Aging Baby Boomers and the Generational Housing Bubble: Foresight and Mitigation of an Epic Transition

Abstract

Problem: The 78 million baby boomers have driven up housing demand and prices for three decades since beginning to buy homes in 1970 and continuing up the housing ladder. What will happen when boomers begin to sell off their high-priced homes to relatively smaller and less-advantaged generations?

Purpose: This article presents a long-run projection of annual home buying and selling by age groups in the 50 states and considers implications for communities of the anticipated downturn in demand.

Methods: We propose a method for estimating average annual age-specific buying and selling rates, weighting these by population projections to identify states whose growing proportions of seniors may cause an excess of home selling sooner than others. We also analyze the likely supplier responses to diminished demand, and recommend strategies for local planners.

Results and conclusions: Sellers of existing homes provide 85% of the annual supply of homes sold, and home sales are driven by the aging of the population since seniors are net home sellers. The ratio of seniors to working-age residents will increase by 67% over the next two decades; thus we anticipate the end of a generational housing bubble. We also find that younger generations face an affordability barrier created by the recent housing price boom. With proper foresight, planners could mitigate what otherwise could be significant consequences of these projections.

Takeaway for practice: The retirement of the baby boomers could signal the end of the postwar era for planning, and reverse several longstanding trends, leading decline to exceed gentrification, demand for lowdensity housing to diminish, and new emphasis on compact development. Such developments call planners to undertake new activities, including actively marketing to retain elderly residents and cultivating new immigrant residents to replace them.

Research support: The Fannie Mae Foundation.

Keywords:

• baby boomers
• aging
• housing bubble

Notes

1. Occasionally housing industry analysts will take a 10-year view. For example, in a recent special effort, a consortium of leading housing trade organizations forecasted the market for a full decade, from 2004 to 2013 (Berson, Lereah, Merski, Nothaft, & Seiders, 2006). Unfortunately, a 10-year forecast horizon was not long enough for this group, the Bureau of Labor Statistics, or the Economic Development Administration to anticipate the effects of the retirement of the baby boom generation. A 10-year forecast launched from 2007 might now begin to detect these consequences.

2. According to the 2000 Census, only 1.4% of owner-occupied households in the United States were headed by a person under age 25. Age 25 is also generally regarded as the lower boundary of prime working age, when individuals are most likely to hold secure employment.

3. In 1980, the highest per capita spending on housing occurred at age 45; accordingly, Mankiw and Weil (1989) concluded that once the baby boomers passed age 45 they would begin to spend less on housing, reducing housing demand and lowering prices as the boomers aged. But it turned out that lower spending on housing after age 45 was not a result of older adults cutting back. Rather, members of the generation that was over 65 in 1980 were unlike their younger counterparts in that they had never spent much housing. The fallacy of Mankiw and Weil's reasoning was revealed once researchers followed individual cohorts' behaviors as they grew older (Pitkin & Myers, 1994).

4. A recent study by the Federal Deposit Insurance Corporation (Angell & Williams, 2005) reviewed historical housing booms in all U.S. metropolitan areas for which data were available and concluded: “In over 80% of the metro-area price booms we examined between 1978 and 1998, the boom ended in a period of stagnation that allowed household incomes to catch up with local home prices.”

5. Those buying for investment purposes increase their market activity when prices are rising most rapidly and withdraw when prices flatten or decline. For example, between 2002 and 2005, when the recent boom began to crest, the investor share of mortgage originations grew from 6.8 to 10.9% in Los Angeles, from 5.8 to 13.8% in Austin, and from 8.1 to 15.8% in Miami (Harvard Joint Center for Housing Studies, 2007, Appendix W-4).

6. There was a very strong positive correlation between changes in homeownership rates among heads of households aged 25 to 34 and changes in median house values for states, with r = 0.75 in the 1980s and r = 0.65 in the 1990s (Myers, 2001).

7. This ratio is not a representation of the purchase calculation of actual households, but a relative index. In fact, the median income pertains to both renters and owners, and many young buyers are likely to purchase homes below the median price.

8. This is certainly the conclusion of Robert Shiller (2005, 2006), who plots house prices against other asset classes and shows that the recent run-up in house prices is unprecedented.

9. We converted data for the 5-year interval preceding the census to its equivalent for a full decade by summing the experience of adjacent 5-year cohorts. For example, we aggregated 5-year rates for persons aged 35 to 39 in 1995 to 1999 with those for persons aged 30 to 34 in the same time period, synthetically estimating the experience of a full decade.

10. Complete data on estimated buy and sell rates for all 50 states for each of 13 age groups are available from the authors on request.

11. Death rates climb markedly after ages 55 to 64. Averaged between men and women (weighted for the population at risk), the annual probability of death rises to 2.4% between the ages of 65 to 74, 5.7% between 75 and 84, and 15.5% at 85 and older (U.S. Census Bureau, 2003b). Considering that elderly persons also face higher risks of severe illness or physical incapacity, it is not surprising that they would be two to three times more likely to sell their homes than younger homeowners.

12. The choice of calibration period is important, because the estimated rates of behavior are intended to represent underlying demographically based demand and will be held constant in future periods. For this purpose, we judged the period of 1995 to 2000 to be much better for estimating normal market behavior than the boom years of 2000 to 2005, or the recession years included in 1990 to 1995. Although we expect the annual rates will likely rise and fall in the future, we assume that they will fluctuate around the rates we estimated from the age schedule in the baseline period. Thus we feel our result properly depicts the changes in underlying demographically based demand likely to prevail in future periods.

13. In Arizona, the sell rate without correcting for those who sold before migrating to the state is 4.2 per 100 among those aged 55 to 59, and 2.9 per 100 after making this adjustment.

14. This comment is by an anonymous reviewer of the article.

15. In 2005, the year of peak sales of newly built homes, 1.28 million newly built units were sold and 7.08 million existing homes were sold, a ratio of 5.5 to 1. In the low point of the 1991 recession, 0.51 million newly built homes were sold and 3.22 million existing homes were sold, a ratio of 6.3 to 1 (Harvard Joint Center for Housing Studies, 2007, Table A-1).

16. For example, among homeowners ages 50 to 54 in 2004, the median equity was $92,000 and mortgage debt $85,000 (Englehardt, 2006, Table 1).

17. As an example, during the 1980s the number of households in the Cleveland metropolitan area grew by only 18,000, but 46,700 new housing units were constructed, mostly in suburban areas, and unneeded dwellings were left empty in the central city and older suburbs (Bier & Howe, 1998).

18. We estimated growth after 2000 by comparing data from the 2006 American Community Survey to the 2000 Census.