AGGREGATION BIAS FOUND IN KOREAN DEMAND FOR IMPORTED BEEF

Abstract

The purpose of this paper is to analyze whether the aggregation problem really matters in estimating the Korean demand for imported beef. This paper incorporates the aggregation bias in the demand system and tests whether the aggregation bias is statistically significant with two models: an unrestricted model with aggregation bias variables and a restricted model without the bias variables. The impact of the aggregation bias is also analyzed by comparing the elasticities of the two models. According to the Wald test, the hypothesis of no aggregation bias is rejected and the LR (Likelihood Ratio) test shows that the unrestricted model is statistically better than the restricted model. However, the impact of the aggregation bias on the demand system for imported beef is not so severe in that the elasticities of the two models are not so different and the elasticities of aggregation bias variables are insignificant. This paper utilizes imported beef data of two types (frozen and non-frozen) from USA, Canada, Australia and Rest of World. The aggregation bias is decomposed into country effect, product effect and interaction effect.

The purpose of this paper is to analyze whether the aggregation problem really matters in estimating the Korean demand for imported beef. This paper incorporates the aggregation bias in the demand system and tests whether the aggregation bias is statistically significant with two models: an unrestricted model with aggregation bias variables and a restricted model without the bias variables. The impact of the aggregation bias is also analyzed by comparing the elasticities of the two models. According to the Wald test, the hypothesis of no aggregation bias is rejected and the LR (Likelihood Ratio) test shows that the unrestricted model is statistically better than the restricted model. However, the impact of the aggregation bias on the demand system for imported beef is not so severe in that the elasticities of the two models are not so different and the elasticities of aggregation bias variables are insignificant. This paper utilizes imported beef data of two types (frozen and non-frozen) from USA, Canada, Australia and Rest of World. The aggregation bias is decomposed into country effect, product effect and interaction effect.