The unprecedented surge in household formation and the homeownership rate in the second quarter is not a sign that Americans are fleeing rental apartments in the city to buy homes in the suburbs. Rather, issues with data collection due to the coronavirus explain the rise. Indeed, given the collapse in home sales seen from March to May, we expect the actual homeownership rate just held its ground.
- The unprecedented surge in household formation and the homeownership rate in the second quarter is not a sign that Americans are fleeing rental apartments in the city to buy homes in the suburbs. Rather, issues with data collection due to the coronavirus explain the rise. Indeed, given the collapse in home sales seen from March to May, we expect the actual homeownership rate just held its ground.
- The second quarter Housing Vacancies and Homeownership (HVS) survey contained several incredible results. The most striking was a 2.9 percentage point quarter-on-quarter rise in the seasonally-adjusted homeownership rate to 68.2%. That was four times the previous largest gain of 0.7 percentage points. The total share of vacant homes also plummeted, and as a result an estimated 4.3 million households formed in the year to the second quarter, also comfortably the highest since records began in 1965.
- It is tempting to conclude that the coronavirus led to the largest ever shift in the housing market, perhaps by persuading people to abandon rental apartments in cities to buy homes in the suburbs. But the less exciting conclusion is that the virus has instead made the survey unreliable.
- As a result of lockdowns, the Census Bureau was unable to carry out in-person interviews, and so instead relied entirely on telephone interviews. An obvious problem with this solution is that contacting people in vacant homes by phone is harder than contacting those in occupied homes. The survey is not designed to account for rapid changes in non-response, and as a result the estimate of the total vacancy rate plummeted, and the total number of households therefore ballooned.
- Admittedly, as the Census Bureau pointed out, it is possible that the drop in vacancy reflects the real impact of COVID on the housing market. But a quick check against alternative data sources rules that out. For example, according to the HVS, the number of vacant multifamily units for rent saw a sharp drop in the second quarter. But data from REIS shows an increase. (See Chart 1.) Given no one could move into a new apartment in the second quarter, the REIS data looks far more plausible.
- A further implication of moving to telephone interviews is that rental households are likely to be harder to get hold of than owners. That probably explains the surge in the homeownership rate. After all, it is difficult to square booming homeownership with a collapse in home sales.
- Admittedly, it can take time for changes in home sales to feed through to homeownership. But the surge is inconsistent with sales over the past couple of quarters. (See Chart 2.) Given the low level of transactions, a more plausible outcome is a flat homeownership rate. Once the Census Bureau is able to return to normal operations, we expect the rate will drop back to its previous level of around 65%.
Chart 1: Estimates of Vacant Apts for Rent (Millions)
Chart 2: Homeownership Rate & Home Sales (% y/y)
Sources: Census Bureau, REIS
Sources: NAR, Census Bureau
Matthew Pointon, Property Economist, email@example.com