The econometrics of demand systems with special reference to commodity group data for Bahrain
The mam objective of this study is to estimate the demand relationships among commodity groups using Bahrain quarterly time series data for the period 1979- 98. Three main demand systems are presented and estimated, namely the Linear Expenditure Demand System (LES) which was introduced by Kelin-Rubin (1946-47) and developed by Stone (1953-54), the Rotterdam System (RM) which was introduced by Theil (1965) and Barten (1966), and the Almost Ideal Demand System (AIDS) introduced by Deaton and Muellbauer (1980). Also, the variables reflecting the effects of habit on purchases are incorporated into three main demand systems. Model selection procedures are applied to select the best model to reflect the Bahrain data. Based on procedures, the static Linear Almost Ideal Demand System (LAVAIDS) is selected among other static demand systems, namely, the linear expenditure demand system and the Rotterdam system models. The selection is based on average information criterion. However the dynamic LAYAIDS is selected over its static counterpart and recommended for the future application for Bahrain data. The selection is based on the likelihood ratio test. Further, the dynamic LAVAIDS satisfies all the restrictions implied by demand theory. All the compensated own price elasticities are negative as expected and the expenditure elasticity classifies the food & beverage, clothing & footwear,housing, and transportation as necessity products, while the other commodity group is classified as luxury products. The second main objective of this study is to test for unit roots and order of comtegration in Bahrain commodity group data. The results indicate that most of the time senes that will be used in estimating the linear AIDS model, such as total real expenditures on various commodity groups, prices, budget shares are shown to be integrated of order one. The application of Johansen and Juselius(1990) Full Maximum Likelihood approach in this study confirm that equilibrium relationships exist between the variables that make-up the LAVAIDS model. Second, the homogeneity of degree zero of price, postulated by consumer theory, is rejected by the data. The argument that time series issues are responsible for this rejection is not always true. The study also focuses on a) Engel Curve and b) income distribution and poverty in Bahrain. The advantage of a cross- section study of the Engel Curve is that consumer demand theory is based on micro-demand data (individual's data). Income elasticities that are estimated from cross-section data are useful in predicting changes in aggregate time series data if the effects of price changes have to be eliminated from time series analyses. Therefore, seven Engel Curves for eight commodity groups are estimated using data from the Bahrain household survey for the period 1994-1995. Studying income distribution and poverty in Bahrain is important. Inequality of income and poverty influences the pattern of household expenditure. It also influences the welfare of households in the country. Therefore, studying income distribution and poverty could guide the policy makers in Bahrain to improve the living standard of households. The analyses of income distribution and poverty in Bahrain are based on the household expenditure and income survey data for the period (1984-85 to 1994-95), involving Gini coefficient and other measures.