In previous posts, we saw that real estate has had volatility similar to that of stocks, with correlation low enough to make it a useful portfolio diversifier. But we also want to know what kind of returns to expect from real estate if it is included in a portfolio.

Historical Returns on Real Estate

Just as the historical data on the volatility of real estate is open to a range of interpretations, so is the historical data on the returns to real estate. Perhaps more importantly for investors, the prospective future returns are also subject to a range of reasonable projections.

In this post, we will look at some of the studies of historical returns to real estate, and see what we can learn that might help us form reasonable expectations looking forward.

Recency Bias?

In the United States, real estate, by most measures, has provided excellent returns over the last 30 or 40 years, with a number of studies finding that real estate has outperformed the stock market. This has led to the widespread assumption that such excellent returns are normal. But are they?

There are strong arguments that such returns are not likely to be repeated in the next 40 years.

One important such argument is based on interest rates. One way to think about income real estate (i.e. properties that are built, such as apartments, office buildings, or retail buildings) is as a long-lasting stream of free cash flows.

For example, suppose that an apartment building produces $100,000 net free cash flow under normal conditions, when it is 95% occupied with good tenants paying market rent. The return to owners of the building, over the long run, will be the sum of the net cash flows plus the change in the value of the asset.

The cash flows are essentially constrained on the upside by the change in market rents. (While this might not pertain to individual buildings, which might be improved for example by capital investment to upgrade a building from a “C” to a “B”, it must be true for the market overall.) Thus, the cash flow portion of the return is unlikely to grow by more than inflation. (Over relatively long periods in the US, rents have actually trailed inflation, as measured by the CPI. There is considerable variation among local markets.)

Capital values, on the other hand, given the cash flows, are essentially a function of the cap rate investors are willing to apply to that type of real estate. Various factors affect this cap rate, but the dominant such factor is the level of interest rates. Real estate is a long duration asset, making the market value of real estate quite sensitive to changes in the level of interest rates.

And it is likely the dramatic fall in interest rates between 1980 and 2020 in most of the world that accounted for the phenomenally good performance of real estate over that period.

Dissenting Voices

David Chambers, of the Cambridge (UK) Business School, studied the actual, same-property returns. Chambers and colleagues report: “Real estate—housing in particular—is a less profitable investment in the long run than previously thought.”

They collected financial data for the investment portfolios of “four large Oxbridge colleges over the period 1901–1983. Gross income yields initially fluctuate around 5%, but then trend [downward for agricultural and residential upward for commercial real estate]”.

They conclude that: “Long-term real income growth rates are close to zero for all property types. Our findings imply annualized real total returns, net of costs, ranging from approximately 2.3% for residential to 4.5% for agricultural real estate.” [1]

While not terrible, the low end of their estimate of 2.3% real (i.e. inflation-adjusted) returns to real estate is not too different from the long run estimate of the real return to the US 10-year treasury note. While tax regimes have varied considerably over the period since 1900, it is likely that for most of the time the tax regime for real estate was not worse than for bonds.

More Optimistic Historical Estimate of Returns

Oscar Jorda of UC Davis and the Federal Reserve Bank of San Francisco, and colleagues published The Rate of Return on Everything, 1870 to 2015[2] in 2017.

Jorda et. al. reported a much higher real return to residential real estate than did Chambers. Jorda reported real returns as high as 4.7% (compound annual returns), composed of 4% income and .7% capital gains.

However, the Jorda data, unlike the Chambers data, relies on a variety of assumptions and data sets. For example, Jorda uses data from official statistics, such as components of price indices, and also uses composites based on a variety of sources, including weighting various country components by estimates of the total value of the capital stock of the country. While such methods can have validity, the results do not represent anything that anyone could have actually invested in.

Other researchers have looked at individual markets, and attempted to determine the returns over the long run to owning specific, individual properties. For example, X studied New York commercial buildings during the 20^th century. Using data from about 85 buildings that traded at least twice during the century, he compiled return estimate. He found that the return was basically the net income, and that capital gains just about kept up with inflation.

Conclusions

Should investors include income real estate in their portfolios? Investors who own stocks probably should, because of the expected diversification benefits. However, the evidence suggests that over the long run, unlevered real estate is not likely to produce equity-like returns. Instead, the returns are likely to be lower, and are likely to be dominated by the income component.

This latter point suggests, for example, that if an investor buys an interest in apartments at a cap rate of 3%, the long run expected return is likely to approximate that income return, with the expectation that the capital value will keep up with inflation.

Contact Us

To speak with us further, reply back to this email or call us at 703-437-9720.

---

[1] The Review of Financial Studies 34 (2021) 3572–3607

[2] http://www.frbsf.org/economic-research/publications/working-papers/2017/25/