A system-wide approach to demand analysis for rental housing characteristics in Honolulu SMSA and effects of demographic variables on housing characteristics

Abstract

The central focus of this study is to investigate household demand for Lancastrian housing characteristics in a system-wide approach in the Honolulu SMSA, and to investigate the effect of demographic difference across households on the demand for rental housing characteristics. Interview data from seventeen hundred and forty households are used in an effort to examine the importance of demographic variables in influencing housing consumption patterns. Total rental expenditure is allocated among seventeen housing attributes by applying the hedonic technique. These housing attributes are aggregated into three basic housing characteristics: interior space, structure quality, and neighborhood condition. Three different demand models are utilized. One model, the linear expenditure system, first applied by Stone(l954), assumes additive preferences and is considered in light of the limitations associated with an additive framework. Another model, the translog demand system, introduced by Christensen, Jorgenson, and Lau(1975), assumes non-additive preferences and is considered in light of its flexibility in capturing cross price effects. The other model, the almost ideal demand system, introduced by Deaton and Muellbauer(l980), also assumes non-additive preferences and is applied in light of its ease of estimation in addition to non-additive structure. Homotheticity and additivity of households preference underlying housing consumption are rejected in the translog model. Homogeneity and symmetry in consumption of housing characteristics are also rejected. The almost ideal demand system is preferable to the translog model in terms of economic interpretation and computing cost. In general, the non-additive system yields a better explanation of household consuming behavior of housing services in the Honolulu rental housing market. Five demographic variables reflecting household composition and size are incorporated into the three demand systems. The inclusion of household size, age, race, and sex of household head, and the level of education of head reveals that such demographic variables play a crucial role in shaping household preferences for housing services and affect the recognized prices of housing characteristics.