가산자료모형을 이용한 연령별 활어수요함수 추정에 관한 연구

Abstract

This thesis aims at estimating age-stratified demand functions, using count data which is conducted by interview survey method, and their consumption behaviour analysis for raw fishes which is an important portion of fisheries industry. The consumption behaviour analysis for raw fishes is largely composed of choice factors of consumption places for raw fishes (such as convenience, freshness, sanitary condition, price acceptance, service level and variety of species), external ornament frequencies, preferences for raw fishes compared with fresh fishes and meats, consumption satisfaction based on tastes, health and convenience for raw fishes. As a result, freshness comes prior to the price level as the priority of choice factors of consumption places for raw fishes. Rather than the younger, the older customers are more tended to select sanitary condition factor. Regardless of the age groups, consumers prefer eating out for raw fishes to consuming at home, while the overall amount consumed is higher in older age group than the younger. In addition, the consumption of raw fishes is preferred to that of fresh fishes and meat in a field of food consumption preferences, of which raw fishes have a comparative advantage over other foods. A main purpose of this thesis is to estimate age-stratified demand functions for raw fishes using count data model, in which consumption frequencies for raw fishes is a dependent variable; expenses per raw fish consumption, income, age and consumption preferences are independent variables. A Poisson model and a negative binomial model in count data models are applied for estimating demand functions for raw fishes in order to account characteristics of count data (non-negative discrete data.) The survey was conducted for 64, 66, 69 and 63 consumers of twenties, thirties, fourties and fifties, respectively, using personal interview method in Busan Jagalchi Fish Market and other large fish markets during May 2008. Respondents were asked about how often they eat raw fishes, income, age, consumption preferences for raw fishes, and so on. Multiple analysis of variance is employed for testing hypothesis of which null hypothesis indicates mean equality of every dependent variable of each age-group. As a result, the null hypothesis is statistically rejected at the 1% level, which means mean vector for every dependent variables per age-group is different; thus, age-stratification demand analysis has a significant implication. In a Poisson model and a negative binomial model, all parameters estimated are statistically significant at the 1%, 5% and 10% level, and theoretically valid. Using the results based on a negative binomial model, the marginal effect of consumption frequencies for raw fishes on the increase of one million won in annual income is estimated as 0.32 in the forties group which is more sensitive than any other in all age groups. It implies that an increase of one million won in annual income affects an increase of 9,415 won in annual consumption cost for raw fishes. In a different view of marginal effect, an increase of 10,000 won in cost per consumption makes annual consumption frequencies be reduced to 3.2 frequencies in thirties group, and then it makes annual consumption cost for raw fishes of 84,180 won down.