This paper extends the nonparametric methods developed by Samuelson (1948), Houthakker (1950), Afriat (1973), Diewert (1973) and Varian (1982, 1983) to latently separable models. It presents necessary and sufficient empirical conditions under which data on the market behaviour of a price-taking consumer, and a hypothesised allocation across latent groups are nonparametrically consistentwith latent separability (Gorman (1968, 1978), Blundell and Robin (2000)). It considers homothetic latent separability and weak separability as special cases.
Centre for microdata methods and practice