MUCH ADO ABOUT 2 - RECONSIDERING RETRANSFORMATION AND THE 2-PART MODEL IN HEALTH ECONOMETRICS

In health economics applications involving outcomes (y) and covariates (x), it is often the case that the central inferential problems of in terest involve E[y\x] and its associated partial effects or elasticiti es. Many such outcomes have two fundamental statistical properties: y greater than or equal to 0; and the outcome y = 0 is observed with suf ficient frequency that the zeros cannot be ignored econometrically. Th is paper (1) describes circumstances where the standard two-part model with homoskedastic retransformation will fail to provide consistent i nferences about important policy parameters; and (2) demonstrates some alternative approaches that are likely to prove helpful in applicatio ns. (C) 1998 Elsevier Science B.V. All rights reserved.