Discover more from CalculatedRisk Newsletter
Lawler: Update on Demographic Trends, and Population and Household Projections Through 2025
From housing economist Tom Lawler:
Executive Summary: Higher Net International Migration (NIM) and lower deaths have improved the US “demographic” outlook, though population growth from 2022 to 2025 will still be historically on the low side – in part because of continued very low birth rates. Household growth, on the other hand, will likely be materially slower than the apparent very rapid growth following the pandemic, and will probably not be that different from the annual average growth experienced last decade.
The Census Bureau’s “Vintage 2022” population estimates suggested that the US resident population increased by 0.38% from July 1, 2021 to July 1, 2022, very low by historical standards but up from the 0.16% from the previous year. The uptick in growth reflected a modest increase in the number of births (from very low level) and a significant increase in estimate net international migration (from very low levels), offset by a slight increase in the number of deaths. (Note that CDC data suggest that the Vintage 2022 estimates for deaths over the 12-month period ending 6/20/2022 is understated by almost 20,000.).
Available data since July 1, 2022 suggest that (1) births over the 12-month period ending July 1, 2023 will probably be down very slightly from the year ago 12-month period; (2) deaths will be significantly lower over the 12-month period ending July 1, 2023 compared to the year ago 12-month period, mainly reflecting substantial declines in Covid related deaths; and (3) net international migration over the 12-month period ending July 1, 2023 should be higher than the year ago 12-month period.
Provisional CDC data on births show that US births from July 1, 2022, through March 31, 2023, were down 0.9% from births in the July 1, 2021, through March 31, 2022, period. These data suggest that births from July 1, 2022, to June 30, 2023, will be down slightly from the July 1, 2022, to July 1, 2023, period.
CDC provisional data on US deaths show that US deaths from July 1, 2022 to April 30, 2023 were down about 10% from the July 1, 2021 to April 30, 2023 period. While because of reporting delays the 2023 data will be revised up slightly, the data clearly show that US deaths over the 12 month period ending June 30, 2023 will be down significantly from the year-ago 12-month period. (Though note again that the Vintage 2022 estimates for deaths in the last reported year were understated by almost 20,000).
Net International Migration:
While there are no timely data on net international migration, the limited data available suggest that net international migration over the 12 month period ending June 30, 2023 will be higher than the year ago 12-month period. For example, from July 2022 to May 2023 immigrant visa issuance totaled 556,611, compared to 443,733 from July 2021 to May 2022. In addition, while quarterly data on immigration from the DHS is horrifically lagged (and more so than before), the number of new arrivals obtaining legal permanent status in the quarter ending 9/30/2022 totaled 135,550, compared to 104,597 in the comparable quarter of 2021. These data suggest that over the last year net international immigration rose again, though it is difficult to quantify by how much.
Taking all of these data into consideration, my “best guess” is that the US resident population as of July 1, 2023 (assuming the Vintage 2022 population estimates are correct) will be 334,984,923, up 0.51% from July 1, 2022. Here are Vintage 2022 components of change for 2021 and 2022 as well as my assumptions for 2023.
Here is a chart showing the annual growth rate of the US resident population since 2011. Note that Census has not released its updated intercensal population estimates for 2010-2020, but the growth rates in the chart below take into account Census’ already announced updated estimates of net international migration over this period.
To estimate the July 1, 2023, resident population by age one needs estimates of (1) deaths by single year of age; (2) net international migration by single year of age; and (3) the resident population on July 1, 2022, by single year of age. In terms of deaths, the CDC has pretty timely data on deaths by age groups, as well as cumulative deaths by single year of age since the end of 2019. As such, one can produce reasonable estimates of deaths by single year of age over the year ending 6/30/2023.
In terms of net international migration, there are no timely data on immigration by age and no data at all on emigration by age. I am assuming that the distribution of net international migration by single year of age for 2023 will be the same as Census assumed for 2022.
For my projections of the US resident population by age through 2025, I am assuming that (1) births will increase, but very slowly, over the next two years; (2) death rates by single year of age will revert back to pre-pandemic levels; and (3) net international migration will edge slightly lower from the apparent elevated level in 2023.
Here are my aggregate assumptions of the components of change for the US resident population through 7/1/2025:
And here are my resident population projections by age.
What Are the Implications for Household Growth?
If one had assumed that headship rates by age had remained constant from 2020 through 2023 (using Decennial Census household counts by age), household growth over that period would have been relatively low. We know, however, that headship rates were not constant but instead increased significantly, though we don’t know by how much, as Census produces multiple and more often than not conflicting estimates of US households. (I’ve written about this “household conundrum” for decades and won’t focus on that here). But rather than focus on what we don’t know, let’s focus on what some surveys suggest.
Data from the American Community Survey (ACS) suggested that US headship rates increased significantly from 2019 to 2021, with especially large increases in young (under 35) adult headship rates. (Census did not release ACS estimates for 2020 because of data quality concerns related to the pandemic). Data from the CPS/ASEC suggested that US headship rates increased from March 2021 to March 2022, again especially amount young adults.
In a study using IPUMS data from the ACS, Ozimek and Carlson found that increases in remote were associated both with increases in spending on housing and increases in household formations. (“Remote Work and Household Formation,” 4/11/2023)
If in fact an increase in the number of employees working from home (and other pandemic related factors) were behind the jump in headship rates from early 2020 to early 2022, then the fact that the share of workers working remotely full-time has been declining since then (and other pandemic-related fears have eased considerable) would suggest that headship rates since early 2022 have stopped increasing, and in fact may have edged slightly lower (though still remaining well above 2020 levels). If that were the case, then one would expect that household growth since early 2022 would no longer have been “goosed” by higher headship rates, but instead would be more in line with, and perhaps even somewhat slower, than increases in population.
Translating all of this information into actual and projected household growth, however, is challenging given the lack of actual and timely historical data. Based on the limited data available, below are my “guesstimates” for household growth from 2020 to 2023, as well as my projections through 2025. In terms of headship rates, I assumed that headship rates by age gradually declined from 2022 to 2025, but that over that period the “reversal” in headship gains was only 30%. (Yes, I am aware that is arbitrary.). Also shown are what household growth would have been if headship rates by age had been unchanged from 2020 (I used Decennial Census data to derive 2020 headship rates, though that data may have issues.)
These estimates and projections assume that the “Vintage 2022” population estimates are accurate, though as I noted before, that may not be the case, and they are subject to possible significant revisions. In addition, I made some somewhat arbitrary assumptions, which may not be valid. I’ll be updating these estimates and projections as more data/information become available.
For folks who wonder what the 7/1/2022 population estimates might look like if Census had used the Decennial Census results (which, again, were not available when Vintage 2022 estimates were released) as a starting point, here is my estimate:
This was from housing economist Tom Lawler.
CalculatedRisk Newsletter is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.