What do we learn from recall consumption data?

In this paper, we use two complementary Italian data sources (the 1995 ISTAT and Bank of Italy household surveys) to generate household specific nondurable expenditure in the Bank of Italy sample that contains relatively high-quality income data. We show that food expenditure data are of comparable quality and informational content across the two surveys, once we properly account for heaping, rounding, and time averaging. We therefore depart from standard practice and rely on the estimation of an inverse Engel curve on ISTAT data to impute nondurable expenditure to Bank of Italy observations, and show how we can use these estimates to analyze consumption age profiles conditional on demographics. Our key result is that predictions based on a standard set of demographic and socioeconomic indicators are quite different from predictions that also condition on simulated food consumption, in the sense that their age profile is less in line with the implications of the standard consumer intertemporal optimization problem.