CUSTO DE VIDA

In this research cross-sectional data refering to ?per capita? monthly income and ?per capita? monthly consumption or expenditures, for several categories of food were utilized. The data were collected through questionaires applied to a random sample of 255 family units in the urban area of Piracicaba during the month of July, 1971. The general objective was to study the functional relation-ships between income and food consumption. Three types of elasticity coefficients were calculated: income-elasticity, expenditure-income elasticity and expenditure-total expenditure elasticity. Four methods of estimation of the elasticities were tested and compared. Two of them were intended to eliminate estimation bias, caused by utilization of the least squares method on data with measurement errors in the independent variable. The first was the classic method, wich assumes that either the value of the ratios between the variances of errors of observation of the two variables or the value of the variance of the error of observation of ?per capita? monthly income is know. The second was the method of grouping of data classified by the independent variable. The third method consisted of adjusting the regressions to the average consumption and average ?per capita? monthly expenditures of six groups classified by ?per capita? monthly expenditure. The double log, semi-log and inverse-log functions were tested through this method. The fourth method aimed to obtain variable elasticity coefficients according to the levels of monthly ?per capita? income or expenditure utilizing a double log function. This method consisted of adjusting a poligonal to the individual data, each segment corresponding to a monthly ?per capita? income or expenditure stratum. The model, through ?t? tests, permitted to determine whether the elasticity differences between one stratum and another were statistically significant.