The American Dream, a phrase coined by author James Truslow Adams in his Depression-era bestseller The Epic of America, hasn’t always been marked by economic achievement packaged in the form of a rags-to-riches story. Adams envisioned a “dream of a better, richer and happier life for all” and “a dream of social order in which each man and each woman shall be able to attain to the fullest stature of which they are innately capable.” But for better or worse, the American Dream has become synonymous with achieving material wealth and upward mobility, manifest for many in the form of homeownership.

The pursuit of homeownership is deeply embedded in the American psyche, and for good reason: homeownership has been one of the primary channels for wealth accumulation in the US. According to the Federal Reserve's latest Survey of Consumer Finances, the median net worth for US homeowners was 40 times more than that of renters: $255,000 for homeowners vs. $6,300 for renters as of 2019.¹

While demographic and socioeconomic characteristics explain much of the housing tenure choice, we show that the transition from rentership to ownership in the US is based on one primary life cycle trend in particular: having children. The decision to have children is based on many factors that tend to follow distinct patterns within a country, largely based on cultural norms and national policies that run the gamut from federal support for subsidised childcare to the availability of paid maternity/paternity leave.

As a result, we focus on childbirth patterns and their implications for single-family housing demand in the US, which allow us to separate a secular decline in birthrates globally from country-specific policy decisions that can affect household formation and the decision to have children. We note that a subset of all Millennials – those aged 25-342 – may comprise the largest 10-year cohort of the population today at 14% (Chart 1), but delays in marriage and lower fertility rates suggest that as this group ages into the 35-44 cohort over the next decade, they will have significantly fewer children than prior generations, with births likely to be below the replacement fertility rate.³

In the subsequent sections, we model cumulative probabilities that women aged 25-34 today give birth in ten years’ time based on recent childbearing patterns and apply these probabilities to the estimated future female population. We then categorise these women (both with and without children) as belonging to one of ten household types, each of which has an associated probability of occupying a single-family home, thereby enabling us to derive an estimate of the demand for single family housing in the future. We conclude that even with the broad-based decline in female fertility rates, the sheer size of the Millennial population implies a significant increase in the demand for single-family homes over the next decade. However, the demand largely will come from childless households, which has implications for the amenities and location of single-family housing (both owned and rented) going forward. We note that our analysis focuses on US households with 35–44-year-old females, which excludes single male householders and other households without women in this age cohort. This is not to ignore a relevant component of single-family occupancy, but to highlight the importance of children and fertility on housing demand.

Source: Moody's and Census.

All things being equal, US homeownership rates are highest for non-Hispanic whites and generally rise with:
1. Age
2. Income
3. Education
4. Marriage (and rises further for married couples with children) 

Even as these factors don’t fully explain differences in homeownership rates over time4 (rates are also affected by credit/mortgage availability, changing securitisation trends, financial shocks and consumer expectations of future asset prices, among other factors), homeownership patterns have followed a similar pattern over the past several decades. As shown in Chart 2, homeownership rates have historically risen with age, with the largest change in ownership (corresponding to the steepest part of the homeownership “curve”) occurring as young adults age from 25-34 to 35-44. (Note that homeownership rates as of 2020 are lower than historical averages for all age cohorts—particularly young adults—reflecting the impact of the global financial crisis, where an estimated 3.7 million homes were lost to foreclosure.⁵)

Source: U.S. Census Bureau.

Fertility rates for women of child-bearing age have been on the decline for decades, not just in the US, but in other developed economies as well, whether due to increasing urbanisation, better job opportunities for women, a lack of affordable childcare or other reasons, including evolving cultural pressures and societal mores.⁶ Whatever the catalysts, the chart below shows aggregate fertility rates across all women of child-bearing age though 2019 (Chart 3), while Chart 4 shows the birth rate in the US by mother’s age. In 2007, the fraction of US women aged 20-24 giving birth was higher than the fraction of women aged 30-34 giving birth; by 2020, the opposite was true. The upshot is that birth rates have fallen the most for younger women (who typically have more children than their older counterparts), reducing the birthrate in aggregate.

In 2020, birth rates in the US declined for women in all age groups 15–44 – a period that largely excludes the impact from Covid-19, which is thought to have further suppressed the birthrate.⁷ The general fertility rate in the US fell 4% in 2020 and stood at 55.8 births per 1,000 women aged 15–44 – a record low. Similarly, the total fertility rate was 1,637.5 births per 1,000 women in 2020 – also down 4% from 2019 and a record low. Birth rates are likely to have continued to fall in 2021, and while any Covid-induced impact should moderate as the disease morphs from a pandemic to something endemic, the structural decline in fertility rates that has been in place for the prior several decades is not expected to reverse in the near term.

Source: The World Bank.
*The number of children that would be born to a woman if she were to live to the end of her childbearing years and bear children in accordance with age-specific fertility rates of the specified year; a proxy for the average number of children per woman.

Source: NCHS/CDC.

Using ACS data,⁸ we model the conditional probability that a woman aged 25-34 today has a child under age 18 in ten years’ time, based on her current child status today.

We use the birth rate experience of women aged 25-44 in 2019 as the basis for our probabilities (noting the number of women who had a child or didn’t have a child in 2019 as well as the age of any additional children) and model each year of age independently. We also use unique probabilities to reflect women’s differing fertility status, noting that the probability a childless woman gave birth in 2019 is different from the probability that a woman of the same age with an older child had an additional child in 2019. The results of these conditional probabilities are shown below (Chart 5) and on Chart 6, where we focus on the presence of under-18 children in ten years’ time, corresponding to scenarios B and D. Because a woman with a child under age 8 will continue to have a child who will be under age 18 in ten years’ time (scenario E) we add these women to our predicted count of women with under-18 children in the future.

Source: ACS 2019 and CBRE IM.

Source: CBRE IM.

While the above conditional probabilities based on 2019 data may appear low – with fewer than half of childless women aged 25-34 expected to have a child over the next decade – the figures are entirely consistent with recent survey data from the Pew Research Center (Charts 7a and 7b) that show 54% of parents who said that they are “not at all likely” to have more children someday, and 44% of non-parents who said that they are either “not too likely” or “not at all likely” to have any children ever.

Source: Pew Research Center
Note: Share of respondents who didn’t offer an answer not shown. Survey of US adults conducted Oct. 18-24, 2021.

In Table 1, we apply these child-bearing probabilities (columns B and C) to the 2019 population of 25-34-year-old women (column A). The product of these figures (columns D and E) is subsequently applied to our forecast for the number of females aged 35-44 in 2029 (Table 2).⁹ Census projections suggest that there will be 3.7% more females aged 35-44 in 2029 than there are females aged 25-34 today due to net international migration alongside a small loss due to deaths – for an increase of 828,000.

Source: CBREIM, ACS and Census.

Source: Census 2017 projections and Census 2019 Population Estimates, ACS 1-year estimates, 2019 and CBREIM.
*Assumes growth in single-year age cohorts over time from census’ projected figures released in 2017. Example: The percentage change in the estimated 25-year-old population in 2019 vs. t he estimated 35-year-old population in 2029 is applied to the known 25-year-old population in 2019 from the Census 2019 Population Estimates release to derive the future 35-year-old population in 2029.

As noted previously, the increase in household income that comes with age (and greater work experience) is one reason that homeownership rates rise with age; the other is family formation and a greater probability of having children as one ages from mid-20s through mid-40s. Whether it’s the need for more living space, a backyard or a better school district, married and unmarried couples with children are significantly more likely to live in single-family homes (whether owned or rented) than their childless counterparts based on ACS 1-year data,¹⁰ as shown in Charts 8a and 8b and Charts 9a and 9b. We partition households by age, the presence or absence of children, and the householder’s relationship status. We focus on the type of housing structure occupied and designate both detached and attached single-family homes as “single-family” vs. all “other” housing structure types, which primarily include apartments and condos. By analysing structure type (and not whether a home is rented or owned) we can determine housing preference independent of income to estimate single-family housing demand. While these age cohorts (25-34 years old and 35-44 years old) are but one subset of the total count of households, it is the jump in single-family housing occupancy as one moves between these two cohorts that is a critical component of the demand for single-family housing.

Source: ACS 1-year estimates

Source: ACS 1-year estimates.

Among householders aged 25-34, single-family home occupancy is:

• 12 percentage points (pp) higher for married couples with children than for childless married couples (74% vs. 62% averaged over 2000, 2005, 2010, 2015 and 2019 data); 
• 10 pp higher for unmarried (but partnered) couples with children vs. childless partnered couples (55% vs. 45%); 
• 13 pp higher for single parents with children vs. childless singles (45% vs. 32%).

For householders aged 35-44, single-family home occupancy is:

• 11 pp higher for married couples with children than for childless married couples (84% vs. 73% averaged over 2000, 2005, 2010, 2015 and 2019 data);
• 8 pp higher for unmarried (but partnered) couples with children vs. childless partnered couples (68% vs. 60%);
• 14 pp higher for single parents with children vs. childless singles (59% vs. 45%).

Because our analysis hinges on the count of future households with and without children (since the propensity to live in a single-family structure vs. other structure types and the propensity to own vs. rent depends on household choices, rather than the choices of single individuals), we break households down into ten different types (five types with children, five types without children) and analyse the share of single-family owners, single-family renters and other (apartment, condo, etc.) owners/renters.

In Table 3, we show households aged 35-44 in 2019, as an example, segmenting the households by partnership status, child status, housing structure type and housing tenure. These are households that may be headed by a man or woman aged 35-44. However, our analysis forecasts the count of women aged 35-44 with and without children under 18 years in 2029, and we note that not all women aged 35-44 will necessarily live in a household headed by someone aged 35-44. Each woman needs to be “placed” into a household type, and while single women aged 35-44 and single female parents aged 35-44, by definition, belong to a 35-44-year-old household (since they are the householder), approximately 22% of females aged 35-44 today reside in households where the household head is younger than 35 or older than 44.

In Table 4, we show the total count of females aged 35-44 in 2019 residing across all household types and ages. Inherent in our analysis is the assumption that women who will be 35-44 years old in 2029 will have partnership and household status that resembles that of women aged 35-44 today (2019). (Unlike the decline in birthrates, there is evidence that women’s household situations have become more stable over time.) As a final step, we apply the share of women aged 35-44 – both with and without children – across owned single-family structures, rented single-family structures and all other housing types and tenures to our projected female population figures in Table 5.

Source: ACS 1-year estimates, 2019
*A “35-44 household: means the householder (head of household) is between the ages of 35-44 years; if a female is not the householder, she can still be part of a household where she is not the head, which is why the count of 35-44 households with females aged 35-44 is a subset of the total count of 35-44 households.
**This household category includes a child but no parent, so for purposes of allocating our projections of females with and without children, we consider this category most relevant for females without children.

Source: ACS 1-year estimates, 2019.

Source: ACS 1-year estimates, 2019.

Are we set to see an increase in the share of 35-44 women who occupy single-family homes in a decade’s time? As shown in the combined tally of both women with and without children in Table 6, the answer is “yes”, but as shown in Charts 10a and 10b where we bifurcate the results by child status, the answer is more nuanced.

Source: ACS, Census and CBRE IM

Prior to the pandemic in 2019, Pew Research estimated that Americans were roughly evenly split between those that would prefer to live somewhere with larger houses that are “farther apart, but schools, stores and restaurants several miles away,” as compared with somewhere with smaller houses that are “closer to each other, but schools, stores and restaurants within walking distance.”¹¹ Among the 53% that preferred the suburbs, families have traditionally made up a disproportionate share of those willing to trade off the costs of longer commutes and greater dependency on cars for the benefits of a larger home and yard plus better schools, among other suburban amenities.

Since the emergence of the Covid-19 pandemic, however, and in response to the same survey question in July 2021, Pew Research estimated that 60% of Americans preferred the more suburban neighborhood descriptor. The sudden and marked shift in reported preferences follows the first decade of increased densification since at least the 1980s,¹² and surely reflects the mood of the time during the pandemic, which at its worst included multi-week shelter-in-place orders across most communities and complete closure of businesses. Many of the pandemic-related hits to urban live-work-play neighborhoods will be temporary, such as restrictions on indoor dining and other indoor activities.

But the rise in prevalence and acceptance of work-from-home as a permanent feature of the employment landscape opens a novel possibility: the ability to untether one’s choice of residence from that of their employer. Moreover, remote workers will need more space with less distractions (including roommates), which is easier to afford in the suburbs. Could remote work be a catalyst to motivate even those without children, who project to form the bulk of net new households over the next decade, to forgo the urban core for the suburbs? Is there any possibility that a heightened preference for the suburbs will stick, and what, if anything, can we glean from migration patterns during the pandemic thus far?

First, it must be recognised that the pandemic has largely accelerated pre-existing US domestic migration trends, rather than motivated new ones. While the pandemic has exacerbated net move-outs from the highest cost coastal markets of New York City and San Francisco in particular, metros with positive in-migration prior to the pandemic have largely maintained and strengthened that in-migration during the pandemic (Figure 1), while those that were losing households largely continued to do so. Figure 1 shows an 84% correlation between the rate of net household migration in the past 20 months (April 2020 to November 2021) vs. the comparable pre-pandemic period (April 2018 to November 2019), as measured by United States Postal Service (USPS) COA forms. The USPS COA data has admitted deficiencies, particularly in accounting for the migration of recent college graduates (who are least likely to file such a form). Nonetheless, the data provide one of the timeliest real-time indicators of migration trends, particularly since March 2020.

Source: CBRE Investment Management and USPS.
Note: Net household migration measured as families and individuals that filled out a USPS COA with the given metro as their destination, less those with the metro as their origin. Rates measured as a percent of pre-existing households in the metro.

Figure 1 is also indicative of the pre-pandemic trends that have recently defined domestic migration: households moving away from higher-cost urban centres to lower-cost sunbelt cities. This is certainly a trend that remote work should only further support, not diminish. But beyond the preference for lower cost cities, what can we say about migration patterns between suburban and urban neighborhoods within these metros?

Source: CBRE Investment Management and USPS.
Note: Net household migration measured as a 12-month trailing average of permanent moves. Permanent moves are used to help control for seasonality in temporary moves, particularly around the school year. They include business, individual and family moves, though in practice business moves are a very small (less than 5% on average) share of total permanent moves. The central city is most often, though not always, synonymous with the metro name. For the 29-county Atlanta metro, for example, “within central city” denotes moves within the Atlanta City limits, while “within metro but outside central city” captures all moves within the Atlanta metro but outside the Atlanta City limits.
* Data aggregated across the top 11 metropolitan destinations for net in-migration from April 2018 to November 2019, which includes Atlanta, Austin, Boise, Charlotte, Dallas/Ft. Worth, Houston, Jacksonville, Las Vegas, Phoenix, Raleigh-Durham, and Riverside.

Source: CBRE Investment Management and USPS.
Note: Net household migration measured as a 12-month trailing average of permanent moves. Permanent moves are used to help control for seasonality in temporary moves, particularly around the school year. They include business, individual and family moves, though in practice business moves are a very small (less than 5% on average) share of total permanent moves. The central city is most often, though not always, synonymous with the metro name. For the 29-county Atlanta metro, for example, “within central city” denotes moves within the Atlanta City limits, while “within metro but outside central city” captures all moves within the Atlanta metro but outside the Atlanta City limits.
* Data aggregated across the top 11 metropolitan origins of net out-migration from April 2018 to November 2019, which includes Boston, Chicago, Los Angeles/Orange County, Minneapolis-St. Paul, the Greater New York Metro, Philadelphia, San Diego, San Francisco/Oakland, San Jose, South Florida and Washington, D.C.

Figures 2 and 3 use the limits of the central city within each metro (e.g., the Atlanta City limits within the 29-county Atlanta metro) to differentiate between net moves within the urban core and closest inner-ring suburbs as compared with outer lying suburbs and exurbs. Figures 2 and 3 aggregate across the top 11 metros by absolute net in-migration and out-migration prior to the pandemic, respectively, largely correlating with sunbelt and gateway markets, to reveal that the trend away from the urban core is not just specific to the largest and most expensive coastal metros. Even in the top (largely sunbelt) destinations for in-migration, there was a persistent net exodus from the central city limits prior and during the pandemic, as measured by USPS permanent COA forms, which increased slightly during the pandemic (Figure 3). Net-positive migration trends to these top sunbelt metros, and the pandemic-related boost to in-migration, was driven entirely by moves outside the central city.

Source: CBRE Investment Management and USPS.
Note: Net household migration measured as a 12-month trailing average of permanent moves. Permanent moves are used to help control for seasonality in temporary moves, particularly around the school year. They include business, individual and family moves, though in practice business moves are a very small (less than 5% on average) share of total permanent moves. The central city is most often, though not always, synonymous with the metro name. For the 29-county Atlanta metro, for example, “within central city” denotes moves within the Atlanta City limits, while “within metro but outside central city” captures all moves within the Atlanta metro but outside the Atlanta City limits.
* Data aggregated across the top 11 metropolitan origins of net out-migration from April 2018 to November 2019, which includes Boston, Chicago, Los Angeles/Orange County, Minneapolis-St. Paul, the Greater New York Metro, Philadelphia, San Diego, San Francisco/Oakland, San Jose, South Florida and Washington, D.C.
** Data aggregated across the top 11 metropolitan origins of net out-migration from April 2018 to November 2019, which includes Boston, Chicago, Los Angeles/Orange County, Minneapolis-St. Paul, the Greater New York Metro, Philadelphia, San Diego, San Francisco/Oakland, San Jose, South Florida and Washington, D.C.

Figure 4 helps to put the evolution in migration trends during the pandemic into sharper relief by focusing on three periods: the full calendar year immediately preceding the pandemic (2019), the peak migratory impact corresponding with the first year of the pandemic (April 2020 to March 2021), and the most recent trailing 12-months (December 2020 to November 2021). Two key conclusions emerge. First, in most cases the impact of the pandemic on migration trends appears to have peaked in the first year of the pandemic and by late-2021 has already begun to wane, though we still have room to go to return to pre-pandemic rates. Second, the key outlier with respect to the first point is the suburbs of major gateway metros (as proxied by moves to the non-central city in the metros with the largest net outmigration pre-pandemic).

In 2021, however, net outmigration from major gateway suburbs has reaccelerated with more households and business moving away from these suburbs than prior to the pandemic. By contrast, the exodus from the central cities of major gateways has largely reversed (though still has room to go) and that from the central cities of the top sunbelt cities has renormalised. We draw a few lessons from the trends in the USPS COA data, particularly with relation to potential suburban preference and single-family demand. First, very few central cities have been immune to the recent trend in outmigration that existed prior to the pandemic. Households are not just moving to lower cost metros; they are also moving to lower cost suburban locations within these metros.

Second, the impact of the pandemic on migration trends already appears to be dissipating in many instances. The evidence so far is that the impact of the pandemic on domestic migration could be transitory and resolve within the next year or two, particularly to the extent that those with the highest propensity to be motivated by the pandemic to move have already done so.

Finally, the most potentially persistent, and growing shift in preferences so far, as revealed by migratory data, is away from the suburbs of the highest cost gateway metros. While it has been widely noted that most moves during the pandemic were short and within the same metro (as is generally the case), the expanding increase in net outmigration away from gateway suburbs potentially points to a more permanent shift in locational preference, perhaps spurred by remote work possibilities. Households that had moved out to the gateway suburbs to raise families or secure lower cost housing, now have more options in a world of remote work, which in theory should provide tailwinds to suburban housing demand in sunbelt markets, given metro migration trends and that these households have already demonstrated a preference for suburbs. While it has moderated from the pandemic peak, suburban demand in the top sunbelt migration destinations appears as strong as it has ever been.

Source: CBRE Investment Management, CBRE Econometric Advisors, NAR and local realtor associations and USPS.
Note: Net household migration measured as families and individuals that filled out a USPS COA form with the given metro as their destination, less those with the metro as their origin. Rates measured as a percent of pre-existing households in the metro.

The differential impact of migration patterns on home price appreciation and apartment rent growth further supports the notion that the pandemic will have only a limited impact on long-term geographical preferences. There has been a much stronger relationship between migration trends, as measured by the USPS COA data, and apartment rent growth than home prices (Figure 5). First, Figure 5 shows a 78% correlation between net household migration rates and apartment rental rate changes by metro during the pandemic, but just a 46% correlation between the same migration rates and home price appreciation in the same metro. Second, the estimated slope coefficient in simple bivariate regressions of rent growth and home price appreciation on net household migration rates is nearly 80% stronger for the former. Even in New York and San Francisco, the metros most impacted by outmigration, cumulative home price appreciation measured 12% and 19%, respectively, from the start of the pandemic to Q3 2021. Nearly every metro has experienced home price appreciation of almost 20% or higher since the start of the pandemic, regardless of migration rate.

In contrast to renters, who are more transitory by nature, for-sale home values are much more likely to reflect future demand expectations, lending further credence to expectations that the impact of the pandemic on migratory trends will be transitory. Moreover, the robustness of home price appreciation during the pandemic, driven in large part by limited inventory for sale, suggests persistent demand for both major gateway markets and single-family homes, even with the possibility of remote work and fewer households with children.

Dharmasankar, S and Mazumder, B (2016), Have Borrowers Recovered from Foreclosures during the Great Recession? Chicago Fed Letter, No. 370, 2016. Available at https://www.chicagofed.org/publications/chicago-fed-letter/2016/370#ftn1

Goodman, L and Mayer, C (2018), Homeownership and the American Dream, Journal of Economic Perspectives, Volume 32, Number 1, Winter 2018.

Gomez, V (2021). “More Americans now say they prefer a community with big houses, even if local amenities are farther away”, Pew Research Center, 21 August. Available at https://www.pewresearch.org/fact-tank/2021/08/26/more-americans-now-say-they-prefer-a-community-with-big-houses-even-if-local-amenities-are-farther-away/

Kolko, J (2021). “The Downtown Decade: U.S. Population Density Rose in the 2010s”, The New York Times 1 September. Available at https://www.nytimes.com/2021/09/01/upshot/the-downtown-decade-us-population-density-rose-in-the-2010s.html

Nargund, G (2009), Declining birth rate in Developed Countries: A radical policy re-think is required. Facts, Views & Vision in ObGyn, 1(3), 191–193. Available at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4255510/

Office of New York City Comptroller (November 2021), The Pandemic’s Impact on NYC Migration Patterns, Bureau of Budget. Available at https://comptroller.nyc.gov/wp-content/uploads/documents/The-Pandemics-Impact-on-NYC-Migration-Patterns.pdf

Picker, L. (2015) “The U.S. Foreclosure Crisis Was Not Just a Subprime Event”, The Digest, Issue No. 8, (August 2015). Available at https://www.nber.org/digest/aug15/us-foreclosure-crisis-was-not-just-subprime-event

Ruggles S, Flood S, Foster S, Goeken R. et al, IPUMS USA: Version 11.0 [dataset]. Minneapolis, MN: IPUMS, 2021. Available at https://doi.org/10.18128/D010.V11.0

Footnotes

¹ Changes in U.S. Family Finances from 2016 to 2019: Evidence from the Survey of Consumer Finances (September 2020), Federal Reserve Bulletin. Available at https://www.federalreserve.gov/publications/files/scf20.pdf.

² While Millennials are often defined as those born between 1981 and 1996 and would therefore be 23-38 in 2019 (our last year of detailed data), we focus on the subset aged 25-34 for purposes of simplification.

³ Melissa S. Kearney and Phillip Levine (2021), “Will Births in the US Rebound? Probably Not,” Brookings (24 May). Available at https://www.brookings.edu/blog/up-front/2021/05/24/will-births-in-the-us-rebound-probably-not/

⁴ Laurie Goodman and Christopher Mayer (2018), Homeownership and the American Dream, Journal of Economic Perspectives, Volume 32, Number 1, Winter 2018.

⁵ Sharada Dharmasankar and Bhash Mazumder (2016), “Have Borrowers Recovered from Foreclosures during the Great Recession?” Chicago Fed Letter, No. 370, 2016.

⁶ "Bye, bye, baby?” Birthrates are declining globally – here’s why it matters”, World Economic Forum (15 June 2021). Available at https://www.weforum.org/agenda/2021/06/birthrates-declining-globally-why-matters/

⁷ “COVID-19 Accelerates Existing Decline in Fertility Rates”, New Security Beat (23 June 2021) Available at https://www.newsecuritybeat.org/2021/06/covid-19-accelerates-existing-decline-fertility-rates/

⁸ American Community Survey or ACS data are collected via annual surveys that sample approximately 1 in 100 people who are asked questions that do not appear in the decennial census. See The Importance of the American Community Survey and the Decennial Census United States Census Bureau (7 January 2022). Available at https://www.census.gov/programs-surveys/acs/about/acs-and-census.html

⁹ The forecasts used to project the size of the female population in 2029 come from census projections that were released in 2017. The advantage of using (even dated) census projections is that they incorporate the impact of positive net international migration as well as the low probability of death among today’s 25-34 cohort, without which, we would only be able to “roll forward” today’s population count. (“Rolling forward” would assume that the number of 35-44-year-olds in 2029 would be exactly equal to the number of 25-34-year-olds in 2019, leading to a likely undercount of the future population.)

¹⁰ For purposes of our analysis, we define “children” as minor children only (under age 18) and ignore the impact that having adult children has on the decision to live in a single-family residence. We note that while a majority (58%) of 18-24 year-old children lived in their parental home in 2021 as did 17% of adults aged 25 to 34 (see: https://www.census.gov/newsroom/press-releases/2021/families-and-living-arrangements.html), if these young adults have children themselves, they are still included in our analysis (“parent and child in another’s household”.)

¹¹ Gomez, Vianney (2021), “More Americans now say they prefer a community with big houses, even if local amenities are farther away”, Pew Research Center, 26 August. Available at https://www.pewresearch.org/fact-tank/2021/08/26/more-americans-now-say-they-prefer-a-community-with-big-houses-even-if-local-amenities-are-farther-away/

¹² Kolko, Jed (2021), “The Downtown Decade: U.S. Population Density Rose in the 2010s”, The New York Times, 1 September. Available at https://www.nytimes.com/2021/09/01/upshot/the-downtown-decade-us-population-density-rose-in-the-2010s.html