Economists are engaged in a debate about whether the geography of housing demand is changing as a result of the COVID-19 crisis and the recession of 2020. Count me firmly in the camp that sees not only ongoing evidence of this change but also expects it to continue to a degree in the post-vaccine economy.

These changes are a continuing evolution of the market due to pre-2020 factors. Declining housing affordability in high-density markets were due to limited resale inventory and insufficient levels of home building. The impact of COVID-19 on prospective home buyer concerns and preferences have accelerated these changes. For example, the direct impact of the virus on housing demand is a competitive disadvantage for residences that require elevators or utilize a great deal of common space.

As a result of these factors, some buyers (not all) are voting with their feet by seeking housing farther out from urban cores due to concerns over residential density, wariness of city governance and tax burdens, and demands for more space. The limits of an acceptable commute have also been extended by an expected permanent change for telecommuting, with some research suggesting that roughly 20% of workers will be allowed to work online for the majority of their work hours. And employees who can telework permanently—admittedly representing less than 10% of all jobs—can effectively choose to live anywhere in the U.S. with a reliable internet connection.

The evidence of this suburban shift is already in view. Rents are falling in low-affordability, high-density markets like New York and San Francisco. And while the evidence is mixed in the resale market, the pass-through impact of declining rental demand in urban core areas should be causing home construction to expand at faster rates in low-density markets to accommodate the change in housing demand.

This is exactly what the data indicates is happening. According to the second quarter NAHB Home Building Geography Index (HBGI), the economic region posting the worst performance with respect to single-family construction was the set of large metro urban core counties, which recorded an almost 17% year-over-year decline. Large metro suburbs and exurbs fared better, with declines of just over and just under 9%, respectively.

While the exurbs were the best performing areas within large metros, small metro geographies registered smaller declines. Small metro core counties (cities with less than 1 million people) declined by only 6% year over year. And the best performing region was small metro suburbs, which had an estimated growth rate of almost 3%, the only region in the HBGI posting a year-over-year gain for the second quarter.

This isn’t just a single-family trend. Multifamily construction fared better than expected in the first half of 2020, in part due to construction in outlying areas. While large metro areas host most apartment construction, the market share for multifamily permits for large metro cores and suburbs fell from 67% to 66% in the second quarter alone, a significant change as market shares tend to be fairly stable over time.

A vaccine will slow these changes and return some strength to urban core rental demand, but this recession will leave lasting impacts on where Americans live and work. Single-family housing, townhouses, and garden-style multifamily will prove to be more resilient during a public health crisis. In turn, these changes are boosting demand for new construction in lower-density markets.