by Barry Vinocur
Photographs by Dave Bergeland

Pension funds have long accepted the notion that real estate is a separate asset class. In recent years, however, they have been slow to embrace real estate investment trusts as a proxy for the direct ownership of real estate. It’s a curious thing. In the early to mid-1990s, many pension funds abandoned the direct ownership of real estate via vehicles such as commingled funds and group trusts because they didn’t offer "real" liquidity. Put simply, by the time the large institutions realized that property markets in the United States were a shambles, the redemption queues of the funds that they were invested in were so long that getting out in a timely fashion was impossible for most investors.

For a time at least, it seemed as if publicly listed REITs were made to order. Not only did they offer liquidity and real-time pricing, which was preferred over annual appraisals, but REITs also offered pension funds the opportunity to invest alongside some of the real estate industry’s major players, who in late 1992 began taking their real estate empires public. From office buildings to regional malls to industrial properties, the REIT market offered it all.

As so often happens, however, institutions and many financial advisors were slow to pull the trigger. By the time they did, or were close to doing so, REIT prices began to plummet. Real-time pricing no longer seemed the boon that many had thought it was. Worse still, the volatility in REIT prices made some investors question whether there was something different - very different, in fact - about real estate in a public wrapper.

When it comes to understanding real estate returns, or looking at such things as the correlation of direct property investments vs. stocks, bonds, and public REITs, no one has spent more time analyzing the data, or is more highly regarded than S. Michael Giliberto.

Giliberto, who earned a Ph.D. in finance, has worked at Salomon Brothers (prior to its acquisition by Smith Barney) and at Lehman Brothers. Currently, he heads the real estate research for J.P. Morgan Investment Management in New York.

Over roughly the past decade, Giliberto has either authored or co-authored (frequently with his wife, Anne Mengden, a former REIT analyst and currently a portfolio manager at European Investors in New York) nearly every one of the most-often-quoted papers and presentations on the subject of real estate returns vis-à-vis their correlations with stocks, bonds, and REITs. In late November 1999, Giliberto presented his findings to attendees at a conference - "A New Era for Real Estate Investing" - sponsored by the Association for Investment Management and Research (AIMR).

How do you view real estate?

We look at it as having features of both fixed income and equity.

Is it a separate asset class?

We think it is, but not solely because it’s a hybrid between two recognized asset classes. Another defining characteristic is the correlation of returns vis-à-vis financial assets. Even after adjusting for the fact that stocks and bonds, and even REITs, are publicly traded and real estate isn’t per se, the returns aren’t correlated. Further, real estate cash flows aren’t in sync with cash flows from, say, corporate America or corporate earnings. All of those factors, as we see it, make real estate its own asset class.

You’re asking a series of questions?

That’s right. First, how does one earn the return - what combination of capital appreciation and income? What’s the correlation with broadly accepted asset classes? And then is there something fundamental to lead us to believe that the correlation normally will be relatively low, suggesting there’s a diversification benefit?

There are those who agree that real estate is a separate asset class, but they argue when it’s put inside a public wrapper that somehow changes it.

The wrapper doesn’t have anything to do with determining the cash flows generated by the asset. Whether you own a shopping center outright or you own a REIT that owns shopping centers, all of those shopping centers are operating in the same physical or space market. So when you’re looking at operating earnings - from a macro point of view, which we did - they’re pretty highly correlated, which, by the way, is what common sense tells you ought to be the case.

Nevertheless, the wrapper does appear to make a difference. Is it a real difference or is it an illusion? Here’s where we run into the problem of using very high frequency data, which we certainly have in REITland; we can look at it moment by moment. On the other hand, we don’t take the temperature of direct real estate nearly that often. Because we have one that’s a publicly traded entity and one that’s traded infrequently and privately, you can be fooled by trying to do correlations month-to-month, or even quarter-to-quarter, into thinking the two aren’t related or are not related enough to be regarded as the same thing.

We find that looking over longer periods helps sort that out. For instance, the five-year returns from private real estate and the five-year returns from REITs are almost identical. If you had invested in either category five years ago, you would have traveled very different paths. But if you were a buy-and-hold investor, you would have ended up with roughly the same rate of return.

What about the correlation between direct real estate and REITs over that time period?

The correlation between the two hasn’t been terribly high over that time period. Again, we believe that’s the result of the path not the returns. In other words, we believe that the public market anticipates; it’s very forward-looking. Private market returns, on the other hand, are much more rearview mirror-oriented. They’re based on appraisals and transactions. No matter how recent the appraisal data or the sales data that an appraiser is using, it’s never perfectly comparable. It’s always got some element of staleness in it, so the appraisals are almost playing catch-up with the realities of the markets, and the REITs are anticipating where the market may be going.

If we look at longer-term return patterns, we see that typically the public market is roughly a year to two ahead of the private market. Again, you have to look at the returns over a multiyear horizon. It’s not just one year.

How are those conclusions impacted, if at all, by the fact that prior to 1993, it was really a much different industry than it is today?

Any careful researcher has to take that into account. You also have to take into account that until 1998, REITs were in a bull market. We haven’t seen what happens to the stocks during a recession, for example. At the same time, you don’t want to simply discard the older data either. The fact that the companies that were around prior to 1993 constituted a much smaller segment of the real estate industry doesn’t change the underlying business those companies were in. It doesn’t invalidate the data. It just means that we perhaps wouldn’t weight the earlier data as heavily as the more recent data.

Not long ago, I received an e-mail from a financial advisor concerning REITs. For the last several years, his group had used REITs to fill its real estate allocation. He now says, "We recently concluded that REITs no longer, if they ever did, represent a real estate allocation. Further, we’re not sure what they do represent and we don’t want an unknown in our portfolio." How would you respond to him?

First, as I mentioned, we like to look over long periods of time. Second, we also prefer to look ahead. Looking forward is very important when you’re doing asset allocation, in part because of the issues that we have discussed with regard to the data. So, while you have to understand what history is telling us, you don’t want to be fighting the last war. Keep in mind as well that a correlation is a statistical attempt to measure some sort of fundamental relationship. What you’re seeing is a statistic that’s one number that encapsulates a lot of economic information, both in the real economy and in the financial economy. What we need to do is to drill down below that simple number and ask what are the patterns that caused this? What was happening in the real estate market, in the economy at large, and in financial markets that leads to these patterns of correlation? Because if we think that the future is going to look different in some way than the past or the constellation of those various factors lines up differently in the future, we’re going to observe a different correlation pattern going forward.

Your findings?

Historically, the correlation between private and public real estate - and we don’t make any adjustments - is very, very low. But if we look at it on a rolling basis, it would seem the correlation between private and public real estate has fluctuated. If we look at these longer horizon returns currently, the correlation between public and private real estate is about 0.6, which is pretty high.

The conventional wisdom, as we discussed, is that real estate has a very low correlation with stocks and bonds. What does your research show?

When people have looked at the correlations between direct real estate, using data from The National Council of Real Estate Investment Fiduciaries (NCREIF), and stocks and bonds that’s the conclusion. But again, you have to keep in mind that you’re mixing a private market measurement - the NCREIF data - with public market measurements. If you take that into account, and we do, we believe the correlation between direct real estate and financial assets - forget about REITs for the moment - is much higher than the historical data would indicate.

The story here is a very simple one. In general, if interest rates go up, asset values go down and vice versa. Real estate is generally going to be beholden to that same fundamental principle, so we really think there’s a positive correlation. It’s modest, but there’s a positive correlation between bonds and direct real estate and, let’s say, large- cap equities and direct real estate.

How does that impact - if at all - the argument that real estate has a positive impact on a mixed asset portfolio?

You don’t need to have zero or negative correlations to have benefits. Look at it this way. Very few people who work with institutional portfolios would advocate being entirely in fixed income or entirely in equities, because the correlation between fixed income and equities is generally perceived to be around 0.3 or 0.4 or somewhere in that range. Those correlations are considered low enough to have huge benefits in terms of diversifying a portfolio and risk reduction.

We have something that we call the real estate framework. It’s a forward-looking projection-based system for doing real estate portfolio design and strategy. Within that we posit similar levels or correlation between direct real estate and stocks and bonds; it’s on the order of 0.2 to 0.3 and a little bit higher for REITs. So, there’s still huge diversification benefits.

It’s also worth mentioning, because it’s frequently misunderstood, that correlation is not what is deemed transitive. In other words, if something is highly correlated with one thing, say A is highly correlated with B and B is highly correlated with C, it doesn’t necessarily follow that A is highly correlated with C.

When institutions ask you about the correlation between REITs and stocks, what do you tell them?

Utilizing our forward-looking framework that I mentioned, we tell investors that the correlation of REITs with stocks is going to center around 0.4, with the normal range being 0.3 to 0.5. Currently, it’s a bit higher than that. If you look at the last couple years of data, it has been near the upper end of our range, and that’s using daily data, which we don’t necessarily advocate. So longer term, we think the 0.3 to 0.5 number will be good. On the other hand, the correlation between direct real estate and stocks is going to be more like 0.2.

What about the volatility of those returns?

If you go back to 1997, which was an up year for both the Standard & Poor’s 500-Stock Index and for REITs, and you used the daily data, REIT volatility was about 9 percent that year and the S&P was about 18 percent. That’s taking the standard deviation of daily returns and then annualizing it.

Normally, we think that REITs are going to have roughly 90 percent of the volatility of the broad market, and that core direct real estate will be about 55 percent of the volatility of the broad market. Those are rule-of-thumb numbers based on our own research and that of others we find credible. Keep in mind that the volatility of direct real estate is on an unleveraged basis. If you add in, say, 35 percent leverage, then the volatility of direct real estate starts to look pretty similar to REIT volatility.

Does it matter which indices you use? On the private side, the gold standard is the NCREIF Index. But on the public side there are a variety of indices; there’s the National Association of Real Estate Investment Trusts indices, the Morgan Stanley REIT Index, the Wilshire indices, etc.

When choosing an index there are several things we believe are important. First, a good public market index is one that an investor can replicate with very little cost, in other words, an index that can be passively managed. Not that we would advocate an investor taking that approach, but if an investor wanted to take that approach, you would want an index that offers a reasonable degree of liquidity in the underlying securities. Second, we believe it’s important to have very good disclosure on how the index is constituted, and that it be what we would call a rule-based index, meaning there’s very strict criteria for when a security goes in or goes out of the index.

Generally speaking, all of the indices that you mentioned meet those criteria. Each, however, has its pluses and minuses. The NAREIT Index, for instance, is broadly based, but that also means that it includes some stocks that are relatively thinly traded, so that’s a drawback. You also want an index that is widely quoted in real time. The NAREIT Index isn’t available in real time except via the NAREIT Web site. Most of our clients want to see the index quoted on the same screen that they see their other quotes on.

Since you brought up passive management, if you look at Vanguard’s REIT fund, which tracks the Morgan Stanley REIT Index, there are quite a few actively managed funds that outperform it every year. Why do you think that is?

Generally speaking, indexing on a purely passive basis is always a viable strategy. The opportunities to outperform the benchmark - and this is sort of a philosophical view - depends on how efficient the market is, whether one really has the ability to get information before it’s broadly disseminated. When you move into the highly liquid large-cap stocks, you would think it becomes harder and harder to do that, although there’s some evidence that it can be done, at least by some people. As you move into the less-efficient segments of the stock market - not that they’re inefficient, but they’re perhaps less informationally efficient - there may still be returns to be gleaned by an investment in information and by creating an information edge. I think REITs can still fall into that category. Over time will that change? It’s likely. We would certainly expect the market for REITs to become more efficient over time, and hence for it to become harder but not impossible to beat a purely passive strategy.