In this study we estimated empirically the impact of U.S.D.A funded school feeding program on household food expenditures and food behavior in the case of Indonesia. The focus was on children associated food expenditures an aspect of food behavior which is particularly relevant for households' day-to-day decisions affecting their food expenses and consumption. The basic data indicated a positive correlation (p=.65) between household whose children receiving milk at school and household food expenditure: participating households have higher child-related food expenditures, more milk consumption at home, and more percentage of food expenditures to income.
We used a reduced-form model of household pattern of expenditure, which controls for relevant household and location characteristics, to estimate the impact of program components to household food expenditure and food behavior. The underlying structural equations treat school feeding program's outputs (milk ration and advocacy) as an input, together with relevant household characteristics, in the household's decision-making process of food consumption. The effects of milk consumption by the children operate through (at least) three mechanisms: sharing of information in peer group, influencing household meals provider with their preference, and broadened household food preference. The magnitude of the children food preference effect was found to be similar to that of program directed to the parents.
Differences between two groups of sample are found for most household expenditures, but to be quite the same for most nominal characteristics such as education, location of living, and magnitude of information. With few exceptions, there is no difference in non-food expenditure between these groups.
The increasing demand for milk, both in experimental group and control group, suggest the potential of a large Indonesian market for dairy products in the future. But it may leave behind domestic supplier (dairy farmers) in the competition because of free market policy, intensive and integrated marketing efforts by foreign supplier whose efforts are integrated in the program. Consistent and integrated policies on agricultural products are the conditions needed.
The differences in consumption patterns between these two groups, and among areas groups imply that development in aggregate consumption patterns in Indonesia will be affected not only by aggregate income growth but also by changes in the distribution and channeling system of information. Using children as channel of promoting behavior change in food pattern is promising.