REITs operate like specialized mutual funds that invest in a wide variety of companies that build, own and manage shopping centers, condominiums, office buildings, housing developments, hospitals, parking garages, factories and all sorts of other real estate that makes money.

From 1972 through November 30 of this year, nearly 36 years, REITs compounded at an annual rate of 13.2 percent, vs. 11.2 percent for the Standard & Poor’s 500 Index. (That may not seem like a very big difference, but over this many years it means a lot. For example, $1,000 invested for 36 years at 13.2 percent would grow to $86,790; at 11.2 percent it would grow to only $45,685.)

Although the experts don’t expect such high returns in the future (and they expect less difference in future returns), REITs still deserve a place in your tax-sheltered portfolio. (Because they pay generous dividends that may be taxed more heavily than dividends from most mutual funds, REITs aren’t a good choice for taxable accounts).

The main reason to add REITs to a portfolio is not to improve return, though that would have happened over the past several decades. The main reason is to improve diversification, with the opportunity for higher return. This is what I want to show you in this article.

First, a brief refresher on what I call “smart diversification.” In a nutshell, it’s the combination of two or more investment assets that go up and down at different rates and at different times. This phenomenon allows you to achieve the long-term results of (historically) high-performing asset classes without getting beaten up when those asset classes are underperforming.

(My son, Jeff Merriman-Cohen, has illustrated this concept and described it very well in an article available at FundAdvice.com called “The perfect portfolio.”)

If you invest in two asset classes with identical long-term returns, their zigs and zags can sometimes cancel each other out, giving the combination lower risk than either of the individual components – while still capturing the full long-term returns. This is particularly valuable for retirees, who are best served by relatively low-risk portfolios that produce relatively high returns as they withdraw money.

In Table 1, you’ll see year-by-year results comparing U.S. REITs with the S&P 500 Index. The “difference” column measures the size of the difference in returns of these two asset classes. For purposes of reducing risk a higher difference is better. For example, look at the line for 1977. The stock market was down that year, but REITs were up, for a net difference of nearly 25 percentage points. You’ll see in the table that there were 15 years when the difference in return between these two asset classes was more than 15 percentage points.

(REITs return data used in this article reflect the Dow Jones Wilshire REIT Index.)

Table 1

Click here to open Table 1 in a new window.

Combining these two asset classes reduces volatility, thus making the combination more acceptable than either one by itself. If you were primarily an investor in the S&P 500 Index, you would have really appreciated REITs in several years, including 1977 and 1979 – to say nothing of 2000-2004. If you were primarily a REITs investor, you would have appreciated having the S&P 500 Index in several years, including 1998 and 1999 and, farther back, 1985, 1989 and 1990.

When we put a portfolio together, we use multiple asset classes, not just the S&P 500 Index. REITs have similar benefits when compared with several other asset classes. You’ll see that in Tables 2 through 5.

Compared against U.S. large-value stocks (Table 2) REITs had 12 calendar years in this period where return differences were more than 15 percentage points. In two of those years (2000 and 2002), the difference was more than 30 percentage points.

Table 2

Click here to open Table 2 in a new window.

In Table 3 you’ll see that REITs were at least 15 percentage points different from U.S. small value stocks in 15 calendar years. In three of those years the difference was more than 30 percentage points.

Table 3

Click here to open Table 3 in a new window.

Table 4 shows a similar pattern for REITs compared with U.S. small-cap stocks. You’ll find six years with return differences of more than 20 percentage points and eight more in which the difference was more than 15 percentage points.

Table 4

Click here to open Table 4 in a new window .

Finally Table 5 shows that, against international large-cap stocks, REITs had 19 years of differences over 15 percentage points. In some of those years, the differences were really huge.

Table 5

Click here to open Table 5 in a new window.

Now I’ll call your attention to one more thing that’s particularly interesting to me about these tables. If you go through year by year, you’ll find only four of these 36 years (1974, 1988, 1994 and 2005) in which REITs didn’t have a 15-percentage-point difference in return from at least one of these other asset classes.

This shows me that, if you are looking for diversification in an equity portfolio, REITs are a good way to get it. I am sure that the returns and the patterns of these asset classes will be different in the future. But I believe REITs will continue to perform differently from other equity types while they produce favorable long-term returns.

In my book, that’s a winning combination.