This content was first published on the Zillow research blog on November 22, 2019.

• Over the next 20 years, more than a quarter (27.4 percent) of the nation’s currently owner-occupied homes are likely to hit the market as their current owners pass away or otherwise vacate their homes.

• Places likely to be most impacted by this upcoming Silver Tsunami include both retirement hubs (Miami, Orlando, Tampa and Tucson) and regions where young residents have left (Cleveland, Dayton, Knoxville and Pittsburgh). The impact of the Silver Tsunami is also likely to vary greatly across different areas within metros.

• The places likely to be least impacted include those with vibrant economies featuring fast growth and affordable housing that act as magnets for younger residents (Atlanta, Austin, Dallas and Houston).

• Housing released by the Silver Tsunami will provide a substantial and sustained boost to housing supply, comparable in magnitude to the fluctuations that new home construction experienced in the 2000s boom-bust cycle.

• It seems likely that, in the coming two decades, the construction industry will need to place a greater focus on updating existing properties, in addition to simply building new homes.

The massive Baby Boomer generation has already begun aging into retirement, and will begin passing away in large numbers in coming decades – releasing a flood of currently owner-occupied homes that could hit the market. That could help end the last few years’ inventory drought, as well as a more fundamental shortage of homes in certain places.

This Silver Tsunami of homes coming to market could be a good substitute for new home construction, which has been in short supply for the past decade in large part because of difficult-to-overcome challenges faced by builders.

Currently, 33.9 percent of owner-occupied U.S. homes are owned by residents aged 60 or older, and 55.2 percent by residents aged 50 or older. As these households age and begin vacating housing, that could represent upwards of 20 million homes hitting the market through the mid-2030s.

But while virtually all areas will feel the effects to some degree, this wave won’t hit all at once and won’t strike all markets equally. Certain markets will be more impacted than others, as will certain kinds of areas within a given market.

The Silver Tsunami will rise progressively in the 2020s and 2030s

An estimate based on mortality data suggests that the Silver Tsunami will hit in earnest as the number of seniors aged 60 or older who pass away each year rises during the 2020s and 2030s. An alternate estimate using data on recent homeownership trends by age and cohort suggests that homes released by 2027 will comprise about one-in-eight (12 percent) of today’s owner-occupied housing stock,[1] and that by 2037 that share will more than double to 27.4 percent.

Older communities and retirement hubs hit hardest; Affordable, growing cities attractive to the young least-affected

The Silver Tsunami will strike nationwide, impacting between one-fifth and one-third of the current owner-occupied housing stock in every metro analyzed.

Well-known retirement destinations, including Miami, Orlando, Tampa and Tucson, will experience the most housing turnover in the wake of the Silver Tsunami. If the number of future retirees choosing to make these places home during their golden years fails to match generations past and local housing demand fades, these areas may end up with excess housing.

Download the metro area statistics here

Regions including Cleveland, Dayton, Knoxville and Pittsburgh are also more likely to see bigger effects from the Silver Tsunami. Younger residents have tended to leave these areas in recent decades, in many cases pursuing better job opportunities elsewhere, leaving older generations to make up a larger share of those who remain.

And some regions will see smaller-than-typical shares of the owner-occupied housing stock released. These include Salt Lake City (where fewer than one-in-five householders (18%) is 55 or older, compared to 28% nationwide), Atlanta, Austin, Dallas and Houston – all of which tend to be fast-growing and relatively affordable. Because they are both economically vibrant and largely willing and able to expand their developed footprints, these metros offer an affordable alternative to expensive coastal cities. The expensive coastal cities are also attractive to many residents, but they impose burdensome housing costs – especially to younger residents just starting out. In contrast, cheap but vibrant areas generally tend to attract the young and have a younger age distribution as a result, reducing their exposure to the Silver Tsunami.

Small, long-exclusive enclaves in many cities will experience rapid turnover

Download the local area statistics as well as their national and metro-specific rankings here

Still, the differences in the share of homes released by seniors between metros are small compared to the differences within them. The most-impacted areas within a given region often contain formally-defined or informal-but-de-facto retirement communities, typically on the fringes of larger metro areas in locations rich with second homes (and former second homes turned primary residence in retirement). Examples include Palm Springs, east of Los Angeles; The Hamptons, at the far eastern end of Long Island in New York; and the island of Nantucket off the coast of Cape Cod, southeast of Boston.

But in many cases, the most-impacted areas appear to be pricey, relatively exclusive enclaves of older residents within otherwise mixed surroundings, including the Upper East Side in Manhattan and the Lafayette-Orinda-Walnut Creek area east of San Francisco. These areas have been more affluent and expensive compared to their surroundings for decades, which has made them less affordable to the young. Younger, more-affordable neighborhoods are likely to have a younger mix of residents and can therefore expect to be less impacted by the Silver Tsunami.

Unlike popular retirement communities that may suffer from a dearth of housing demand down the line, these older, long-affluent enclaves may see more demand pour in from neighboring areas and more residents churn when the Silver Tsunami brings more homes on the market.

Silver Tsunami as substitute for new home construction?

The release of older residents’ homes into the market adds net supply to the market, similar to new construction (most existing home sales involve a seller who is almost simultaneously also a buyer, contributing to both the supply and demand for housing).[2] That is important because new construction has failed to return to its historical production levels in the years following the Great Recession, and home builders face challenges that are likely to endure for a long period. These include construction labor issues and difficulties accessing capital, as well as issues related to the shift in demand towards locations closer to the metropolitan core – where vacant lots are scarce and re-development is more constrained than ever.

New home construction activity skews towards the metropolitan periphery. But much of the existing housing set to be released by seniors in coming decades is better-located to meet growing demand for living closer to the center.

Finally, more homes coming onto the market in high-demand areas – if combined with local changes to land-use policy that allow for more density – could spur more construction of small- and mid-sized multi-unit properties by providing developers more opportunities to acquire lots (and by making it easier to assemble them into larger ones).

How will the Silver Tsunami affect the housing market at large?

In the decade from 2007 to 2017, roughly 730,000 U.S. homes were released into the market each year by seniors aged 60 or older. From 2017 to 2027 and from 2027 to 2037 that number is set to rise to 920,000 and 1.17 million per year, respectively, an addition of about 440,000 homes annually by the second decade.

There is no doubt the Silver Tsunami will substantially boost the supply of housing. But will that be enough to counteract rates of new construction that are currently low in historical terms? New homes peaked at 1.28 million sales per year in 2005 and hit bottom in 2011 at 310,000 new homes sold, a difference of almost 1 million per year (in 2018, it clocked in at 620,000 – double recent lows, but still well off peak levels). A swing of that magnitude is larger than the Silver Tsunami’s boost of 440,000 homes per year.

But because the Silver Tsunami’s boost to the housing supply will be sustained over the course of several decades, rather than the sometimes-sharp year-to-year fluctuations in new home construction, it makes more sense to compare longer time periods. From 2000-2009, spanning the last decade’s housing boom and the early years of the subsequent bust, an average of 900,000 new homes were sold each year. Between 2009 and 2018, spanning the bulk of the bust and most of the sluggish recovery that followed, there were only 450,000 new home sales per year, on average. The difference of 450,000 new home sales per year is comparable in magnitude to the Silver Tsunami’s boost.

The Silver Tsunami will boost the supply of housing on a magnitude comparable to the fluctuations that new home construction experienced in the 2000s boom-bust cycle. Of course, whether that means there will be a glut or a continued shortage of housing depends on how new construction fares and, crucially, on future demand for housing. Whether housing released by the Silver Tsunami is appropriately located, priced and to meet future demand could be an important issue as well.

What seems most likely amid all the uncertainty is that, in the coming two decades, the construction industry will need to place a greater emphasis than before on updating existing properties, in addition to building new ones. The construction industry may be bracing for a tsunami of old-new housing supply that crowds out new development, but renovation could be where the real silver is at.

Notes

[1] The year 2027 was chosen because it is one decade on from the most recent owner-occupied housing number available from the 2017 American Community Survey.

[2] The sale of older residents’ homes as well as newly-built ones also have a disproportionate, positive effect on the volume of home sales, because they set off vacancy chains in which their buyer often turns around to sell their old home, whose buyer does the same. Estimates with respect to new construction indicate that on average every new home built corresponds to 2.43 home sales, i.e. the new home plus 1.43 existing home sales of those moving along the chain. Houses released by seniors are likely to do the same.

Data and methodology:

The number of owner-occupied homes whose residents will be over 60 and pass each year is estimated using a mortality-based approach:

• This approach uses the 2017 1-year American Community Survey (ACS) in conjunction with the 2016 Social Security Administration period life table for the U.S. to select the individual over age 25 with the longest expected remaining lifespan in each owner-occupant household based on age and sex observed in 2017. It then assigns the household the selected individual’s hazard of death each year thereafter, and sums the (weighted) hazards of individuals aged 60 or more in each future year to obtain the national number of currently owner-occupied homes whose resident with the longest expected remaining lifespan will be 60 and is estimated to pass away each year through 2040. The results reported in the first chart, above, are based on a random 10% national sample of households.

• This approach does not estimate the number of homes released into the market by senior residents passing, because it abstracts from all the reasons that households may exit owner-occupancy aside from mortality. Nevertheless, the approach is informative with respect to the timing of the Silver Tsunami.

• For simplicity, this approach abstracts changes over time in mortality rates, taking those from 2016 as fixed, as well as the case in which the individual with the longest expected remaining lifespan in a household is not the last to die. It also abstracts from international migration, and from differences in mortality rates across race and across locations within the U.S. Finally, it is constrained by observed homeownership as of 2017: Households whose residents will be 60 or more and pass away by 2040 only influence the estimate if they were already owner-occupants as of 2017, but not if they would become owner-occupants after 2017. Similarly, exits from owner-occupancy prior to passing by households that were owner-occupants as of 2017 is not reflected in the estimate.

• The number of owner-occupied homes whose residents were 60 and passed each year from 2007 to 2017 cannot be observed directly because deaths cannot be observed in the ACS. Instead, the mortality-based approach is applied to the data observed in the 2007 1-year ACS and again to that observed in the 2017 1-year ACS, and the number of owner-occupied homes whose residents were 60 and estimated to pass is average across those two years and taken as the estimate for the 2007-2017 period.

The number of owner-occupied homes to be released into the market by residents aged 60 or more is estimated with an owner-occupancy retention-based approach:

• This approach assigns every owner-occupant household observed in the 2007 and 2017 1-year editions of the ACS an age based on its youngest member over the age of 25, and uses that age to assign each household to one of the following cohorts “born before 1927”, “born 1927-1936”, “born 1937-1946”, … , “born 1987-1996”. Following a recent study by Fannie Mae, the 2007-2017 retention rate for each cohort is defined and observed as the ratio of the pooled number of owner-occupant households in the cohort as of 2017 and as of 2007. The retention rate reflects households whose members have passed away, as well as those whose members have moved into or out of homeownership and into or out of the region. The future retention rate of every cohort in the 2017-2027 and 2027-2037 periods is then estimated, by assumption, to be the same as that of the correspondingly-aged cohorts circa 2007-2017. For example, the retention rate of the 1957-1966 cohort during the 2017-2027 period is assumed to be the same as the retention rate observed for the 1947-1956 cohort during the 2007-2017 period. To estimate the number of homes released by seniors in the 2017-2027 and 2027-2037 periods, the number of owner-occupant households in each cohort is observed in 2017 for the relevant geography (nation, metro area or public-use microdata area) and then projected forward using retention rates estimated nationally, because retention rates for finer geographies increasingly reflect mobility of households into and out of region, which is not of interest here. The number of homes not retained is then summed up across all cohort-by-period cells in which the cohort is aged 60 or more at the start of each period.

• This approach, too, abstracts from changes over time in mortality rates, taking those from 2016 as fixed. It is also muddied by mobility into and out of the U.S., and by the Great Recession, which occurred during the 2007-2017 period in which retention rates are observed—to the extent that older cohorts lost their homes. Having said that, members of cohorts over 60 at the time were less likely to lose their homes than members of younger cohorts who were more likely to have purchased during the years preceding the Great Recession.

Both approaches are conservative in that they assign households’ death hazards (mortality-based approach) or cohort (owner-occupancy retention-based approach) based on the member with the longest expected remaining lifespan or the youngest member (over age 25), respectively. Both these assignments are both likely to influence the resulting estimates. Inasmuch as the householder tends to be the oldest member of the household, these choices are likely to reduce estimates. Whether a home is in fact likelier have its residents pass and/or be released into the market when the oldest adult—typically the householder—or the youngest adult—typically the spouse—passes, is unclear. Using the householder’s age is no more and no less correct than the alternative assignment used here. The matter becomes more complex when the youngest adult is not the householder or their spouse, but a member of another generation. In such cases, the family relationship between household members and their choices after inheritance, for example, are also not clear.

Metro-level estimates of the cumulative number of owner-occupied homes whose residents will be over 60 and pass by 2037 using the mortality-based approach and metro-level estimates of the cumulative number homes released into the market by residents aged 60 or more by 2037 using the owner-occupancy retention-based approach differ. The former tends to be lower than the latter because households often exit owner-occupancy prior to passing. However, the estimates are very similar in terms of the ranking of metro areas, having a correlation of 0.87.

The metro areas referenced in estimates and maps are Combined Statistical Areas (CSAs) where applicable, and Core-Based Statistical Areas (CBSAs) elsewhere. The list of metro areas is limited to those with population greater than 1 million as of the 2010 Census.

New home sale numbers are drawn from the U.S. Census’ new residential sales series (C25), obtained via Moody’s economy.com service.