Special Studies, February 11, 2009
By Paul Emrath, Ph.D.
Report available to the public as a courtesy of HousingEconomics.com
Builders and manufacturers of building products are often interested in how long home buyers actually stay in the homes they purchase—for example, to gain insight on what types of building products to use and how to market them. Buyers who intend to remain in their homes for at least thirty years are likely to be interested in siding, windows, flooring, roofing, heating equipment or other items that are durable, require relatively little maintenance, and conserve on energy use to reduce operating costs in the long run. Buyers who intend to remain in their homes for less than five years may still be interested in durability, low-maintenance, and lower operating costs, but they tend to think of these characteristics more in terms of their impact on resale value. A caveat here is that the length of time buyers ultimately remain in their homes may be different from their expectations when they first move in, but the expectations will govern what prospective buyers are looking for when shopping for a home.
Housing policymakers and advocates also may be interested in how long buyers tend to remain in their homes. In the current economic and political environment, the issue arises when evaluating the cost to the American taxpayer of policies such as subsidized mortgage. In this context, it’s necessary to estimate how long the average home buyer will stay in a home in order to estimate how long the average subsidized mortgage will last.
American Community Survey and the Simplest Question
One simple way to gather information about how long people remain in their homes is to ask a representative sample of home owners how long it’s been since they moved in. This is, in fact, one of the questions the U.S. Census Bureau asks in its American Community Survey (ACS) - the survey based on (and that has replaced) the long form questionnaire formerly used in the decennial census. Although it takes the ACS five years to collect enough data to produce the geographic detail formerly available in a decennial census, that still translates to sample of about 1.3 million housing units a year for the current version of the survey, and about 500,000 a year in 2003 and 2004 during its initial phase when the survey was being ramped up to its present form.
According to data from the most recent (2007) ACS, a little over 22 percent of single family home owners have been in their homes 10 to 19 years, 12 percent have been in their homes 20 to 29 years, and a little over 15 percent have been in their homes at least 30 years. Added together, this comes to roughly half of all single family home owners having lived in their homes for at least 10 years. There has been very little change in this percentage since 2003.
These ACS percentages can’t readily be used to produce a true average length of time buyers remain in a home, however. As discussed in a 1992 NAHB article by Michael Carliner and David D’Alessandris, data on home sales has traditionally shown that about five percent of the owner-occupied stock turns over in a given year. If all homes sold exactly at that rate, it would imply an average stay of 20 years, about the same number you would get if you assumed that single family owners in the ACS have on average been in their homes for half the time they will ultimately live there. But this would overstate the typical length of stay because it neglects the effect of frequent movers—that is, those who will move more than once in a ten year period and bring the average down. In order to calculate a true average, it would be necessary to begin with a representative sample of people buying a home in a given year, and then track them year after year until the last one moves out. Information has not been collected in a consistent manner for a long enough period to produce a data set like this.
American Housing Survey and Home Buyer “Survivors”
But this is essentially the same data problem statisticians face when trying to calculate an average life expectancy. The typical solution is to use death rates for people of various ages, which can be computed from recent annual data, to construct survival tables. The survival tables show a “life expectancy” in the sense that half the people who are or would have been that age are expected to still be alive.
Similarly, data on mobility rates of home owners by year moved in can be used to estimate an expected “length of stay” in the sense that half of the households who buy a home are still expected to be in that home. In the aforementioned 1992 article, Carliner D’Alessandris pioneered a methodology based on these principles, using data from the American Housing Survey (AHS). The AHS is based on a representative sample of U.S. housing units, is funded by the Department of Housing and Urban Development, and is conducted in odd-numbered years by the U.S. Census Bureau.
Because the AHS revisits the same housing unit every two years, it’s possible to match up owner-occupied units in one survey with the same units in the next survey and see if the owners have moved out. In order to deal with the every-other-year nature of the survey, Carliner and D’Alessandris assumed that the number of owners moving in one year was half the number moving in the two years between surveys. They originally applied this to data from the 1985-1987 and 1987-1989 AHS panels. The result of the same method applied to data from the 2005 and 2007 AHS is shown in Table 1.
AHS data indicate that 7.6 percent of home owners who bought single family homes in 2005 had moved out by 2007, so the table shows a mobility rate of half of this (3.8 percent) for those who had lived in their homes for a year. In other words, 3.8 percent of households who buy a home will move out within a year, but 96.2 percent will “survive” - i.e., still be living in the home they bought a year ago. Similarly, the AHS data produce a mobility rate of 5.5 percent for owners who purchased their homes two years ago. Subtracting 5.5 percent of the 96.2 percent of owners who remain after one year generates a 91 percent “survival” rate after two years, and so on. In general, mobility rates peak around 3-4 years after the move-in and then decline gradually thereafter - although the decline is not perfectly smooth, as might be expected given the relatively small number of observations in some of the cells in Table 1.
The result of this decline in mobility after year four is that many of the buyers who remain in their homes after that time will still be there years later. The survival table shows that about three-fourths of homebuyers are still expected to be in the homes at the end of the fifth year. Over half of these (38 percent of the total who originally moved in year 0) are still expected to still be living there 25 years after they moved in.
Based on the data in Table 1, it will take about fifteen years for half of current single family home buyers to move out of the homes they purchased.
Effect of Housing Characteristics
A natural follow-up question is whether the length of time buyers remain in their homes varies based on characteristics of the homes purchased. The extent to which the survival table method and AHS data can be used to investigate this is limited, because the AHS provides only recent data on housing characteristics. It therefore makes sense to look for differential lengths of stay only based on housing unit characteristics that are not likely to change even over an extended period of time - such as the type of structure, or region in which it’s located.
In addition to single family home buyers, Table 1 also shows survival rates for buyers of multifamily condominiums. Move-out rates tend to be considerably higher among the condo buyers. For example, the table shows that over sixteen percent of condo buyers move out the very first year. The result is that, based on the 2005 and 2007 AHS data, it will take it will take only about six years for half of current home buyers to move out of the homes they purchased.
Two caveats apply to this conclusion. One is that the table of condo survival rate is based on a much smaller sample. According to the 2007 AHS, there were fewer than 4 million owner-occupied multifamily condos, compared to more than 66 million owner-occupied single family homes. This translates into many of the cells in the condo survival table that are based on relatively few observations. The correspondingly large sampling error helps explain why the condo move-out rates are so variable from year to year.
The second caveat is that the 2005 and 2007 AHS data may be capturing a historically atypical period of condo market activity. In 2005, sales of existing condos soared to nearly 900,000 a year—almost double the historical average. Since then, condo sales consistently declined while the inventory of unsold condos on the market swelled from under 5 months’ to a record of over 14 months’ supply.
To investigate the issue of historical patterns, Figure 2 shows the estimated time it will take half of single family and multifamily condo buyers to move out based on pairs of AHS data sets from 1985 through 2007. The year shown on the horizontal axis of the chart is the most recent year of the pair of surveys used to generate the results. For example, “1987” means that the bars of the chart are based on data from home owners who moved or stayed put between 1985 and 1987. For each year, the figure shows results for both single family and multifamily condo home buyers.
The results show that the recent period has indeed been atypical. The longest expected time until half of current single family home buyers move out-15 years-is based on data from the 2005 and 2007 surveys. The shortest expected time until half of current single family move out—10 years—is based on data from the 2003 and 2005 surveys. The 2003-2005 period was characterized by hot single family housing markets, rising home prices, and record number of home sales, so it’s not surprising that mobility rates were unusually high and “survival” rates unusually low at this time.
The low mobility and high survival rates based on 2005-2007 data reflect the general slowdown in single family housing activity that occurred after 2005. Single family mobility rates tend to decline and single family homeowner survival rates tend to be somewhat higher when housing markets are weak, as in the early 1990s.
Until recently, condo markets followed a similar pattern. Although the expected length of stay in a multifamily condo is always shorter than the expected length of stay in an owner-occupied single family home, the single family and condo series tended to track one another. The expected time owners remain in either type of structure tended to decline during periods of hot housing markets and increase when housing markets were weakening. The peculiar combination of increasing move-out rates for condos at a time of weak housing markets when single family mobility is declining is confined to the 2005-2007 period and results in an unprecedented difference between the expected time owner-occupied single family multifamily buyers will remain in their homes that shows up in the most recent data.
The AHS sample of single family home owners is large enough so that it can be broken down by the four principle census regions.
The results show that the expected length of time single family buyers will remain in their homes is only 13 years in the South or West, compared to 17 years in the Midwest, and 22 years in the Northeast. These results conform to the regional pattern of home buying activity, with shorter lengths of stay in the regions with the most activity. In 2008, over half of single family sales occurred in the South, about a quarter occurred in the West, 14 percent in the Midwest, and only 7 percent in the Northeast. Although the overall level of single family sales has varied over time, the general pattern of the greatest number of single family sales taking place in the South, followed by the West, with a relatively small share in the Northeast has prevailed for decades.
First-time vs. Trade-up Buyers
A related issue is whether characteristics of the buyers influence how long they remain in their homes. Again, the number of buyer characteristics that can be investigated is limited, because we have only current data and many characteristics of the household may have changed since the home was purchased. Fortunately, the AHS collects data on whether current home owners had owned a home previously, and this can be used to construct separate survival tables for first-time and trade-up buyers.
The behavior of first-time buyers is often a relevant policy issue, as some types of federal assistance (such as the Mortgage Revenue Bond program or the recently enacted Homebuyer Tax Credit are targeted (and restricted) to first-time buyers. Also, first-time buyers have, on average, certain characteristics that distinguish them from buyers who have owned homes previously. As described in a previous article compared to trade-up buyers, first-timers tend to be younger and have somewhat lower incomes. This suggests that first-time buyers might experience changes in job or family status causing them to move out sooner.
Table 2 does, in fact, show a tendency for first-time buyers to move out earlier. Compared to the rates for trade-up buyers, the move-out rates first-time buyers are consistently higher for the first eight years. The result is that it will take about thirteen years for half of first-time buyers to move out of their homes—two years faster than it takes home buyers in general to reach this point—and seventeen years for trade-up buyers.
Table 2 is based on 2005 and 2007 data. Figure 2 extends the results by showing the estimated time it will take half of both first-time and trade-up buyers to move out based on pairs of AHS data sets from 1985 through 2007. Again, the year shown on the horizontal axis is the most recent year of the pair of surveys used to generate the results.
The relationship between first-time and trade-up home buyers varies across the time period captured in Figure 4, but not drastically. The estimated time until half of buyers move out of their home is always shorter for first-time buyers than for trade-up buyers. The difference averages takes less time for half of first-time buyers to move out throughout the period considered in Figure 2. The average difference is just about three years and varies from one year in 1989, 1995, and 2001 to six years in 2003.
The longest expected times until half of both types of buyers move out occur in the most recent period (based on moving tendencies between the 2005 and 2007 data). The shortest expected lengths of stay - 13 years for trade-up buyers and 9 years for first-time buyers - are based on data from the 2003 and 2005 surveys. The general tendency for homeowner survival rates tend to be lower when housing markets are strong and lower during periods of comparative weakness is present among both first-time and trade-up buyers.
Summary and Conclusion
This article has presented information on how long home buyers stay in their homes from two surveys produced by the government. The American Community Survey, based on a very large sample, shows that about half of single home owners report having lived in the same home for at least ten years. During the limited time this survey has been conducted, there is little variation in this statistic.
The American Housing Survey, although based on a smaller sample, collects more detailed information about homes and home buyers. Based on the most recent data from this survey, the expected length of time home buyers remain in the homes they purchase (in the sense that half of the buyers will still be in the homes and half will have moved out) is 15 years for all single family home buyers, 13 years for first-time buyers, and 17 years for trade-up buyers. The most recent data come from an atypical period when housing markets were cooling down from an overheated condition, however. A peculiar aspect of the most recent period is that, at a time when the expected length of stay in a single family home has risen to the longest ever in the period analyzed, the expected length of stay in a condo has fallen to the shortest ever—only six years.
Averaged across all eleven data points shown in Figures 2 and 4, the average length of stay in a single family home is a little over 12 years for all home buyers, 11 years for first-time buyers, and a little under 14 years if the analysis is confined to trade-up buyers.
In the current economic and housing downturn, information like this is particularly useful when evaluating policies designed to stimulate home buying. For example, the “Fix Housing First” strategy that NAHB is advocating includes a federal government buy down of the interest rate on 30-year fixed rate mortgages. In order to estimate the cost of the buy down to the American taxpayer, it’s necessary to estimate how long the average subsidized mortgage will last. Because buyers are less likely to refinance or prepay a strongly subsidized mortgage before they move out, the cost of the average subsidized mortgage can, to a first approximation, be based on the expected length of stay in a home reported in this article.
For more information about this item, please contact: Paul Emrath at 800-368-5242 x8449 (firstname.lastname@example.org)
Table 1 (Excel)
Table 2 (Excel)
See previous Special Studies
 Throughout this article, the single family category includes single family attached as well as single family
detached homes. Single family attached homes are side-by-side townhomes separated completely by a wall that runs from roof to basement and is not broken by any pipes or wires for shared utilities. The number of single family attached units is relatively small and adding them to a sample of single family detached units tends not to change the results drastically.
 “Home Owner Mobility and Mortgage Prepayment” Housing Economics, September 1992.
 Trade-up buyer is used as a convenient term to denote all buyers who have owned a home previously, even though some of these buyers will be moving to smaller or less expensive homes.