An implication is that, in the short run, particular segments of the housing market may be affected more by changes in the status of existing structures than by the addition of new ones. Rental markets, for example, may be affected by demolition or the conversion of existing structures from one use to another. Markets for housing that serves low-income renters may be affected by these factors, as well as by changes in rents charged for existing units.

The Office of Policy Development and Research (PD&R) in the Department of Housing and Urban Development has commissioned Econometrica Inc. of Bethesda, Md., to produce a series of reports that analyze changes in particular rental units over time. These “Rental Market Dynamics” reports are part of the “Components of Inventory Change” system that studies the way housing markets evolve by comparing data from two consecutive installments of the American Housing Survey (AHS). The AHS is based on a representative panel of housing units in the U.S, is funded out of the Office of PD&R, and conducted in odd-numbered years by the U.S. Census Bureau. Because the AHS revisits the same units each time the survey is conducted, it can be used to track changes in particular units over time.

The latest Rental Market Dynamics report, based on data from the 2005 and 2007 AHS, was released in June of 2009. This article summarizes information from that report.

Where Does New Rental Housing Come From?

According to AHS data, there were 39,756,000 rental units in the U.S. in 2007. This total includes both units that are renter-occupied and those that are vacant for-rent. Most of these (nearly 85%) also were part of the rental housing stock in 2005. Where did the other 15% (a little more than 6 million housing units) come from?

About 650,000 of the 6 million were added as the result of new construction. This is only a little more than half of the number added through units being converted from seasonal to rental use, and a little more than one-sixth the number added through conversions from owner-occupied to rental housing (Figure 1).

Even “other additions to the housing stock” (a category that includes mobile home move-ins and structures converted from non-residential use) accounted for more than half a million units, a number not much smaller than the number added through new construction.   

Where Does Old Rental Housing Go?

“Rental Market Dynamics” also attacks the problem from the opposite angle, beginning with the 2005 data and investigating what happened to specific units that had disappeared from the rental stock by 2007. 

The AHS data produce an estimate of 38.4 million rental units in 2005. Slightly more than 85% of these were still rental housing units in 2007. What happened to the other 15%? The greatest number (3.2 million) were housing units that changed tenure—going from rental units to units intended for owner occupancy. Next in line were units that changed from being rental to seasonal (1.5 million), followed by units that were lost to the housing stock entirely (about 900,000). The losses include units demolished or destroyed, as well as housing units converted to non-residential use. Some of the conversions eventually could be recovered by being re-converted to residential use at a later date. Each of the three major components of losses was larger than the amount added to the stock through new construction (Figure 2).

It may seem peculiar; but, if the information on additions in Figure 1 and losses in Figure 2 is netted out, the result is not the difference in the estimated stock of rental units between 2005 and 2007. This has to do with non-response problems that cause weights on particular housing units to change between surveys. The final section in “Rental Market Dynamics” shows the results of Econometrica, Inc.’s procedure for reconciling the differences. The separate forward and backwards results are shown here primarily in order to take advantage of greater detail available in certain categories of additions and subtractions. The results should not be too confusing, provided readers do not expect the numbers to add up perfectly.

The results show that roughly 15% of particular rental housing units observed in 2007 consisted of units added to the rental stock between 2005 and 2007, and only a small share of these were added through new construction. Similarly, 15% of particular rental housing units observed in 2005 were lost to the rental stock between 2005 and 2007, and this would be enough to overwhelm the less than 2% added through new construction without counterbalancing conversions from previously owner and seasonal units.

Rental Housing to Serve Low Incomes

The situation for affordable rental housing is even more fluid. In addition to the factors considered above, additions and subtractions to the stock of rental housing affordable to particular families can change as rents either go up or down—a process known as filtering in the academic literature. 

“Rental Market Dynamics” investigates changes in affordable rental units in some detail, classifying them in eight affordability strata, based on how much income it takes to afford them. For this purpose, the standard AHS data set is supplemented with local information on median family income, and a rental unit is classified as affordable if gross rent (which includes utilities, irrespective of who pays them) is less than 3% of an income threshold expressed as a share of the area median. 

In order to simplify the presentation in this article, the numbers are collapsed into a single affordable category, which includes any rental unit flagged as being subsidized, paying no cash rent, or where the gross rent is affordable to a household with 60% of area median family income or less. That 60% of area median income is a standard used in the Low-Income Housing Tax Credit program, and most other rental housing programs target incomes at or below 60% of the area median.    

By this standard, there were 26.4 million affordable rental units in the U.S. in 2007. About 6.5 of these had not been affordable rental units in 2005. About 2.9 million of the 2007 affordable rental units were generated by rental units filtering down (either reducing rents or holding rents relatively constant while local incomes are rising) to serve lower-income households. Another 2.2 million were added by owner-to-rental conversions (Figure 3). The number of affordable units added through new construction over the same two-year period is small by comparison—240,000.

Starting with the rental units that were affordable (at 60% of area median income or lower) in 2005, 8.1 million were no longer affordable rental units in 2007. Of these, 4.1 million remained part of the rental housing stock but filtered up to the point where they were no longer affordable to a median income family. Another 2.2 million were lost through rental-to-owner conversions (Figure 4). Again, these numbers are large relative to the number of units added to the affordable rental stock through new construction.

Conclusion

In the long run, all housing is created by new construction. However, the amount of construction taking place in any one year is small relative to the magnitude of the total stock. As a result, other factors—particularly owner-rental-seasonal conversions and filtering of existing units to serve different income brackets—have the potential to have a greater impact on rental markets than new construction in any one year. 

The latest rental dynamics study produced by the Department of Housing and Urban Development shows that the numbers of units added to, and subtracted from, the rental housing stock through conversions and filtering were each in the range of 6 to 7 million units over a 2-year period. These numbers are much larger than the 240,000 affordable rental units added through new construction over the same period.

It is necessary to bear this in mind when evaluating the need for programs that support the construction of affordable rental housing, especially during times of economic weakness.

A temporary credit crunch or period of declining house prices may inhibit home sales, and force properties that would normally be owner-occupied into the rental market, causing a spike in rental vacancies. A temporary drop in oil prices may drive down utility costs, making existing units more affordable to low-income renters.

Short-run changes like these may easily be reversed. When economic growth and employment are increasing, housing units that were being offered for rent on a temporary basis may be sold to owner-occupiers. Rents on existing units may rise to the point where low-income families cannot longer afford them. 

It’s not safe to assume that owner-to-renter conversions and downward filtering of particular units will be sufficient to offset renter-to-owner conversions and upward filtering when economic fortunes change. Such conversion and filtering of the existing housing stock has the potential to create or exacerbate a shortage of affordable rental housing that, given the magnitude of new construction relative to the size of the existing rental stock, could take some time to overcome. [ return to top ]