This thesis focuses on the estimation of demand function through a system-wide approach, involving five different commodity groups.
Demand studies can be classified into two broad categories. One group of studies is concerned with finding a model for explaining the relationship among the quantity concerned of a single commodity, its price and price of its related commodities, and income. Such a demand model is considered in some of those studies as a part of a system of demand functions, and tests of theoretical properties such as homogenity of degree zero in price and income are conducted. Another group of demand studies is concerned with the problem of allocating total expenditure to an exhaustive set of different commodities. These studies, which usually assume that the problem of deciding how much to consume at any given point of time has been solved, concentrate on the problem of allocation. In this thesis we shall be concerned not with a demand for a single commodity, but with demand systems designed to explain the allocation problem of a rational consumer.
This thesis considers three demand systems: the Linear Expenditure System proposed by Stane (1954) & Powell (1966), the Rotterdam Differential Demand Model developed by Barten (1976) & Theil (1980), the Indirect Addilog Model established by Jorgenson & Lau (1975).
The aim of this thesis is to provide a comparison of these model of both their theoretical properties and their empirical usefulness on the basis of the same set of data. Furthermore the results can be used as basic data for prediction of changes in consumption structure.
The data used for these estimation of all these models are prices, quantities, and total consumption indicated in the National Income Account compiled by the Bank of Korea for the period 1962 to 1982, and consumption expenditure per household by consumption expenditure group of all households in cities shown in the Annual Report on Family Income and Expenditure Survey conducted by the Economic Planning Board for the Period 1970 to 1982.
The results of the estimation indicate that the characteristics of consumer behaviour can be better explained by the data in the Annual Report on Family Income and Expenditure than by those in the National Income Account and that the Rotterdam Model is serperior to the Linear Expenditure System or the Indirect Addilog Model.