03262nam a2200181 a 450000100080000000500110000800800410001910000200006024501080008026000160018830000100020450000510021452027300026565000200299565000250301565300180304065300220305810916532007-12-12       1979    bl uuuu  m   00u1 u    #d1 aTEIXEIRA, S. M.  aIncome, family composition and social factors as variables in an expenditure modelbthe case of Brazil.  a1979.c1979  a158p.  aTese  (Doutorado) - Purdue: Purdue University,  aThis study is concerned with the estimation of the effects of socio-economic variables on household consumption. The procedure used incorporates household composition into a household utility function.The effects of household composition and social factors are then estimated by adult equivalent scales with the demand equations derived from the utility maximization problem subject to income constraints. The utility function incorporates the socio-economic variables namely income, family composition, formal education of the family head and house-wife, and employment status of the head and housewife, resulting in a linear expenditure model to be estimated. Due to nonlinearities in the coefficients, an interative procedure is  employed to estimate the adult equivalent parameter which is incorporated into the final specification of the model. The data comprise a sample of 2,542 urban households from three capital cities in Brazil: Rio de Janeiro (1,000 families), Porto Alegre (726 families), and Recife ( 726 families), obtained from the ECIEL Survey by the Brookings Institution in connection with a Brazilian research foundation- Getulio Vargas Foundation (FGV) performed in 1968. The model is estimated for a group of 16 comodity categories. The overall results of the estimation appear reasonable, although the coefficients for the adult equivalent scale components for nonfood commodities presented unexpected signs and large standard errors. The income coefficient proved to be highly relevant and estimates of commodity coefficients for the city of Rio de Janeiro resulted in  extremely low values as compared to the other two cities. Further analysis was undertaken for that specific parameter and the large dispersion of the income distribution suggested the deletion of data for five very high income families. The analysis, after deletion of these five observations, presented more consistent results when compared to the other two samples. Both results for the overall sample and the deleted obsrvations are presented. The functional form of the model is based on the work of Buse and Salathe (9) for expenditure on food commodities in the U.S. The approach in this study modifies and extends their work by changing the age ranges and applying the procedure to other commodity categories. Illustrations are given of the adult equivalent scales using the results of the study for the 16 commodity groups. For the nonfood commodities, the hypothesized functional form the adult equivalent scale generated unexpected results. This was regarded to be due in pate to the unrestricted signs for the coefficients in the equations, and in part to the unreliability of the data for selected commodity categories.  aeconomic policy  aPolítica Econômica  aFamily salary  aSalário família