This report documents the research results from modelling Japanese household food consumption behaviour by using household-level data for 1997. The FIES data provide a rich source of information for estimating food demand in Japan. The report presents only two empirical food demand models: one for 11 aggregate foods, including rice; and one for seven meats. More can - and will - be done to learn more about Japanese food consumption behaviour from this database and from similar data for 1996.
This report presents analyses of descriptive statistics for food consumption by demographic groups and econometric estimations of the two food models. Several single equation demand models and the AIDS model are estimated.
The econometric results indicate that rice is not an inferior good in Japan, and that demographic variables are important indicators of the future trend of Japanese rice consumption. According to the meat demand analysis, Japanese dietary patterns have become increasingly Westernized.
One puzzling set of results is related to the estimated own-price elasticity of rice demand. The report's estimates of this elasticity range from -1.5 to -1.9, which are very high. Since this elasticity has important implications for the impacts of Japanese agricultural and trade policies, it needs to be assessed carefully. The reliability of the estimates can be seen from the fact that the range is relatively robust among several model specifications. It should be noted, however, that the survey data are observations of purchase behaviour, which may not be the same as consumption behaviour. This is especially true for rice because households buy rice in response to changes in price: substantial variations in rice prices among households and months were recorded. These differences might have been caused by quality differences; high-income households tend to buy higher-quality rice than lower-income households do. An attempt is made to address this question by estimating various demand models by income level, but the estimates of the own-price elasticity of rice remain very high. This report does not answer the question regarding differences between purchase and consumption. The issue is important and a more elaborate modelling framework, based on the concept of infrequency of purchase, should be used to address it.
Some of the findings contradict those of previous studies of aggregated time series data and shed more light on the continuing debate about the income and price effects of food demand, especially for rice in Japan. The results presented in this report should also be useful to agricultural policy-makers in assessing the impacts of various agricultural and trade policies.
