Joint Choice of Tenure and Dwelling Type: A Multinomial Logit Analysis for the City of Chongju

Abstract

This paper explores the plausible variables that influence choices in the Chongju housing market. A joint model of tenure and dwelling type is estimated within the multinomial logit framework. The logit model of housing choice is fitted to the whole sample and to two sub-samples, which are categorised on the basis of neighbourhood quality. The empirical results do not exhibit a uniform pattern over the three sample cases. In the case of the whole sample, housing choice is influenced by the age, educational level and occupation of household head, a dummy of school-age children, and the ratio of housing price to household income. For the households locating in the high-quality neighbourhoods, the age, educational level and occupation of household head, and the housing price to income ratio appear significant. In the second sub-group, which comprises the households choosing a poor-quality neighbourhood, the age and occupation, a dummy of school-age children, and the housing price to income ratio appear to be the responsible variables. The findings suggest that the supply of housing should be diversified to meet the diverse segments of demand in the Chongju housing market. An increasing priority should be placed on the provision of rental housing, whether detached or multiple, rather than owner-occupied apartments.

Japanese Housing Tenure Choice and Welfare Implications after the Revision of the Tenant Protection Law

“…Histogram of three tenure by planned length of time to own house 12Cho (1997) analyzed four tenure choices.…”

Housing prices and price endogeneity in tenure and dwelling type choice models

“…Most of these studies relied on the workhorse of MNL models when it comes to estimating the joint choice of tenure and dwelling type/ location, without careful consideration of the error structure as to whether assumptions on the distribution of error terms were satisfied. Examples were Cho (1997) and Skaburskis (1999) who both investigated the joint choice of dwelling and tenure type. Although the rationale for the use of a genuine NL model was explicitly recognised by Cho and Skaburskis, they used MNL models due to the limitation of employed software and estimation technique at the time.…”

“…Numerous studies have focused on two dimensions of housing choice, with the most common being tenure choice (Cho, 1997;Rapaport, 1997;Skaburskis, 1999;Tu and Goldfinch, 1996). Most of these studies relied on the workhorse of MNL models when it comes to estimating the joint choice of tenure and dwelling type/ location, without careful consideration of the error structure as to whether assumptions on the distribution of error terms were satisfied.…”

<b>A Block-level Estimation of Residential Characteristics </b><b>Using Survey and Spatial Microdata </b>

“…(1) In general, dwelling type and size are strongly associated with demographic attributes. Studies on housing demand found that the aged, householders who are professionals, and households with school-age children are not inclined to choose rented apartment housing 34) and that an increase in household size, age, or income is associated with demand for larger dwellings. 35) However, this trend is not uniform across the city.…”