by Mike Kirby, Jon Fosheim, and Warner Griswold

There are two credible schools of thought on the best way to value the shares of real estate investment trusts. One school holds that REITs are no different than any other public company and should therefore be valued with a dividend discount or discounted cash flow model, in which the focus is on the current cash flow or dividend-generating ability of a given REIT and its growth prospects.

While this technique is always an appropriate way to value the equity of any company, many tend to oversimplify its use, and the results tend to be heavily biased in favor of companies with either high initial yields or healthy near-term cash flow growth. The shortcomings are often found in the form of problems inherent in incorporating longer-term growth prospects or, even more importantly, different levels of risk into the model.

Risk can be found at the property level, within the corporate structure, or on the balance sheet, and discounted cash flow models often do a poor job of explicitly addressing either long-term growth or risk issues.

Ways in which a dividend discount model can be “fooled” include: (1) Credit is often not given for real estate portfolios that offer better-than-average growth or lower-than-average risk, and vice versa; (2) such models may give insufficient (or no) credit for projects under development or land being held for future development; (3) a REIT that utilizes leverage can be wrongly rewarded for the FFO-boosting impact of leverage; and (4) structural differences at the corporate level, including conflicts of interest, are often not explicitly addressed.

While a discounted cash flow model, properly conceived and utilized, could address these issues, we find a different approach to be more intuitively appealing. [In fact, within the next several months, Green Street will publish a discounted cash flow (DCF) pricing model, which will be used as a supplement to its NAV-centric pricing model.]

Green Street’s pricing model breaks the pricing equation into two major components. The first of these is NAV—derived by marking to market the value of the real estate and other balance sheet items. The second is the value added or subtracted due to the ability of management working in the current business environment to create value, and/or structural or balance sheet features that may detract from value.

To an extent, an assumption is made that a REIT is not unlike a closed-end mutual fund where share values are closely tied to NAV, but where high costs, conflicts, or managerial incompetence cause funds to trade at discounts to NAV, and where solid management sometimes warrants premium pricing. While it may be a bit of a stretch to take this analogy too far because REITs are generally better positioned to “add value” than a typical closed-end fund, it is certainly true that REIT prices should be more closely tied to NAV than is true of a typical operating company.

While considerable attention must be de-voted to the non-real estate determinants of value, at the end of the day, a REIT consists primarily of a collection of physical assets whose values can be determined with relative ease and precision.

A discounted cash flow model is clearly the better way to value the shares of, say, a carmaker, software firm, or retailer; but a value-based model may well be superior for a REIT.

Critics of the NAV-based approach often gripe that “investors don’t use NAV to value Wal-Mart, Microsoft, or GM, so why should they look at it with REITs?” Our answer is that if they could, they would. REITs lend themselves well to an analysis that delineates between the value of the assets/equity in place today vs. the value of future external growth prospects.

The Economic Value Added (EVA) approach to valuation, currently in vogue, is an academically “pure” technique that starts with the basic premise that a firm should be worth the value of its capital in place, plus the value of its future growth prospects. In most industries, determining the value of the capital in place is no easy exercise. Analysts utilizing an EVA-based methodology are forced to look at some flawed proxy, such as book value, for the value of the capital in place. What a wonderful blessing for REIT analysts/investors that the value of the existing capital base is relatively easy to quantify.

Other objections to a heavy reliance on NAV are that “NAV is backward-looking” or “NAV estimates by different analysts vary too much to be of use.” These criticisms are misguided. In reality, NAV is just as forward-looking as any stock valuation methodology, and one must remember that it is a valuation estimate, not a precisely defined number. In order to properly estimate NAV, an analyst needs to form an opinion on the risk/reward profile of a given real estate portfolio, and then compare that profile with the pricing dynamics (e.g., cap rates, internal rates of return, price per square foot, etc.) of the numerous real estate market transactions that are constantly occurring. Since real estate pricing is presumably determined by parties who, like stock investors, will pay more for better growth or lower risk, the market comps implicitly contain an assessment of how future growth/risk should be priced.

Of course, a properly implemented cash-flow-based pricing model should generate the exact same results as a well-implemented value-based model. We prefer our method because the model breaks the pricing equation into a number of separate steps, each of which represents a variable that has material pricing ramifications for a REIT.

By addressing qualitative issues with regard to the real estate first, the model explicitly addresses an area that arguably does not receive enough attention from many analysts. After taking into ac-count the growth and risk aspects of a given company’s real estate portfolio, the question of how much value manage-ment adds, or how much value is impaired due to structural features, can be address-ed separately.

It should be noted we do not utilize our pricing model in evaluating all REITs. For healthcare and other triple-net-lease REITs, we have found that a model based on AFFO yield works better, which may be attributable to the fact that they, to a greater degree than other REITs, tend to be financing vehicles.

Also, where a company’s external growth story accounts for a very large portion of its overall valuation, cash-flow-based pricing models probably work better. In the REIT boom of 1996-1997, several REITs (e.g., Crescent and Vornado) received a great deal of credit in the marketplace for the value they stood to create through future external growth. In hindsight, we now know that investors should have kept their values more closely aligned with the value of the existing capital base. But if and when growth prospects are extremely lucrative, our NAV-based pricing model is not as helpful as in a more normal growth environment.

Implementing the Model

The first step in implementing the model is to derive an estimate of the marked-to-market value of the real estate portfolio for a given REIT, which affords the ability to calculate NAV. We typically adhere to the approach of applying an appropriate cap rate to the real estate net operating income, or NOI, generated by the portfolio. As such, the choice of an appropriate cap rate is critical, and it is a function of a number of factors, some of which are: (1) property type; (2) location; (3) quality and age of properties; (4) expected demand growth for the property type; (5) local employment growth; (6) the difficulty or ease of replication; and (7) the foreseeable outlook for new supply of that property type. While the list could go on, very simply, the cap rate is a function of initial yield requirements in the capital markets (as measured by, say, the 10-year Treasury bond); cash flow growth expectations at the property/property type; and the risk profile of a given property/property type.

We rank the property portfolios of the various REITs based on what we call economic cap rates. Economic cap rates equate to the unleveraged yield on a property after deducting a normal-ized reserve for recurring capitalized maintenance and re-leasing costs (capitalized expenditures, or cap ex) from real estate NOI. Because of this adjustment, the economic cap rate is always lower than the more commonly touted (nominal) cap rate, and the difference between the two numbers is largest where cap ex items are sizable. (For more on cap ex, see the story on page 50.) Since cap ex items represent a very real cost and are, for all intents and purposes, akin to a current period expense, the economic cap rate is a better gauge of the true unleveraged yield. It is also a cap rate that allows for an appropriate comparison across property sectors (some property types warrant bigger cap ex reserves than others, making a comparison utilizing nominal cap rates a fruitless exercise), among the different REIT portfolios, as well as against other financial instruments, such as Treasury bonds.

In addition to the analysis we do regarding portfolio growth rates, our cap rate rankings are based on information derived in our visits to the properties and interviews with local brokers. We often utilize price-per-square-foot figures and/ or price vs. replacement cost as a cross check for our cap rate selections. These alternative benchmarks have particular relevance when it comes to valuing properties—such as office buildings—that have long-term leases that may, for a limited time, be well above or below market.

The implicit NOI growth rate (based on in-place rents vs. market rents) present in a given portfolio can vary substantially, and this factor can weigh heavily on our final valuation estimates. Once an appropriate cap rate is utilized to value the real estate portfolio of a given REIT, that valuation can be inserted, with a couple of other adjustments, into the most recent balance sheet to derive NAV. The other adjustments include marking to market any below- or above-market debt, ascribing value to property management or service businesses, as well as eliminating intangible assets, such as capitalized leasing commissions, or financing costs.

The estimate of NAV is, importantly, only a starting point in determining the appropriate share price for a particular REIT. As we have noted frequently over the years since we introduced our pricing model, there are a number of variables that impact the magnitude of any premium or discount to NAV at which a given REIT should trade.

Model Variables

We initially identified the variables that comprise our pricing model by attempting to answer a simple question: Why do some REITs trade at premiums to NAV while others trade at significant discounts? Our answer to this question consists of seven variables outlined below. The model variables are:

Franchise Value—This is both the most important variable in the model and the most subjective. Put simply, fran-chise value is the ability of a management team to create value over and above the current value of the company’s existing portfolio. As the REIT industry has been transformed into an industry where the majority of companies are fully integrated and self-administered, the ability of REITs, as a group, to add value has improved. That said, the unimpressive capital allocation track record (see “Tough Choices,” Property, January/ February 2001, page 63) demonstrated by many REITs has served to partially offset the otherwise strong gains made on this front. It may be easiest to summarize franchise value by stating that a REIT that does not have it should never trade at a premium to NAV. While management talent and ability are very important determinants of franchise value, they are not the only ones. In fact, franchise value is probably more precisely defined as the ability of a given REIT to create value (which may or may not coincide with growing FFO, AFFO, or NAV) through intelligent capital allocation. This may be in the form of acquisitions or development (remember, internal growth at the property level is already taken into consideration in our cap rate assumptions) or, during a discounted pricing environment, it may be tied to shrewd property sales and/or share buybacks.

Remember, even a good management team operating in a tough environment may warrant a low franchise rating. Franchise value is a function of the following:

Is the company fully integrated and capable of adding value if all other conditions are suitable? Most REITs in our coverage universe now receive an affirmative answer to this question, but there are still a fair number that have shortfalls in this regard. While some REITs lack certain key elements, such as property management skills or the ability to add value at the property level, others fall short because they are “one-trick ponies.” The most obvious example here would be REITs that prospered in the lucrative acquisition environment of recent years but lack other attributes crucial to success now that the real estate cycle has matured (e.g., development and redevelopment abilities).

At least as important as whether a REIT is well-integrated is the job it does as a capital allocator. A REIT with demonstrated acumen at knowing when to “open the floodgates” on the acquisition or development front vs. those times when either real estate or capital market constraints make a slower pace the more prudent course warrants a high franchise rating. Numerous REITs that looked brilliant for buying lots of property during the mid-1990s buyers’ market are now smarting because they continued to buy long after pricing became rich. Of increasing importance is a demonstrated ability to add value through a self-funded growth program, or, when appropriate, to engage in a partial liquidation (sell properties/buy-in shares) to capitalize on the fact that real estate is worth more on Main Street than on Wall Street.

Is the company in a property sector that offers lucrative opportunities? As real estate markets recovered during the 1990s, some did so ahead of others. For example, apartment acquisitions were a great story early in the decade, followed several years later by the attractive acquisitions that could be made just ahead of the office market rebound. During the “sweet spots” in the cycle, apart-ment REITs, and later office REITs, were in a much better position to add value than other companies, and, therefore, were accorded much higher franchise values. It is now fair to say, however, that just about all real estate markets have recovered to an equilibrium point, and it is difficult to argue that there is a huge advantage on the external growth front to operating in one property sector vs. another.

REITs with stronger-than-average balance sheets are better positioned to grow profitably, regardless of conditions in the capital markets. By contrast, a highly leveraged balance sheet can serve as an impediment to growth if equity markets are temporarily unfriendly. Thus, while a strong balance sheet by itself cannot serve as a vehicle for creating value, a weak one can certainly impair a company’s ability to do so.

While REITs with high franchise value typically exhibit solid growth in FFO, AFFO, and NAV, the mere ability to do deals accretive to these benchmarks does not translate directly into a high franchise rating. High multiple REITs can often buy just about any piece of real estate in an accretive manner, but this does not necessarily mean the company is adding value for shareholders. Rather than focusing on FFO/AFFO or NAV accretion, the ability of a REIT to add value can only be measured by looking at whether it is doing deals that have a positive net present value (NPV). This is a particularly important point as it pertains to developers; while just about every development project is accretive, a far smaller number generate returns in excess of appropriate risk-adjusted hurdle rates.

Focus—Over the years, certain REITs and a number of real estate companies have attempted to diversify their portfolios across large geographic regions and/or a number of different property types. Historically, this has been done in a poorly conceived manner because a company did business in a lot of markets where it had no real expertise. Common sense tells us this approach is flawed, and this has generally been confirmed by the poor performance of most REITs that have tried this approach.

Real estate is, and always will be, a local game in which a “local sharpshooter” will outperform less knowledgeable outsiders. For this reason, specialized REITs have long outperformed their diversified brethren. While diversification is important, investors can efficiently accomplish this goal by investing in a diversified portfolio of focused REITs. Therefore, geographic and property type diversification at the REIT level adds little value at the company level.

While local expertise will always remain important, some of today’s better REITs have become geographically diversified; yet by building critical mass in given locales, they have been able to obtain many of the benefits of being a local player. By way of example, Equity Residential owns as many apartments in some of its key markets as the recognized local players, and this affords the company the ability to hire top-notch managers in these markets. While senior management still does not have the local knowledge of each market present in the more focused companies, the company gives up nothing (and may even be better) at the operating level. For this reason, the rating on geographic focus also is impacted by the level of expertise and depth that a company has in most of its markets, as opposed to merely being a reflection of how many markets it is in.

While geographically dispersed REITs with strong local market knowledge may warrant a high mark on the focus rating, there is no reason to believe similar benefits are present for REITs operating in several different property types. Diversification across property sectors may not be a serious negative if well-qualified people are involved to oversee the different sectors, but it is hard to argue that it is a material positive.

Insider Ownership—The logic of why high insider ownership is desirable is obvious. The large number of old line, entrepreneurial real estate companies that have become REITs in recent years has increased the average on this variable to a very impressive 13.0 percent, and the median to 11.1 percent. While these figures have been trending downward over the last few years (see graph on this page), they represent a large increase from the averages of several years ago, and it is almost certain that the REIT industry stacks up very favorably vs. most other public companies on this score.

Balance Sheet Strength—REITs that leave themselves some flexibility to finance their growth with debt are in a much better position than highly levered REITs when the equity window closes. While this line of thinking may run somewhat contrary to academic theory (theory suggests that investors should be indifferent to leverage for a non-taxpaying entity such as a REIT), it is generally true that REITs with lower leverage and solid balance sheets tend to trade at higher premiums/lower discounts than their more highly levered peers.

We measure balance sheet strength with three distinct gauges. Total leverage is measured by dividing liabilities (including preferred stock, which common shareholders should view as more akin to debt than equity) by the estimated value of the assets derived in our NAV calculation. By using the estimated value of the assets in the denominator—as opposed to the more widely cited, but flawed, method of using market capitalization—our measure is not affected by changes in the share price.

Exposure to variable rate debt is measured by dividing the amount of unhedged (including out-of-the-money hedges) variable rate debt (net of offsetting cash balances) by the value of the assets. Finally, exposure to debt maturity risk is measured by dividing the net amount of short-term debt (defined as cash balances less debt maturing within the next two years) by the value of the assets.

Taken together, these three ratios provide a meaningful snapshot of the health of a REIT’s balance sheet at any point in time. As of mid-August 2001, the averages for each of these ratios were: liabilities as a percent of assets, 52.1 percent (see graph on page 44); variable rate debt as a percent of assets, 10.4 percent; and short-term debt as a percent of assets, 4.2 percent. Total debt is roughly similar to last year’s level but has risen by a wide margin since the 31 percent level of just a few years ago.

Overhead Expense—Measuring overhead for REITs is admittedly an iffy proposition. Accounting conventions are vague as to which expenses constitute general and administrative costs, and because of this, it is difficult to make an apples-to-apples comparison between companies. Nevertheless, REITs with high overhead deserve a larger discount/lower premium to NAV than REITs that are run lean. Additionally, it is necessary to include this variable in the pricing model in order to “catch” REITs that inflate real estate NOI (and, therefore, estimated NAV) by categorizing an undue amount of property level expenses as corporate overhead instead of property costs.

As newer, larger REITs have entered the industry, average overhead figures have dropped over the years, and the average for our coverage universe as of mid-August 2001 was 0.54 percent of our estimated asset value (see graph on page 46). This is slightly higher than last year’s figure, and down from 0.73 percent in 1992.

Potential Conflicts of Interest—Conflicts of interest have been among the root causes of some of the largest debacles in the history of the REIT industry, making REIT investors particularly sensitive about this issue. Fortunately, the relatively recent “explosion” of new REITs has generally included REITs with high insider ownership stakes and a number of investor-friendly structural safeguards. Nevertheless, potential conflicts remain plentiful.

Among the more troublesome conflicts are instances where insiders retain the right to sell properties to a REIT, a feature that is quite common, though in all but a few cases not overly material. Also, some features of the newer REIT struc-tures, such as the leasing of properties to entities owned by management (a common structure among hotel REITs), are nothing but rehashings of fundamentally flawed structures that have failed in the past.

An additional—and increasingly important—area of concern pertains to corporate democracy issues and anti-takeover provisions in bylaws. Most of the newer REITs have chosen to incorporate in states such as Maryland, where anti-takeover statutes are little more than management-entrenchment provisions. Additionally, many of the umbrella partnership REITs, or UPREITs, contain clauses that give insiders veto power over major asset dispositions, mergers, etc., even if a majority of shareholders and the board votes in favor of such a move. Where REIT insiders feel compelled to entrench themselves with such devices, it will have a negative impact on share pricing.

Liquidity—Our own past studies indicate that larger, more liquid REITs tend to trade at larger premiums to NAV. This is consistent with findings throughout the broad stock market, where larger companies tend to have lower long-term costs of equity capital (i.e., all else equal, they have higher stock price multiples). Whether investors are paying up for size or liquidity is difficult to ascertain, but since the two almost always go hand-in-hand, the answer may be of only theoretical interest. We measure liquidity based on trading volume over the prior 12 months.

Computing Warranted Share Price

Perhaps the most important feature a user of our model should be aware of is that it is a relative pricing model. In other words, it is designed, at any point in time, to show an equal number of underpriced REITs vs. overpriced REITs, and as such, it will never provide an indication as to when REITs, as a group, are mispriced.

A common misperception is that, because it is based on NAV, the model is biased against companies trading at large premiums to peer group averages. This is not valid as the model currently suggests that some REITs should trade at premiums to NAV of around 20 percent despite the current (mid-August 2001) average premium to NAV of 3 percent for our coverage universe.

The first step in determining an appropriate premium or discount is a ranking of the REITs, relative to each other, on each of the seven variables in our pricing model. The variables are weighted according to their importance in the model, and each REIT’s score determines its position in our relative rankings of the REITs. The output at this point effectively amounts to a ranking of the REITs, as companies, from “best” to “worst,” at least in terms of their current ability to add value for shareholders.

The model works by taking a statistical snapshot of the REIT industry at a given point in time; i.e., what is the average premium or discount to NAV, and how widely (as measured in terms of standard deviation) are REIT prices spread out vs. this average? In other words, at a given point in time, an average REIT may be trading at a 5 percent discount to NAV, but some may be trading at premiums as high as 20 percent, while others are trading at discounts of 30 percent. This statistical “description” (a normal distribution is assumed) of the market is then overlaid onto the results of the pricing model rankings to determine where each REIT should be trading. For example, if the most richly priced REITs at any moment are being awarded 10 percent premiums to NAV, the model will award a warranted premium of roughly that magnitude to the companies that score the best on our model.

Once appropriate premiums/discounts to NAV are determined for each of the REITs, the premium/discount is applied to our NAV estimate to derive a warranted share price. The warranted share price can then be compared to the existing share price to determine which REITs are overpriced and which are underpriced. The output of the model provides a very useful guide as to where prices should generally be set, although it is important to note that we override the model from time to time in making our Buy/Sell or Hold recommendations. For example, if we are anticipating a price-changing piece of news (e.g., a dividend cut, quarterly earnings surprise, etc.) on a company that might otherwise warrant a higher/lower rating, we might delay implementing any recommendation until after the market has digested the news.

A Glossary of Terms Commonly Used by Green Street

Adjusted Funds From Operations—Equates to FFO less a normalized reserve for capitalized leasing and maintenance costs, less straightlined rent, less gains on property sales. Many analysts subtract only a reserve for “cap ex,” and our figure therefore differs in its treatment of straightlined rents and gains on sales. We feel strongly that neither of these items belongs in a performance measure, and therefore feel it is appropriate to subtract them in computing AFFO. AFFO is a much better performance measure than FFO and should serve as the foundation for any performance multiple or yield analyses. As evidence of how materially “wrong” FFO can be, as of late January 2001, FFO exceeded AFFO by an average of 19 percent, and for 20 of the 68 companies in our coverage universe, it overstated performance by more than 25 percent. Cap Ex—Short for recurring capitalized expenditures related to maintenance and re-leasing activities. Examples would be virtually all tenant improvement dollars (except those associated with an expansion), capitalized maintenance costs, and other capitalized costs associated with turning over a tenant. A normalized “cap ex” reserve should be deducted from funds from operations to better gauge cash flow (see Adjusted Funds From Operations), as well as from real estate net operating income (NOI) prior to computing current value. While many other analysts now embrace this concept, certain of these analysts are guilty of generalizing by applying a standardized reserve. This is not appropriate because each REIT’s accounting policies are different and therefore have an impact on the appropriate reserve; an appropriate reserve can thus be determined only by looking at each REIT separately. Cap Rate—The initial, unleveraged yield on a property prior to deducting cap ex costs from the return. It is calculated by taking the estimated net operating income (NOI) for the next 12 months divided by the all-in (including commissions and closing costs) purchase price. The cap rate is a function of both the risk profile and the growth prospects for a given property. Economic Cap Rate—The cap rate that is applied to NOI after deducting a cap ex reserve, as opposed to the more traditional calculation of just applying a cap rate to NOI. It is a better gauge of the true yield provided by a property, since it implicitly adjusts for the very real expense of cap ex. It is also a yield number that can better be compared to other capital market yield measures such as bond yields. Implicit Cap Rate—The cap rate that would have to be applied to the real estate NOI to derive an NAV that is equal to the current share price. Warranted Premium/Discount to NAV—The output of our REIT pricing model is the “appropriate” premium or discount to NAV that should be applied to determine a warranted share price. Warranted Share Price—The “appropriate” share price indicated by our REIT pricing model. As a general rule that we sometimes break, we never recommend purchase of a stock unless it is trading at a discount of at least 5 percent to this level (10 percent for an IPO).

Online Version