Discrete choice modeling and demand estimation for diapers (in Russian)

The paper examines the demand structure for diapers in one of Russian cities with the use of a discrete choice model with random coefficients in consumer utilities. As a first step, logit and nested logit specifications are estimated. The market-level information for sales volumes, prices and other product characteristics is used. An unbalanced bimonthly panel for the period from 2008 to 2011 is available. At the second stage, the full model adapted from Berry, Levinsohn & Pakes (1995) and Nevo (2001) is estimated. Additional features are added to the model to capture the market specifics.