DOES IT MATTER HOW WE SURVEY DEMAND FOR FOOD - EVIDENCE FROM KENYA AND THE PHILIPPINES

This article examines the reliability of food quantity information collected from expenditure surveys, and finds that measurement errors will very often be systematically correlated with income, such that the response of food intakes to increases in income is seriously overstated. This may be, firstly, because food quantities are not measured independently of income if total expenditures are the proxy used for income, and secondly, because food transfers in the form of guest and hired worker meals are under-recorded. These biases are illustrated using household surveys from Kenya and the Philippines.