MultiFamily Market Outlook - October 17, 2006

Apartments Have Positive Impact on Property Appreciation Rates

One of the more common barriers to multifamily housing is a presumption on the part of local stakeholders that it will have a negative impact on their property values, especially values of single-family homes. New data recently released from the Census Bureau confirm, however, that this presumption is yet another myth about multifamily housing. In fact, NAHB’s latest analysis of the 2005 American Housing Survey data shows that, on average, single-family homes located near multifamily communities actually appreciated at a higher rate than homes that did not have multifamily neighbors.

Interestingly, there is no evidence that multifamily housing causes single-family house prices to decline over the long-term, although there is some indication of slightly reduced appreciation rates close to the time multifamily housing is first built. Compensating for this, however, is a somewhat higher rate of appreciation in subsequent years, so that after three or four years the cumulative appreciation on single-family homes is essentially the same whether or not an apartment building had been built in the neighborhood in the meantime.

The Source of the New Data

The AHS is conducted in odd-numbered years by the U.S. Census Bureau, and receives funding from the Department of Housing and Urban Development’s budget. The survey is based on a nationally representative sample of U.S. housing units. Because the AHS revisits the same homes every two years (augmenting the sample with new construction) and records the value of the same property at different points in time, it can be used to compute rates of appreciation.

One caveat: Home values recorded in the AHS are owners’ estimates of how much their properties are worth at the time of the survey, which may raise questions about the accuracy of the estimates. Several academic studies have looked into this, and have invariably found that although owners’ estimates of value are not completely accurate, they tend to be off by a percentage that does not vary in any systematic way with characteristics of the house or the persons occupying it. (The more recent studies find that owners tend to overstate the value of their homes by a relatively constant percentage.) This implies that appreciation rates computed from the AHS values will be unbiased.  (For example, if a home’s value is overstated by 8% in two different years, the rate of appreciation measured between those two years will not be affected by the overestimate.)¹  There are other possible sources of error in the data, and the appreciation rates reported in this article are calculated after taking several steps to screen out inaccurate values.²

Another feature of the AHS is that it identifies whether or not certain structures are present within a half block (defined as approximately 300 feet) of the unit surveyed. This enables us to calculate different appreciation rates for single-family homes based on the presence of neighborhood features such as nearby apartment buildings.

Appreciation if Multifamily is Present at the Start of the Period

Table 1 and Figure 1 show appreciation rates for single-family detached homes in various time periods between 1997 and 2005. They also show different rates, depending on whether apartment buildings were present in the neighborhood at the start of the period or not. The 1997-2005 period in particular is examined, because the Census Bureau changed the way it collected neighborhood data in 1997. Starting with that year will produced an analysis based on data collected in a consistent manner. 

The numbers in the figure and table are average annual appreciation rates. For example, the appreciation rate of 6.9% for 1997-2005 means that the increase in the price of an average single-family home between 1997 and 2005 was the difference that would result if the price increased at a rate of 6.9% every year. Of course, it’s no surprise to anyone who follows housing markets that appreciation rates were higher than this in the more recent past.  The average annual appreciation rate for single-family homes between 2003 and 2005, based on the AHS data, was 12.1%.

For each of the time periods shown in Table 1 and Figure 1, appreciation was higher for single-family detached homes if they were in neighborhoods where apartment buildings were present at the start of the period. For example, single-family homes in neighborhoods where multifamily structures were present in 1997 appreciated at a rate of 8.3% between 1997 and 2005. Single-family homes in neighborhoods where multifamily structures were present in 1999 appreciated at a rate of 10.2% between 1999 and 2005, compared to 8.1% for single-family homes that were not near apartment buildings, and so on. 

This is consistent with results reported previously by NAHB based on earlier releases of AHS data. This doesn’t actually prove that building a multifamily building will cause nearby single-family homes to appreciate at a faster rate. An alternate explanation that is also consistent with the data is that areas where markets are growing and housing demand is generally high are places where house price appreciation continues to be above average and where there has been pressure over the years to use land more intensively and build taller structures, such as apartment buildings. Even under this alternate scenario, though, apartment buildings are likely to be fulfilling an important function as part of a healthy economic environment—providing housing for a certain segment of the population in areas where high and increasing prices may price them out of the market for single-family detached homes.   

Appreciation if Multifamily is Built During the Period

The above statistics may not satisfy all critics, however.  A possible concern is that, while single-family homes appreciate at a healthy rate once an apartment building already exists in the neighborhood, the initial building of the building triggers a one-time reduction in house prices. Even if appreciation were later to resume its normal rate, this hypothetical one-time reduction would be a concern to owners of existing single-family homes in the neighborhood.

To investigate this, Figure 2 compares average annual appreciation rates for single-family homes over the 1997-2005 period based on whether apartment buildings were introduced sometime after 1997 or not. For comparison, the figure also shows appreciation rates if commercial buildings or manufactured housing were introduced during the same time interval.

The first thing to notice is that there is no evidence of house prices declining when one of these features is introduced after 1997, although the comparison does show lower rates of appreciation. For apartment buildings, however, the difference is very small—6.6% vs. 6.9%. The difference in appreciation rates if commercial buildings or manufactured housing are introduced is more pronounced.

Table 2 and Figure 3 take this a step further and break the numbers down based on when during the 1997-2005 the neighborhood features were introduced. As the table and figure show, all of the reduction in appreciation is concentrated among single-family homes where multifamily housing (or manufactured housing or commercial buildings) was introduced in 2004 or 2005.  Because the appreciation is based on a final house price measured at some point in 2005, this may include cases where the apartment building is still under construction, where landscaping is not yet complete, or where the building has not yet become fully occupied. 

Moreover, the effect on appreciation where apartment buildings were introduced in 2004 or 2005 is quite mild compared to the effect observed in when commercial buildings or manufactured housing were introduced.

In summary, the 2005 AHS shows that, in cases where apartment buildings are already present within a half block, single-family detached homes appreciated at a faster rate between 1997 and 2005. This is true no matter which time interval within the 1997-2005 period we look at.

In cases where apartment buildings were introduced into the neighborhood after 1997, single-family appreciation rates were somewhat lower. However, the effect is small compared to the effect of introducing commercial buildings or manufactured housing, and is confined to cases where the apartment buildings were introduced after 2003—so that construction may have been still in progress when home prices were measured in 2005. In cases where apartment buildings were introduced into single-family neighborhoods between 1998 and 2003, any reduction in appreciation that may have initially occurred seems to have been completely offset by subsequently higher rates—so that single-family detached homes appreciated at close to an average annual rate of 6.9% between 1997 and 2005, irrespective of any multifamily structures built nearby between 1998 and 2003.

^1.  Eight percent is the highest figure reported in the published literature.  The reference is Kiel and Zabel, “The Accuracy of Owner-Provided House Values: the 1978-1991 American Housing Survey,” Real Estate Economics, 1999.
^2.  There are well-known examples in the AHS of home values that are off by a decimal place.  Home values in the public use AHS data set are also topcoded (truncated at the top end) to preserve respondent confidentiality.  Before computing appreciation rates, homes with recorded values that are either topcoded or below $20,000 are deleted.  Home values are also checked against their previous values and cases with extremely high changes from one AHS year to the next are also deleted.  Finally, cases where the home value is allocated (i.e., where the respondent doesn’t provide an estimate and the home is assigned the same value as another home in the data set with similar characteristics) are deleted before appreciation rates are calculated.

What Looks Like Rising Starts May Not Be, Really

Preliminary numbers released by the Census Bureau  show starts in buildings with five or more apartments increasing by 8% in August to a (seasonally adjusted annual) rate of 265,000, but this is somewhat misleading. At the same time, the Bureau revised its July estimate for five-plus starts down by about 8%—from what was already a weak number. If the August rate of 265,000 holds up after revisions, it still will represent one of the lowest five-plus rates posted in the past five years, and a 17% drop on a year-over-year basis.

Source: U.S. Census Bureau; NAHB Economics Group

While many data indicate that housing markets in general are cooling off (inevitable after an extended period during which they were unsustainably hot), the main question for multifamily construction is to what extent weakness in condo production will be offset by strength in the production of market-rate rental units, and production of affordable units (which should be relatively stable, as it is sustained by the Low Income Housing Tax Credit program). Based on the recent signs of weakness, NAHB's short-term forecast has been adjusted downward. The forecast now calls for a five-plus starts rate of 270,000 in the second half of 2006, bottoming out at 265,000 in the second quarter of 2007 before gradually increasing and returning to a rate of 300,000 by the end of 2008.

Real Rents on an Modestly Upward Track

The real rent index edged back up in August. After inching down from 107.0 to 106.9 in July, the index—which adjusts changes in residential rents for overall inflation using information from the Consumer Price Index (CPI) series—increased to 107.1.

The components of the change break down as follows: The "rent of primary residence" sub-index of the CPI increased at a fairly healthy (seasonally adjusted annual) rate of 4.3%t during the month, in line with its average performance since March. The overall CPI, meanwhile, increased at a relatively modest rate of 3.0%. 

Based on seasonally adjusted Consumer Price Indices; U.S. Department of Labor, Bureau of Labor Statistics. The annual rates indicate what the percentage change would be if the current monthly rate were sustained over a 12-month period. The real rent index is the CPI for rent of primary residence divided by the CPI for all items and scaled so that January 1995 is 100.

Monthly fluctuations of the CPI often are driven by the volatility of energy prices. In August, prices for both energy services and petroleum-based energy products increased at very gradual rates, and this was a substantial factor in allowing residential rents to rise at a slightly higher rate than inflation during the month.

Gross Domestic Product and Job Creation Slowing

Growth of real Gross Domestic Product (GDP) slipped to an annual rate of 2.6% in the second quarter of 2006, according to the “final” estimate released by the Commerce Department on September 28. This definitely was a below-trend pace with sobering implications for the labor market. Payroll employment rose by only 51,000 in September, but August was revised up substantially and average monthly job growth for the second quarter was a respectable 121,000. Core inflation still is running on the high side and business labor costs still are rising at a brisk pace. However, the evolving slowdown in economic growth and job creation should help keep unit labor costs in check, and the recent impressive fallback in energy costs will help stem the “leakage” of higher energy prices into the core.

The Federal Reserve and financial market participants apparently concur with this assessment. The Federal Reserve held short-term interest rates steady at the September 20 meeting of the Federal Open Market Committee (FOMC), maintaining the 5.25% target for the federal funds rate. This was the second consecutive holding action following 17 consecutive quarter-point rate hikes from mid-2004 to mid-2006.  The public statement issued at the conclusion of the September 20 FOMC meeting highlighted the ongoing “moderation” in economic growth, the ongoing “cooling” of the housing market, and the likely moderation of currently “elevated” inflation pressures over time.

The way things are going, stable monetary policy is the best bet for the balance of this year and the early part of next year, and we’re expecting some monetary easing by mid-2007. The current monetary policy stance, the prospects for stable or even easier policy down the line, the slowdown in economic growth, and well-anchored inflation expectations have combined to generate an impressive bond market rally. Following the movement in the fed funds and 30-day LIBOR rates, yields on long-term Treasury securities have come down significantly from their mid-year highs, and those rates should remain close to current levels for some time.

Multifamily Stocks Continue Record-Breaking Performance

During the month of September, the MFSI increased by slightly more than 42 points—a shade more than one-and-a-quarter percent. With this rise, the MFSI is, yet again, at new all-time high, and is more than 30% higher than it was just 12 short months ago. During the past month, the value of the S&P 500 with dividends reinvested jumped by 2.58% and, as a result, it now finds itself almost 11% above where it was a year ago. 

Because the S&P 500 with dividends reinvested rose by more than the MFSI during the month of August, the performance gap—or percentage difference—between the two indexes decreased from 178 June to 175 percent in September—three points lower than the all-time high of 178 percent set the previous month. Despite the very strong 76% rise in the S&P 500 since its most recent low (set in October 2002), the MFSI has risen a staggering 137% during the same 48-month time period. In addition, the MFSI continues to dramatically outperform the S&P 500 over longer time periods, including the past four, five and six years. Since December 1998, the MFSI has risen by a whopping 237% while the S&P 500 with dividends reinvested has gained a meager 22.7%.

 1. For initial article discussing the MFSI in detail see NAHB Multifamily Market Outlook, January 2002.      
2.  Percent difference is defined as (MFSI minus S&P 500 with dividends)/S&P 500 with dividends.

During the month of September, the price-to-earnings ratio (P/E) of the MFSI rose slightly and now stands at 18.32 while the dividend yield, defined as the total cash dividend payments divided by the current stock price, and which moves in the opposite direction eased to 3.43%. The MFSI is an index of 24 publicly traded US headquartered firms, including 20 REITs, principally involved in multifamily ownership and management.