Does Household “Crowding” Explain Charlottesville’s Apparent Affordability?
CFRP has performed extensive analysis on housing affordability data for Charlottesville, both over time and in comparison to other jurisdictions. We have repeatedly found evidence that Charlottesville is either average or better-than-average on metrics of affordability and change in affordability.
You can see our earlier analysis using Census data in our first-ever presentation from over a year ago here; you can see our work on the affordability trend here; you can see fresher wage-based data here; you can see how Charlottesville stacks up on the National Association of Homebuilders’ Housing Opportunity index here. To assure the robustness of our analysis, we continue to try to poke holes in our own case, something that we think distinguishes our approach from that of pro-upzoning activists. As part of that effort, we wondered if Charlottesville’s apparent affordability at the level of household statistics might be flattered by households “combining” or accepting more crowded accomodation in order to achieve affordability. That is, does some meaningful fraction of Charlottesville households manage to afford a residence only by making the household bigger and/or the residence smaller?
Fortunately, the Census’ American Community Survey looks at the number of residents per room for both owner and renter households. The informal definition of “crowding” tends to be “above 1.0 occupants per room”. That definition, though, appears stale in light of 21st century housing patterns in most of the US. The percentage of renter households at above 1.0 barely reaches mid single-digits for country as a whole. We instead used the the Census’ tabulation cut-off of 0.5 occupants of our room as our metric of interest.
We find that deliberate “crowding” as an affordability strategy cannot explain Charlottesville’s relatively good performance on housing affordability metrics. There is virtually zero crowding among owner-occupied households, so we focused on renter households alone. Charlottesville has an unusually low level of crowding. Among US counties (recall: Charlottesville, as independent city, is tracked as a county by most US government agencies), Charlottesville’s percentage of rental housing with more that 0.5 occupants per room is at the 18th percentile. And while it has been low for at least the last decade, it has actually managed to avoid the trend of increased crowding we see in other cities. We included Gallatin (Bozeman) and Travis (Austin) because they are two cities that have undergone a housing boom that has challenged affordability. We included Santa Clara County (San Jose/Silicon Valley) as an example of a longtime highly expensive area with notoriously restrictive land-use policies. Finally, we put in Hennepin County (Minneapolis) and Multnomah County (Portland) because these two places have been adduced by RHI as examples to emulate.
Another way to look at the data is to consider directly the change in crowding percentage from 2009 to 2020 (the latest data we have). A city experiencing a housing affordability squeeze should see an increase in crowding. The US has gone from 40.1% to 43.1% of rental households above our “crowding threshold”. Charlottesville is at the 28th percentile for change in crowding percentage. Our city again scores well better than average.
Data on rental housing crowding does not by itself establish Charlottesville’s affordability or lack thereof. We have shown elsewhere ample evidence that is highly suggestive that Charlottesville does not suffer an unusual affordability problem at or near the median of the housing market and labor market. What looking at the crowding data does accomplish, however, is to rule out another possible explanation of how the city might present a mere mirage of affordability. We’re poking… but still no holes.
Appendix: Rental Crowding By Census Tract
Below we map the rental crowding percentage in 2020 (latest data) for each census tract in the city. You can see that even the census tract with the highest percentage (tract 2.02. roughly equivalent to 10th & Page and West Main areas) of “crowded rental households” reaches a level only equivalent to 60%th percentile of US counties.