Trends and Cycles in Housing Production
Seiders, David F., Business Economics
Housing is a key element of the national economy and of every local community in the United States. Housing production, as measured by residential fixed investment (RFI) in the national income and product accounts, recently has made up about one-fourth of gross private domestic investment and about 4 percent of gross domestic product (GDP).
The stock of owner-occupied and rental housing produces a flow of housing services that are counted under personal consumption expenditures, accounting for another 8 percent of GDP. In addition, about 8 percent of GDP goes for consumer spending on housing-related goods and services, including furniture, appliances, utilities, and other household operation items. Thus, the housing sector of the economy typically accounts for about one-fifth of GDP, dwarfing all other industrial sectors.
Residential fixed investment "drives" housing's role in the economy over business cycles and on a trend basis. RFI includes the production of new units in permanent single-family and multifamily structures, improvements to existing units in such structures, mobile home production, and brokers' commissions on home sales. This component of the economy, which accounts for about 4 percent of GDP on an annual average basis, contributes a much larger share of the change in GDP at certain stages of the cycle, largely because of the sensitivity of RFI to changes in interest rates. RFI typically accounts for about 20 percent of GDP growth during the first year of recovery following a recession; indeed, RFI accounted for nearly 30 percent of GDP growth in the year following the 199091 recession. RFI also turns down early, usually "leading" the U.S. economy into recession.
The balance of this article discusses trends and cycles in the major components of RFI and presents NAHB's forecasts for the production of new housing units through the year 2005. The discussion also covers some trends in the sizes and major characteristics of housing units produced in the U.S.
About 1.16 million single-family housing units were built in 1996. Nine-tenths of these were detached units built on individual lots, and the balance were attached or "town house" units. Ninety-eight percent of all single-family units were intended for owner occupancy when built, with only 2 percent intended for rent. The heavy prevalence of detached owner-occupied units should dominate the single-family housing market for years to come.
New single-family homes have been getting bigger and better over time, and this trend is likely to be maintained. As shown in Table 1, the median size is approaching 2,000 square feet. Nearly one-third of new homes now have at least four bedrooms, and [TABULAR DATA FOR TABLE 1 OMITTED] nearly half have at least two and one-half bathrooms. More than three-fourths of new homes now have garages for two or more cars.
With respect to construction method, nearly 95 percent of new single family homes are "stick built" on site. Only 3 percent are constructed with factory-built modular components, and the remaining 3 percent are panelized or precut units. Chronic shortages of skilled labor in construction, along with evolving shortages of framing lumber, are likely to force heavier use of modular and panelized construction methods as we move into the next century.
Figure 1 shows single family housing starts (a start occurs when a foundation is begun), back to 1980, with NAHB's quarterly forecast through 1998 to the right of the vertical line. Major cyclical contractions occurred in the early 1980s and during the 1990-91 recession, but volatility has been relatively mild since then. NAHB is anticipating a mild downswing during the second half of 1997 and in 1998, as part of slowdown in the overall economy engineered largely by the Federal Reserve. Single family starts should gravitate back to about 1.1 million units in the final years of the decade as long as the economy resumes trend growth by then.
In 1996, 316,000 units were in new multifamily structures (containing two or more units) started that year. Four-fifths of these units were built for the rental market, while the balance were intended for condominium ownership where individual units are owned by their occupants and there is common ownership of some open space and facilities. The condo share of multifamily production was as high as 40 percent in the early 1980s but has trended down since then. Resurgence of this type of homeownership is not likely to occur in the years ahead.
Table 2 shows that new multifamily units have been getting bigger and better for some time. The median size has exceeded 1,000 square feet in recent years, nearly three-fourths of the new units now have at least two bedrooms, and half have at least two bathrooms.
Table 2 New Multifamily Housing Starts Four Median Average 2+ Bed- 2+ Bath- Floors Sq. Ft. Sq. Ft. Rooms Rooms or More (1) (2) (3) (4) (5) 1979 893 938 59 24 10 1980 915 979 62 30 12 1981 930 980 64 32 16 1982 925 990 62 36 20 1983 893 942 62 33 14 1984 871 914 62 35 10 1985 882 922 61 37 9 1986 876 911 60 36 8 1987 920 980 61 39 10 1988 940 990 63 41 12 1989 940 1000 62 41 12 1990 955 1005 65 44 11 1991 980 1020 69 43 12 1992 985 1040 71 42 14 1993 1005 1065 73 45 10 1994 1015 1035 71 44 8 1995 1040 1080 72 49 8 1996 1030 1155 72 51 6 (1), (2) Square feet. (3), (4), (5) Percent of new homes completed. Source: U.S. Bureau of Census, Construction Reports, Series C-22, Housing Completions.
Multifamily production has gone through some wide fluctuations since 1970, and NAHB is projecting multifamily starts that are modest by historical standards. It's worth noting, however, that previous periods of high multifamily production were spurred by changes in federal policies that subsequently were reversed. In particular, the boom of the early 1970s reflected a sharp expansion of federal subsidy programs for production of low-income rental housing, and the boom of the early and mid-1980s reflected major enhancements of tax incentives for production of rental housing. The 1980s boom, which was scuttled by the Tax Reform Act of 1986, generated high rental vacancy rates that depressed multifamily production into the 1990s.
The forecast presented in Figure 6 assumes that there will be no major changes to federal housing policy through 2005. This is a major assumption that must be constantly reassessed. Current policy support could be reduced in the interests of deficit reduction, and both the tax deduction for home mortgage interest and the low-income housing tax credit are vulnerable on that front. Furthermore, major federal supports for the housing finance system, including the three government-sponsored enterprises (Fannie Mae, Freddie Mac and the Federal Home Loan Bank System) are under constant scrutiny by Congress and the Administration.
On the other side of the policy coin, growing shortages of decent and affordable rental housing units for low-income people are becoming increasingly visible and are receiving more attention in the media. At some point, the federal government may feel compelled to increase support for low-income rental housing, either through expenditure programs or the tax code.
Figure 2 shows NAHB's forecasts of multifamily housing starts through 1998. This market saw massive levels of tax-driven production in the mid-1980s, followed by a seven-year collapse subsequent to the Tax Reform Act of 1986. A reasonable recovery has occurred since 1993; however, the supply-demand balance weakened somewhat in 1996 and a mild downswing is in store for the rest of 1997 and 1998. Multifamily housing starts are not likely to climb much above the 300,000 unit mark during the balance of this century, even on a quarterly basis.
Residential remodeling amounted to an estimated $115 billion in 1996 (current dollars). About 35 percent of this total represented maintenance and repair expenditures, while the balance was "improvements" (additions, alterations and major replacements) that are embedded in residential fixed investment in the GDP accounts.
Figure 3 shows the real value of residential remodeling, on an annual basis, back to 1980; 1996-98 represent NAHB forecasts because the data for 1996 are not yet complete. Contrary to popular belief, the remodeling market is procyclical (rather than countercylical), although fluctuations are quite mild in comparison to the market for new units. NAHB is forecasting slight year-to-year increases in 1997 and 1998 for both maintenance/repair and improvements.
About 310,000 mobile home units were placed for residential use in 1996, about the same as the number of apartment units in new multifamily structures. As in past years, nearly two-thirds of the mobile home placements were in the South region of the country and one-sixth were in the Midwest. Double-and triple-wide units have gained market share in recent years, and the traditional single-wide units accounted for only about half of all mobile home placements in 1996. In spite of the uptrend in size, the average price of mobile homes was only $38,500 in 1996, compared with an average price of $174,500 for permanent new single family homes.
Figure 4 shows NAHB's forecast for mobile home placements through 1998. This market displays pro-cyclical characteristics, but shifts in the regional composition of economic activity are quite important as well. The projected downswing in 1997-98 reflects, in part, a regional economic and housing forecast that involves some loss of market share in the South and Midwest.
RESIDENTIAL FIXED INVESTMENT
Figure 5 shows the RFI component of GDP, with NAHB forecasts through 1998. The cyclical volatility of RFI is legendary, and sharp ups and downs surrounded the deep recession of the early 1980s as well as the brief and shallow recession of 1990-1991. Real RFI hit a record high in 1996, reflecting the movements in the components discussed above as well as a record year for sales of existing homes (brokers' commissions are included in RFI).
RFI should be a drag on GDP growth during the second half of 1997 and most of 1998. RFI should become a roughly neutral factor in GDP growth during future years as long as the economy hangs around trend growth, posting 2 to 3 percent growth in real terms in spite of a rather flat trend for the number of new housing units produced. Positive growth will be generated by a persistent up-drift in the size and quality of new housing units along with freight-train growth in both residential remodeling and existing home sales.
Trends in housing production are primarily set by net household formations, i.e., by changes in the number of households in the United States. Indeed, household growth will account for about four-fifths of the demand for new housing units during the next decade. The remaining one-fifth represents new units needed to replace units lost from the existing housing stock (primarily through demolition) and units that will be vacant (including second homes).
As discussed above, new housing units produced to meet total housing demand can be permanent single-family or multifamily units (those represented by data on housing starts and housing completions), or they can be mobile homes placed for residential use. In 1996, housing starts and mobile home placements totaled 1.772 million units. Permanent single-family units accounted for two-thirds of the total, while multifamily units and mobile homes claimed roughly even shares of the remainder.
Figure 6 shows NAHB's long-term forecast for the production of new housing units, divided into single-family and multifamily units as well as mobile homes. Total production for the 1996-2005 period averages 1.68 million units per year, with 1.12 million single-family units, 280,000 multifamily units and 280,000 mobile home placements.
The average annual production of new housing units for the 1996-2005 period compares favorably with average production levels for the years 1981-95, but current and projected production levels fall well short of the numbers posted in the 1970s (an average of 2.1 million units per year). However, the production of single-family units is expected to equal or surpass all periods except the boom of the late 1970s.
David F. Seiders is Chief Economist and Senior Vice President for Economics, Mortgage Finance and Housing Policy at the National Association of Home Builders, Washington, DC. The data and forecasts presented in this article are contained in various issues of the following monthly publications of NAHB: Housing Economics, Housing Market Statistics, and Home Builders Forecast.…