Virginia Tech’s Yang Zhang has an interesting manuscript in this issue of the Journal of Planning Education and Research on housing choice and risk perception.
Zhang, Y. 2010. Residential housing choice in a multihazard environment: Implications for natural hazards mitigation and community environmental justice. Journal of Planning Education and Research. doi:10.1177/0739456X10381386.
This study examines residents’ housing decisions in a multihazard environment (flood, hurricane, and toxic chemicals). The combination of a hedonic price model and a household survey conducted in Harris County, Texas, indicates that flood hazard was an important factor in households’ housing decisions, but not hurricane and toxic materials hazards. The results have further suggested that home buyers were poorly informed about hurricane and toxic materials hazards when housing decisions were made and came to understand their vulnerability only while living in hazard-prone neighborhoods. Moreover, it appears that low-income households and minorities were the least-informed group in the housing market. These findings provide implications for using hazard information programs to promote community resilience to natural hazards and community environmental justice.
There are some things you can quibble with here: for some reason, the study includes a poorly defined access measure (perceived importance of accessibility)–which I can’t imagine survey respondents answering in a consistent way—and little else about housing characteristics in a hedonic specification. That leaves the study open to criticism about latent variables. Is lack of access (or having access) really a deal-breaker in this Texas County? How much difference in supply does one have to choose from in the housing markets, no matter how much difference there might be in preference?
Nonetheless, the idea is a good one, and his point is more about the difficulties of housing choice in markets about multiple hazards, at least a couple of which are low-probability hazards, when information is not cost-free.
The problem with leaving out the rest of what would be a fuller specification is that we could also explain differences in willingness to accept housing in hazard-prone area by other factors, such as relative school quality. Or a preference ordering of hazards, based on information preferences (or not): so some aspects of air quality are hyped over and over, but information about toxic events is harder to come by and the events are low-profile and infrequent, so households locate according to one set of information but not the other—and even risk people can’t seem to agree on which is the more important source of risk: locating near small but chronic hazards or near low-probability larger hazards.