Diversification is often described as the only free lunch in finance. But after a tough year for equities and a brutal year for bonds, many investors are still recovering from food poisoning.

It’s natural to seek out alternative diversifiers and real estate is an obvious place to look. The total value of global real estate is greater than equities and bonds combined. Many real estate investment structures exist and listed real estate investment trusts (REITs) are liquid, convenient and easily accessible assets for most investors. But are they a good investment?

Not surprisingly, the National Association of REITs (Nareit) is a fan, writing “REITs historically have delivered competitive total returns, based on high, steady dividend income and long-term capital appreciation. Their comparatively low correlation with other assets also makes them an excellent portfolio diversifier that can help reduce overall portfolio risk and increase returns.”

On diversification, they continue “listed REITs help to diversify a portfolio because, as real estate, they are a distinct asset class that has demonstrated low-to-moderate correlation with other sectors of the stock market, as well as bonds and other assets. In other words, REIT returns have tended to zig while returns of other assets have zagged, smoothing a diversified portfolio’s overall volatility.”

That sounds pretty good! But my recently published research piece, Are REITs a Good Investment?, draws a very different conclusion. Here are some key takeaways from our research.

Over the period beginning October 2004 and ending September 2022, a large, listed REIT ETF, which holds over 160 REITs and closely tracks the MSCI US Investable Market Real Estate 25/50 Index realized a compound annual growth rate (CAGR) of 7.2% while a U.S. total market equity ETF realized a CAGR of 9.0% and a long-term bond ETF realized a CAGR of 4.0%. REITs were 0.75 correlated to equities and 0.03 correlated to bonds over this period.

Strong returns—check. Modest correlation—check, sort of. With these two facts, many investors would draw the same conclusion as Nareit, that REITs have provided diversifying, competitive returns.

However, there are more facts to consider. The REIT ETF realized nearly 50% more volatility than equities over this period (22.3% vs. 15.6%). Its worst peak-to-trough drawdown was about 70% vs. about 52% for equities over the same period. Higher volatility and worse drawdowns than equities, with modest underperformance over the period. These risk characteristics raise some red flags, but in and of themselves aren’t conclusive.

But, wait, there’s more! A multi-variate regression over the same periods indicates that the REIT ETF had 114% exposure to equities and 34% exposure to long-term bonds. This means that investing $100k in a broad-based REIT fund was somewhat like investing $114k in equities and $34k in bonds, with a $48k margin loan.

That helps to explain why the REIT ETF was so much more volatile than equities—it has embedded leverage. It also calls into question the claim that “REIT returns have tended to zig while returns of other assets have zagged.”

And herein lies the problem. An investor who has $114k of exposure to equities and $34k exposure to bonds deserves to earn the returns associated with those exposures. The REIT ETF didn’t keep up—it didn’t even keep up with $100k of equity exposure.

Turning back to the multi-variate regression, the REIT ETF realized a -2.8% annualized alpha over this period. That number represents how much—in realized returns—REIT investors were effectively short-changed relative to the risk exposures the REIT provided to equity and bond markets.

Let’s consider an alternative to the REIT to further illustrate the shortcoming. We established that investing in the REIT is somewhat like investing 114% in equities, 34% in bonds, with a 48% margin loan. Let’s create that portfolio as a “synthetic REIT” and horserace it against REITs.

Recall from earlier, the REIT ETF realized a 7.2% CAGR over the nearly two-decade period. The “synthetic REIT” realized a CAGR of 11.3% over the same period, a 4.1% difference in annualized return. What about risk? The REIT ETF realized 22.3% volatility, while the synthetic REIT realized 17.4%. Drawdowns tell a similar story: worst drawdown for REITs was 70.1% vs. 54.4% for the “synthetic REIT.”

The goal here is not to advocate for a “synthetic REIT.” Instead, it is to provide an appropriate benchmark for evaluating a REIT. Listed REITs have historically provided material equity and bond exposure. Contrary to appearance, listed REITs have not lived up to their diversification promise. And when considering these exposures, they’ve underperformed. Less return and greater risk. It’s hard to get excited about this.

Considering these findings, I conclude that historical allocations to listed REITs would have been detrimental to a well-constructed equity and bond portfolio.

But there’s a more important conclusion I would ask you to consider and apply elsewhere. Be careful with analysis that only scratches the surface—it is too easy to be led astray. A deeper analysis is typically required to establish, with confidence, how to construct optimized portfolios.

Roni Israelov is the chief investment officer of NDVR.