4 - A Quasi-Loglinear Model of Neighborhood Choice

Abstract

This chapter reviews a quasi-loglinear model of neighborhood choice. The price of a standardized bundle of housing services varies from one neighborhood to the next, as do the levels of neighborhood goods; households select the neighborhood that maximizes overall utility, subject to their budget constraints. Neighborhoods are not viewed qua neighborhoods but as bundles of characteristics, and households moving to them are seen as revealing preferences for these characteristics. The small number of neighborhood types in the study necessarily limits the number and quality of inferences that can be made concerning taste for neighborhood. The chapter presents an empirical technique involving the distribution of households, one stratum at a time, across neighborhoods, where neighborhoods are positioned in a multiway table of neighborhood variables. The tables are incomplete in the sense that only some combinations of neighborhood categories are represented by actual neighborhoods and have counts in them. The quasi-loglinear model enables the assessment of the effects of the explanatory variables on the individual neighborhood variables and to determination of the interrelationships among the response variables.