A Varied Spatial and Temporal Examination of Vacation Home Rentals and Crime
MetadataShow full metadata
This dissertation examines vacation home rental (VHR) properties in Austin, Texas in relation to residential burglary, substance crimes, and disturbances. The dissertation takes a three-study approach, examining VHRs in Austin with three different units of analysis. The first study uses 2018 data and neighborhoods, operationalized as census block groups (CBGs). CBGs are mutually exclusive regions with non-overlapping boundaries and varied spatial dimensions. The second study assesses result robustness by using CBGs with 2016 data, and by also using egohoods, a method of operationalizing neighborhoods with overlapping boundaries and fixed spatial dimensions. The third study uses months as the unit of analysis with a time-series design to examine VHRs and crime in the city. For the two initial studies, count regression models, social disorganization variables, spatial lag, and geographic analyses are used. For the third study, seasonal autoregressive integrated moving average (ARIMA) models are used to analyze monthly data from November 2014 to December 2019 (n=61). Vacation home rental data are reported and contextualized in different kinds of neighborhoods in the city, and in different manners per study. The most prominent finding is that listing type appears to matter. Room-only rentals were significantly and positively associated with crime, in every model that included them. However, the associations for entire-structure rentals varied by crime type, year, and neighborhood operationalization. Some of the implications are that greater scrutiny should be used to understand renter differences of these properties, as well as the property owners that rent rooms versus entire structures.