An analysis of household consumption pattern : a case study in Kabupaten Kerinci Jambi Province, Indonesia

This study aims to examine household consumption behaviour. The main inquiries covered were: (i) whether there is any systematic pattern of consumption in relation to income differences, and (ii) what are the implications of the consumption behaviour in the economic development? The cross-section data consisted of income, expenditures and socio-demographic factors of the household samples from Kabupaten Kerinci, Jambi Province, Indonesia. Information regarding general socio economic and geographical conditions of the kabupaten were also presented to support the inferences drawn from the empirical results. The empirical analysis was carried out by using a comprehensive approach of the demand function which is called the Extended Linear Expenditure System (ELES), developed by Lluch et al (1977) from the Stone-Geary utility function. The partial demand function in the ELES shows the relationship between any expenditure category as a dependent variable and household supernumerary income and socio-demographic factors as the explanatory variables. All prices of the consumption goods have been well captured in the supernumerary income, and this is the 'unique' characteristic of the ELES. Thus, even in the absence of price data, the effects of prices of all the consumption goods can be examined. This study found that consumption expenditure both aggregate as well as disaggregate are significantly determined by household income, family size, location and occupation. In Kabupaten Kerinci, 60 per cent of total household expenditure is allocated for food items. This portion decreases with an increase in income. The Marginal Propensity to Consume on the total consumption bundle, instead of decreasing, increases with an increase in income, and the Marginal Propensity to Consume for food also increases with an increase in income. These findings contradict most of the previous studies. These results indicate that at low income levels, the socio-demographic factors dominate income in influencing consumption behaviour. It was also found that large households have significantly higher Marginal Propensity to Consume than the small households. Similarly, as urban households have a higher Marginal Propensity to Consume than the rural households, it follows that rural households have a higher Marginal Propensity to Save. However, there is no significant difference between the Marginal Propensity to Consume by farmers and labourers on the one hand and by civil servants and traders on the other. The subsistence level of total expenditure by the average household in Kabupaten Kerinci in 1977 was Rp 35,000/Capita/Year (about 47 per cent of the average per capita income). This amount is slightly lower than the poverty line. The subsistence level across the different types of households ranged from Rp 22,000 to Rp 71,000/Capita/Year. For the majority of household types, the subsistence level is higher than the poverty line. There is a tendency that subsistence expenditure increases with an increase in income. The demand of the consumption bundle is quite sensitive to income and food price changes, but relatively less sensitive to the changing of prices of other consumption goods. Apart from the above results, this study also presents income elasticities, total expenditure elasticities, own and cross price elasticities of every type of expenditure for every type of household. The saving sensitivity for every type of household were also examined. The studies of consumption behaviour like this could give significant contributions to some areas related to economic development, among these are: (i) identifying the stage of economic development, (ii) identifying economically critical regions, (iii) detection of consumption saving decisions, (iv) demand forecasting, (v) detection of the effect of prices on the*welfare of consumers, (vi) guidance and evaluation tools of income distribution policy, (vii) price policy, and many others. Finally this study suggests the direction of further research which would cover: (i) widening the data base and estimation framework, (ii) extending the theoretical model, and (iii) incorporating the results into economy-wide models