The Impact of Market Variations on Cap Rates
Understanding the Idiosyncrasies of Appraisal-Based Pricing .
by Petros Sivitanides, Jon Southard, Raymond G. Torto, and William C. Wheaton
In the investment markets, the key measure of how a stock is viewed is the ratio of its price to earnings. In markets with high liquidity, investors look to P/E ratios as reflecting opportunities, risk, and uncertainties surrounding earnings growth. Higher P/Es signal greater expected growth and less uncertainty; lower P/Es imply the reverse. And, of course, average P/E ratios over time should also reflect the opportunity cost of risk-free government yields—what the investment would earn if put into U.S. Treasurys.
In real estate, the number that corresponds to a stock’s P/E is the capitalization rate, or current net operating income as a percentage of valuation. Since most real estate is privately owned and individual properties are traded only occasionally, there has always been considerable interest in whether real estate cap rates act the way stock P/Es do or should. More specifically, in the face of scant data from actual transactions on property values and the need to rely on periodic appraisals, do those valuations move, like P/Es, with the opportunity cost of capital and reflect expectations about future income growth and risk?
To study this question, we examined average capitalization rates over 16 years and across 14 metropolitan markets. We did this for four property types, obtaining consistent results that exhibit interesting differences. Our data source is well-known, the shared property performance database of the National Council of Real Estate Investment Fiduciaries (www.ncreif.org). Despite the widespread availability of NCREIF data, our study is the first to systematically examine NCREIF capitalization rates at the local level.
Our conclusions are quite clear. Capitalization rates extracted from the NCREIF database move exactly as P/E ratios do, but only if appraisers form expectations about future income growth by looking myopically backward, not forward. Even though real estate markets are widely felt to be mean-reverting (that is, over time they tend to return to a long-run average level), the NCREIF capitalization rates move contrary to this. They tend to assume income growth will continue at the same pace. When the market is at historically high levels, capitalization rates are low rather than high; in anticipation of mean reversion, they should be high.
An additional and more practical conclusion from our analysis is the strong suggestion that it’s possible to reliably forecast appraisal-based capitalization rates based on forecasts of market rents and interest rates.
In the stock and bond markets, prices tend to adjust rapidly to changes in information and market conditions. In private markets, however, the lack of liquidity and the use of appraisal pricing means that measured capitalization rates will adjust much more gradually in response to new information. Some of this “smoothing” is due to institutional factors, but it also reflects fundamental appraisal behavior. To capture both effects, we treat cap rates as adjusting around equilibrium values.
There are two sets of influences on cap rates that deserve considerable discussion: (1) discount-rate influences that reflect both market risk and the opportunity cost of capital and (2) factors that shape property investors’ expectations of income growth.
The nominal discount rate has two central components: the risk-free rate, reflected by the rates on Treasurys, since they are deemed a risk-free investment; and a risk component, which depends on the specific investment and its industry or market environment. Since real estate competes with other investments for capital, rises in risk-free rates should motivate investors in risky assets, such as real estate, to require higher returns. In our analysis, we used the 10-year Treasury rate, which worked better than other maturities we tried and is consistent with the conventional view of real estate as a long-term investment vehicle.
The second component of the discount rate is risk tied to the investment being considered. This component is most likely to vary across metropolitan real estate markets. Risk perceptions should be shaped by the long-term structural traits of metropolitan markets, such as size of the market, employment composition, and the historic volatility of the local economy and property market. For example, smaller markets may be considered riskier because of thinner property demand. Smaller markets also have less capacity to absorb new development projects beyond a certain size. A sector’s employment picture should also affect investor risk perceptions; markets with more diversified economies are widely deemed less volatile.
The second set of influences on cap rates has to do with expectations of income growth. In real estate, the leasing structure leads to a significant lag between market changes in rent and the movement of property income. To gauge the future of property income, investors typically form expectations about market rents rather than examine actual asset income. Thus, virtually any measure of expected income growth is based largely on expectations of where market rents are headed. In our analysis, we looked at nominal rental growth, splitting expectations into two components: growth due to economy-wide inflation and expected growth in real (that is, inflation-adjusted) rents.
Inflationary expectations (as a proxy we used the lagged consumer price index) should have a negative impact on capitalization rates, because expectations of higher inflation and, hence, higher nominal rent growth would motivate investors to accept a lower return when acquiring a property.
In trying to determine the expected rate of growth of real market rents, it is important to develop some idea of how market rents move through time. There is evidence clearly suggesting that real estate markets are not random walks, and that prices and rents revert to the mean, making them highly predictable over time. There are two simple rules for rational prediction.
First, when the market is at or near historic highs, it is more likely, as one goes further out, that prices and rents will turn down and revert to the mean. The opposite holds true near historic lows. With rational expectations, forward-looking appraisers should anticipate lower income growth when the market is relatively high, leading to higher capitalization rates. On the other hand, there is always the possibility that naive expectations could prevail, with appraisers assuming that when the market is up, growth will continue even further. Under those assumptions, higher relative real rents would lead to lower capitalization rates.
Second, the change in rents does indicate whether future rents will be higher or lower. Thus, rent differences do predict future rent growth; in periods of high rent growth we expect cap rates to be lower.
Our findings, summarized in the graphs on this page, suggest among other things that appraisal-based valuations do move with the risk-free rate, as would be true for valuations in publicly traded markets, indicating that real estate investors do require a higher return when the risk-free rate rises. The impact, however, is less than what one might expect. A 100-basis-point, or one-percentage-point, rise in the interest rate does not result in a similar rise in the cap rate. Instead, the increase would generate a rise of only about 25 basis points in the real estate capitalization rate. Now this might mean simply that investors do not necessarily expect a yearly change in the government risk-free yield to become a permanent change in the cost of capital. Yet clearly, appraisals do not react as closely to risk-free rates as, for example, bonds do. More generally, we learned that appraisal-based valuations react with a lag to current market conditions, with varying degrees of magnitude. We find that an increase in expected economy-wide inflation of 1 percent annually lowers the capitalization rate by 46 basis points. A 1 percent increase in last year’s rate of real office rental growth reduces the capitalization rate by only 9 basis points. Perhaps appraisers regard yearly rental growth rates as transitory.
Our findings strongly suggest that investors use a myopic (as opposed to a forward-looking) approach in forming their expectations of market movements. It appears that during periods characterized by historically high real rents, capitalization rates are lower, implying that investors expect continued high income growth. The opposite holds as well: When the market is low, growth expectations are also low. There is simply little evidence that appraisal valuations recognize any degree of market mean-reversion.
Historically, real office rents have varied between their highs and lows by about 50 percent. Again, a 10 percent increase in real rent levels generates a drop in capitalization rates of only 56 basis points. Thus, the typical market sample’s swing in office rents results in capitalization rates that move by slightly less than 3 full percentage points.
Our calculations reveal a broad similarity in the effect the factors we have discussed have on various property types. For example, a 1 percent increase in the inflation rate will cause a 40-basis-point decrease in multihousing capitalization rates, a 54-basis-point decrease in retail capitalization rates, but only a 20-basis-point decrease in industrial cap rates—all compared with the 46-basis-point drop for office cap rates. Some of these differences might be explained by differences in how each property type’s income responds to changes in market rental growth. For example, the income of industrial properties might well be the least exposed to market changes because industrial leases roll less frequently than other commercial leases. It also could mean that investors always place a high risk premium on industrial properties, with little regard to what is happening in the local market, because they are least familiar with this property type.
Our analysis points to a bigger difference in cap rate sensitivity to changes in rent levels, particularly between multihousing and retail. A 10 percent rise in real rents will pare multihousing capitalization rates down by 87 basis points, but it only reduces retail capitalization rates by 14 basis points. Office and industrial capitalization rates respond in between—56 basis points for office and 58 basis points for industrial. A possible explanation for the unusually large impact on multihousing and small impact on retail properties again focuses on the role of leases in buffering property income from market rent movements. In retail properties, who’s the tenant is often as important as the movement in market rents. By contrast, with annual leases and much less importance attached to who the tenant is, multihousing income moves closely with market rents.
Our findings point to three major conclusions. First, capitalization rate levels exhibit persistent differences across markets as a result of variations in market characteristics that influence investor perceptions of risk and income growth expectations.
Second, movements in market-specific capitalization rates strongly incorporate components that are shaped by the behavior of the local market and, more specifically, by the rate of rental growth and rent levels relative to their historical averages. The nature of these relationships suggests that appraisal-based valuation is more backward- than forward-looking.
Third, cap rate movements in specific markets also incorporate common national influences from the capital markets in the form of interest rates and from the economy in the form of expected general inflation.
Overall, the results of our study indicate that movements in market-specific capitalization rates have strong predictable components that allow the development of econometric forecasts (contingent on forecasts of critical local market indicators along with forecasts of interest rates and inflation).
Further research could address the issues of variation of capitalization rates across submarkets within the same metropolitan area or, alternatively, between suburban vs. downtown locations. Additionally, the sources of the sluggish adjustment in capitalization rates across and within metropolitan markets need to be investigated. Examination of these issues would further advance our understanding of the idiosyncrasies of appraisal-based pricing in private markets—a welcome prospect for real estate investors.
Petros Sivitanides is a senior economist at Torto Wheaton Research (www.tortowheatonresearch.com), a commercial real estate forecasting and consulting firm in Boston. Chief Economist Jon Southard is manager of the research group. Raymond G. Torto is a principal and the managing director. William C. Wheaton is a principal and a professor at the Massachusetts Institute of Technology. This article was adapted from an article published in Real Estate Finance, Summer 2001 issue.
Print Friendly Version