An empirical analysis of household choice among motor vehicles

You are here: Homepage > Search

An empirical analysis of household choice among motor vehicles

Manski, Charles F. / Sherman, Leonard

Abstract This paper presents an econometric model explaining the make, model and vintage composition of individual household motor vehicle holdings. We view the household as yearly evaluating its current holdings of vehicles and updating these as desired. The utility of any vehicle, or vehicle pair, to a household is assumed a function of vehicle seating capacity, luggage capacity, weight, acceleration time, noise level, scrappage probability, price, operating cost and of a search-transactions cost associated with entering the vehicle market. Household size, age, education, income, number of workers and residential location condition the utility function. A multinomial logit model probabilistically describes each household's choice among vehicle alternatives. The empirical analysis is based on a random sample of households drawn from a nationwide rotating consumer panel. Two models are estimated, one explaining vehicle choices in households holding a single vehicle, the other explaining the composition of holdings in two-vehicle households. Among the many empirical findings, one prominent result is that each household has an optimal vehicle seating capacity which varies directly with household size. We find that most aspects of vehicle performances have little effect on choices but, counter-intuitively, sluggish vehicles appear strongly preferred to quick ones. Vehicle price, operating costs and transactions costs are all important determinants of vehicle utility but the influence of price and operating cost varies considerably among socioeconomic and demographic groupings.