Exogenous factors of food expenditure variations: A microdata analysis with restricted regression

• Main Authors: De Leon, Madella S., Dimaculangan, Francis Leo L., Giron, Frances Maureen R., Lucion, Angelica Faye M.
• Format: text
• Language: English
• Published: Animo Repository 2005
• Subjects: Food consumption > Philippines; Cost and standard of living; Households > Philippines; Economics
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/14343
• Institution: De La Salle University

This is an empirical study which deals with the variation in food consumption among Philippine households in both rural and urban areas using the 1997 Family Income and Expenditure Survey (FIES) conducted by the National Statistics Office. This detailed survey data contains several food categories (cereals, meat, fish, etc.) disaggregated into more specific types (rice, chicken, fresh fish, etc.) and further classified as being in cash or in kind. The advantage of addressing the effects of several factors upon certain food groups is to better determine the focus or volume of consumption and purchasing decisions of households. Budget shares of food categories were used as dependent variables and each is being referred to as the ith food commodity consumed. A sample set of households with non-consuming members was compared with a set of all consuming ones to determine if differences in statistical measures exist. The behavior of household food expenditures with respect to increases in income was evaluated. Most importantly, a generous mix of socio-demographic and dichotomous variables enhanced the array of predictors. And to further supplement the regression model, prices of some selected food items were incorporated via the provincial consumer price index (CPI). Several runs and stages of regression were performed to obtain relevant comparisons. A descriptive analysis was also utilized to provide added insights on the level of expenditures. The regression process was done by employing OLS and Full Information Maximum Likelihood (FIML) methods. One remarkable feature of this study was the inclusion of restrictions to property analyze demand systems. Household food expenditures were found to be strongly influenced by income and prices. However, this paper's attempt to test the extent of significance of other household attributes in terms of food consumption and preferences had notably achieved positive outcomes as well.