. Finally, I demonstrate the geographical differences in pass-through rates.