Using China’s urban household-level survey data from 1992 and 1998, we study household
food-away-from-home (FAFH) expenditure across the two time periods and across regions. We
use a popular parametric linear specification and a newly developed nonparametric estimation
method (with mixed categorical and continuous variables) to estimate the FAFH expenditure
function. The nonparametric model shows that the income elasticities have increased from 1992
to 1998, while the parametric model suggests the contrary. The goodness-of-fit analysis, a model
specification test, and in-sample analysis all suggest that the nonparametric method gives better
description of the data than the parametric approach. The nonparametric estimation results also
reveal other interesting FAFH consumption patterns which are not detected by the parametric
method