In The Long Run, The House Wins

|
Includes: ACC, AIV, AMH, APTS, AVB, BRG, CLPR, CPT, ELS, EQR, ESS, INVH, IRET, IRT, MAA, NXRT, RESI, RVEN, SUI, UDR, UMH
by: Hubert Wu
Summary

A recent study compared long-run returns for residential housing with other assets.

Over the last 100 years, housing and equity earned similar real returns, but housing was less volatile.

Housing was also less correlated across countries than equity during this time.

These findings suggest residential housing is an important asset class to consider in portfolio construction.

Investing in residential REITs is one way to do this.

Housing accounts for around half the national wealth in most economies and is usually the largest asset on individuals’ balance sheets. Yet the long-run nature of the asset is somewhat mysterious relative to others: unlike equities and bonds, data about housing’s rate of return can be hard to find for more than a few countries and decades.

A landmark paper and database released in late 2017 by prominent macroeconomists solves this problem. It also made a striking discovery: over the last century, residential housing was far more attractive an asset than was previously recognised and is often on par (and better in some respects) than equity.

I’ve detailed the work’s key findings and potential implications for investors below.

Residential real estate: ‘safe as houses’, but with equity-like returns

Jordà et al. (2017): the best public historical database of long run annual asset returns available

In “The Rate of Return on Everything, 1870–2015”, Òscar Jordà, Katharina Knoll, Dmitry Kuvshinov, Moritz Schularick, and Alan M. Taylor build a database of annual returns (measured by yield, capital gains and total returns) for four assets: equities, residential housing, government bonds, and short-term bills. (Here’s a recent short overview by Alan Taylor that explains the project in more detail.)

Advantages of the dataset over others include its:

  • time coverage: returns span almost 150 years
  • country coverage: returns cover sixteen advanced economies
  • inclusion of residential real estate returns, which was non-existent for many countries and years before the authors’ work.

In other words, if stable characteristics of asset returns exist across long swathes of time and space, Jordà et al’s (2017) database is the best current place to find evidence of them. For housing, the authors construct returns by combining new data about house prices and rents. (More details about how this is done are here and here.)

The paper finds three stylised facts about housing returns.

Finding 1: in the last century, housing returns were comparable to equity returns

The long-run data show that total returns for housing were similar to equity in throughout last century (although equity has outperformed housing since World War II and in recent decades on a simple average basis). Average annual real returns for both asset classes are below.

Long run average real returns per annum (percentage points), equity and housing

Asset/time period

Equity

Housing

Full sample

6.89

7.05

1951-2015

8.28

7.44

1981-2015

10.68

5.39

Source: Table 5 (p. 22) in Jordà et al. (2017).

The averages mask meaningful cross-country differences. For the full sample, equity outperformed housing by the largest amount on average in Italy, the US, and UK; the reverse was true for France, Norway, and Portugal. Here’s a chart for those countries from data in the paper:

Cross-country full sample mean returns for equity and housing (bar labels are % percent per annum), countries with the largest positive and negative differences

Long-run returns for equity and housing for several countries Source: data from Table 5 (p. 22) in Jordà et al. (2017).

Finding 2: housing returns are less volatile than equity

Jordà et al's (2017) second finding about residential housing is that its returns were less volatile than equity over the studied period. A central figure from the paper showing decadal moving averages summarises this well:

Decadal moving averages of cross-country returns for equity and housing

Trends in real returns on equity and housing Source: Figure 6 (p. 20) of Jordà et al. (2017).

Housing's lower volatility can also be seen in the geometric return of the two assets (a measure that better captures the effects of compounding relative to simple, i.e. arithmetic, averages) and the standard deviations of the assets' returns:

Standard deviation and geometric mean real returns, equity and housing

Asset/outcome

Equity

Housing

Standard deviation of returns, full sample

21.94

9.98

Geometric mean of real returns, full sample

4.64

6.61

Geometric mean of real returns, post-1950

5.54

7.1

Source: data from Table 3 (p. 13) in Jordà et al. (2017).

This mix of comparable returns (Finding 1) with lower volatility (Finding 2) leads to one of the research's most puzzling, but fascinating, discoveries: over the very long run (circa the last century), housing earned a superior risk-adjusted return than equities. Sharpe ratios for the asset summarise this fact; note too that the pattern is consistent across countries.

Risk and return measures for equity and housing

Risk and return of equity and housing Source: Figure 8 (p. 23) of Jordà et al. (2017).

Finding 3: housing returns are less correlated across countries than equity returns are

A third finding is that residential housing returns are less correlated across countries than equity returns are. And this relationship – which is perhaps due to globalisation and its effect on linkages between capital markets – has only become stronger since 1980.

Rolling decadal correlation coefficients (average of all country pairs) for equity and housing

Correlations between equity and housing returns across countries

Source: Figure 7 (p. 21) of Jordà et al. (2017).

A related finding is that when compared with each other, the covariance between equity and housing is relatively low in comparison. The current rolling correlation coefficient is slightly above 0.20 for the two assets.

Implications for investors

Actionable lessons

For me, the main investing lessons from this work include:

  1. Residential housing is a valuable asset class relative to others (most notably, equity and bonds). While this may strike some as obvious, hard evidence on this was scarce until now, and the fact remains debated. Housing’s high Sharpe ratio may even suggest it is a superior long-term investment to equities – but there are some dangers to this view. Some of these are outlined below.
  2. It may be worthwhile investing in both equities and housing as the low covariance (and similar returns) between the two suggests significant gains to be had from diversifying between the two asset classes.
  3. If feasible, significant gains from diversification across countries can be had from investing in residential housing, more so than applies to equity today.

However, a question remains as to how one would actually go about investing in the market for residential housing given that that individual houses are expensive, less liquid than stocks, and come with many other costs (transaction, search, etc.).

One means of overcoming many of these barriers – and please feel free to add others in the comments – is to invest in residential real estate investment trusts (REITs). Residential REITs provide diversified exposures to the relevant market and are often listed on exchanges. (Note that these are different from commercial REITs which cover a different kind of asset that is subject to different economic drivers.)

Some example REITs are at the bottom of this article.

Caveats: why this idea may be wrong

Here are some counterpoints to the argument in this article:

  • The past is not necessarily the future: long-run historical research like Jordà et al. (2017) is worthless for investing decisions if the next hundred years are not like the last. And, without a theory – which the paper and this piece don’t offer – it’s hard to justify why history should repeat. See the chart about international stock correlation above for an example of where reliance on the past in this context can go wrong. And sharp-eyed readers may also see that the gap between stock and housing returns has widened in the 1980s until now.
  • It may take many decades for returns to materialise due to the long-run nature of this analysis. This may result in years of underperformance that aren’t appropriate or acceptable for many.
  • Equity returns in the paper cover only a small set of developed economies. This may exclude the high potential returns from emerging markets (driven, for example, by higher economic growth rates) and make the comparison between the assets an unfair one.
  • If you are in countries like the US, the residential housing market may be in the midst of a price bubble. This could mean that now is exactly the worst time to invest in residential property.
  • It’s not easy to execute the strategy. For example, residential REITs aren’t available for all countries, high levels of leverage may be required; regulatory and tax considerations may further complicate investing in housing; among others.

Annex: list of residential REITs for selected countries in Jordà et al. (2017)

Country

Example exchange-listed residential REITs

USA

ESS; AVB; SUI; CPT; ELS; MAA; EQR; IRET; AIV; UDR; ACC; NXRT; AMH; INVH; RESI; UMH; APTS; IRT; BRG; RVEN; CLPR

UK

LON:RESI; LON:PRSR; LON:KCR; LON:CSH

Australia

ASX:URF; ASX:INA

Japan

TSE:3269; TSE:8986; TSE:3278; TSE:3282; TSE:3278; TSE:3459; TSE:8979

France

EPA:GFC (a mix); EPA:ALTA (a mix)

Source: returns data are from Table 5 (p. 22) in Jordà et al. (2017) and apply to the full sample; residential REITs from REITNotes.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.