Analyst's Notebook
The Power of Two
Using discounted cash flow analysis alongside its NAV-based pricing model, Green Street Advisors gains added insight into a real estate stock’s risk and long-term growth prospects.

by Mike Kirby, Warner Griswold, and Jon Fosheim

Arguably the best way to value any stock is discounted cash flow analysis. DCF, in a nutshell, applies appropriate risk-adjusted hurdle rates to estimates of all current and future cash flows generated by a business. By taking into account the combination of in-place cash flow, near-term and long-term growth prospects, and the risk in a company’s business lines, DCF represents a thorough and insightful means to value a company.

However, as is true with any valuation model, the devil is in the details. A poorly implemented DCF model will be of no more relevance than a Ouija board, whereas even the simplest metrics can be useful in the hands of a knowledgeable and thoughtful investor. The theoretical underpinning for DCF is solid, but it is deceptively easy to run into a garbage-in/garbage-out problem. Perhaps the biggest shortcoming is that small changes in long-term growth rate assumptions or estimated discount rates can have huge ramifications on share value, especially for a leveraged company. For this reason, users of a DCF model must ensure that these inputs have been derived in a thoughtful and systematic manner.

To value shares of real estate investment trusts, Green Street Advisors has relied for the past 12 years on a pricing model based on net asset value (see “Puzzle Palace,” Property, Sept./Oct. 2001, page 42). This model has proven to be very helpful in valuing shares of the companies we follow and has been a big contributor to our successful long-term stock-picking record. [To view Green Street’s stock-picking record, point your browser to] Based on the simple logic of “if it ain’t broke, don’t fix it,” our NAV-based model will continue to serve as our primary valuation metric.

Yet it’s also worth observing that though a meteorologist can predict the weather using either satellite pictures or radar images, he can do a better job using both in tandem. Therefore, we have developed a DCF approach for use with our NAV-based model. A DCF model can provide insight that is not readily apparent from relying solely on an NAV model, and the converse is also true. For the time being, our firm will place more emphasis on the output of our tried and tested NAV-based model, but as we become more comfortable with the DCF approach, we will likely give it equal weight in our stock selection process.

The primary reason we always have put a lot of weight on NAV is that real estate matters and it matters a lot. But the DCF model relies just as much on the risk and growth attributes of real estate portfolios as our NAV-based approach does. The beauty of an NAV-based approach is that it compartmentalizes the valuation analysis, and, through the information contained in real estate prices, specifically addresses the risk and long-term growth prospects of a real estate portfolio. By first answering the question, “What is the value of what the REIT now owns?” an NAV-based approach allows investors to address separately the question, “How much value will be created or destroyed by the REIT’s growth strategy?” By addressing one question at a time, our NAV-based model does a good job of separating the two biggest determinants of a REIT’s value.

Another key attribute of an NAV-based model is that it relies on prices for real estate in the private market to provide guidance on appropriate pricing in the public market. For the most part, this is a positive attribute, as private market prices contain the collective opinions of thousands of buyers and sellers about the growth prospects and risk of a given type of real estate in a given locale. More often than not these private markets work efficiently, and public-market investors who choose to ignore the information conveyed by thousands of willing buyers and sellers are foolhardy. Securities analysts in other industries tend not to focus on NAV, but this is because most corporations own assets that represent a means to an end, such as plants, machinery, and inventory. REIT analysts, by contrast, are fortunate to work in a sector where the great bulk of the valuation of a company is composed of commodity-like assets. A viable and active private market constantly reevaluating risks and rewards in those assets provides a valuable starting point for determining the value of any REIT.

A final argument for focusing on NAV is that pricing in the REIT sector has a long-term tendency to revert back to NAV parity. The average discount/premium to NAV for our coverage universe for the past 11 years has been roughly 0 percent. The historical evidence suggests that investors believe REITs neither create nor destroy value, in the aggregate.

Yet conceptually, DCF is superior to any valuation technique. All determinants of value are boiled down to quantifiable inputs, and the process of coming up with some of the inputs forces the user to explicitly address tricky issues, such as risk and long-term growth prospects. Some of these factors are conveniently ignored in an NAV-based model based on the premise that private-market pricing benchmarks already incorporate adequate assumptions on issues such as long-term growth prospects for a real estate portfolio and/or the risk of that portfolio. Though it is normally better to ascribe the benefit of the doubt to the collective wisdom reflected in private-market pricing, the fact that a DCF model forces a user to address these issues explicitly results in added insight.

(Download Output of the DCF Model vs. NAV-Based Model)

An additional attribute of a DCF model is that it allows the user to better understand each of the determinants of value. By understanding the implicit assumptions, the analyst is better positioned to form an opinion as to whether private-market pricing makes sense. If, for example, mall capitalization rates were 6 percent in the private market, at a time when office cap rates were 10 percent (this was actually true not long ago), the differential would be justified only if malls were much less risky, offered superior long-term growth, or both. A DCF model forces the user to consider just how much less risk or additional growth is necessary to justify prevailing cap rates. By gaining a better understanding of the assumptions underlying private-market pricing, the user is better positioned to second-guess the private market when appropriate.

Finally, a DCF model can more objectively address external growth issues than an NAV-based model can. Our NAV-based model uses subjective assessments of franchise value and other non-NAV determinants of share value, but a DCF model picks up “franchise value” in the form of earnings growth prospects that either exceed or fall short of what would otherwise be expected given certain structural attributes, such as dividend payout policies and leverage.

Which model is better? Neither. In fact, most of the time a well-implemented DCF model should generate roughly the same output as a well-implemented NAV model. While the relevant valuation inputs are addressed in a different order and sometimes in a different format, the same inputs serve to determine value no matter which model one uses.

Our DCF approach is technically a dividend-discount model. But our dividend-discount model will almost always assume a static dividend payout policy. While we recognize that this assumption runs contrary to the observation that REIT dividend payout ratios are generally falling, this inconsistency should not adversely affect the integrity of the model.

That clarification aside, there are several important inputs that can have a dramatic impact on a DCF model. While some of the inputs, such as this year’s adjusted funds from operations and the outlook for AFFO growth over the next couple of years, are reasonably easy to arrive at based on analyses we already develop, two areas—long-term AFFO growth and appropriate company-specific discount rates—require additional work.

Our DCF-valuation approach is broken into three distinct parts: (1) estimating long-term AFFO growth; (2) deriving the weighted-average cost of capital, cost of equity, and warranted share price; and (3) determining the implied cost of equity and cost of capital as a reality check.

So, what does the model tell us? The table on page 50 compares the warranted prices derived with our DCF model vs. the prices derived with our NAV-based REIT pricing model. The table on page 52 shows the most mispriced REITs. Several observations stand out.


Seven stocks stack up as being overpriced by more than 20 percent, and seven appear to be underpriced by at least 10 percent. At the extreme end, the model suggests that Crown American and Associated Estates are both overpriced by more than 50 percent. The variances between the model’s warranted pricing and actual prices are, on average, much larger than those suggested by our NAV-based pricing model. This is primarily attributable to differences in fundamental assumptions underlying the two models. While we are comfortable with an assumption underlying our NAV-based model that REIT prices follow a normal distribution pattern relative to NAV, the DCF model is not bound by such an assumption. More important, the DCF model suggests that the pricing variances that are observed in the REIT world are not as large as they should be (a conclusion our NAV-based model is constrained, by design, from reaching). In other words, low-quality companies are generally overpriced at this time, while high-quality companies tend to be underpriced. Our NAV-based model confines the range of warranted prices to roughly approximate observed price patterns, but if the market gets this wrong, the DCF model will do a better job of picking it up.

One of the most basic and best-accepted principles of modern finance is that higher returns associated with higher leverage should not result in higher share prices. Instead, investors will demand a higher rate of return to compensate for the added volatility associated with leverage, and this higher return expectation entirely offsets the benefit of increased cash flow. But REIT investors currently appear to be ignoring this tenet. Each of the 10 most overpriced REITs (see table above) has higher-than-average leverage, and the average for all 10 companies is about 64 percent, well above the industry average of 51 percent.

There are exceptions, but the most overpriced stocks listed in the table tend to have property portfolios that are below average. These same companies also tend to have high leverage, so it is difficult to distinguish whether investors are wrongfully ignoring real estate quality or risks associated with leverage. It’s likely that both factors are resulting in mispricing.

While investors appear to be making systematic errors in evaluating the growth/risk tradeoff of both leverage and property quality, the newness of this model causes us to be unable to put this in historic perspective. It is easy to conclude that the mispricing we see today is more severe than it has ever been, since many of the overpriced stocks have outperformed industry benchmarks by enormous margins over the last 10 months. But we are unable to draw a conclusion at this early juncture as to whether investors are normally suckered in by the high yields associated with low property quality or high leverage.

In sum, our NAV-based REIT pricing model has served us well for 12 years, and we expect it will continue to do so. Nevertheless, new perspectives that add insight are welcome. Our new DCF-based model is not intended to represent a better way of valuing REIT shares; it is merely a different way of addressing the value question. From now on, we will rely heavily on both satellite pictures and radar images.

Mike Kirby and Jon Fosheim are cofounders and principals of Newport Beach, California-based Green Street Advisors, a buy-side research boutique specializing in property stocks. Warner Griswold is a senior analyst with the firm.

Print Friendly Version