Abstract:
We consider the possibility that household demographic variables are measured with error. Such errors will arise because income and consumption surveys measure the household's structure at a point-in-time, whereas the demographic composition of the household is constantly evolving over the survey period. We construct and estimate sharp bounds which suggest that the degree of these measurement errors is nontrivial. We also provide evidence that these errors may have important ramifications for recent work on the identification of economies of scale within households.