When we as financial advisors talk with investors about the virtues of diversification, asset allocation, and long-term investing, a question that often comes up is, “What about real estate? Should I buy an investment property?” In fact, the issue of real estate tends to divide investors into camps almost like a personality dimension such as introversion/extroversion. Many investors could not be less interested in the hassles of being a landlord and prefer to keep all their money in easily accessible and transparent brokerage accounts. On the other hand, many others don’t trust the complexity and abstractness of the stock market and prefer to keep their wealth stored in tangible, physical land and buildings. Financial advisors such as myself, who generally earn their bread and butter in the public capital markets, often cast a skeptical eye at real estate investing with its high hidden costs and lack of liquidity and will tend to downplay its merits when talking to clients. Ask a real estate agent, however, and I’m sure you’ll hear a very different story.

But comparing these apples and oranges of the investing world has always been an incredibly daunting task and so there has never been a consensus answer to the question of whether stocks or real estate is the better investment. To be sure, there’s mountains of data on the risks and returns to stocks, bonds, and other publicly traded assets – one of the major selling points of such investments – but details on privately held real estate investments are much more sparse. Enter a bold new paper from three economists at the Federal Reserve Bank of San Francisco: The Total Risk Premium Puzzle. Building on a long line of financial research that documents the very-long-run returns to traded assets like stocks, bonds, and cash, the authors Jordà, Schularick, and Taylor (henceforth JST) add in the dark side of the moon of financial markets: privately held residential real estate. By painstakingly compiling a dataset of historical housing prices and rental rates, JST piece together a time series of the total returns to housing investments for 16 developed market countries going back as far as the 1870s, comparing risk and return properties to those of traded investments. Their conclusions are striking; according to JST “housing returns in the long run are comparable to those of equities, and yet housing returns have lower volatility… often by a factor of 2 or more.” Around the world and for over a century, being a landlord has paid about as much as being a stock investor with about half as much risk.

The paper has quickly made a splash in economics circles and has been picked up in the mainstream financial press. Tyler Cown at Bloomberg, normally known as a cautious and skeptical writer, pithily summed up the takeaway many people had: “Scared of Stocks? Buy a House Instead. New research shows that real estate is both a better and safer investment than previously believed.”

 

Real Estate Too Good to be True?

I am extremely skeptical of this conclusion (of course I am, I trade public markets!) and want to rain on the parade of real estate investors everywhere.

The first clues lie in the very name of the paper and the reason why so many observers in financial circles were surprised by it. For decades financially savvy observers have pointed out that buying stocks has been a great way to build wealth. “Stocks for the long run” goes the slogan for intelligent investors everywhere and the name of the bestselling popular finance classic. In fact, the long-run returns to stocks have been so good that economists have long found it a bit of a puzzle, the “equity risk premium puzzle”, as it’s often called. According to classical economic models, the returns to stocks are much higher than they “should” be. Yes, stocks are much riskier than cash in the bank, but the compensation that investors earn for bearing that risk seems unreasonably high, so much so that it seems irrational that more people don’t have more money invested in the stock market. And so it is all the more surprising then that real estate should earn similarly high returns as stocks with even less risk. JST name their paper “The Total Risk Premium Puzzle” to highlight this fact, and spend much of their paper exploring how startling this finding is from the perspective of standard economic models. They state, “The striking conclusion is that if the equity premium on its own is a puzzle that is hard to explain, introducing into the representative investor’s portfolio a second asset class, of equal weight, and with higher returns, lower volatility, and low correlation to the first, is likely to make the puzzle even bigger.”

Now, the most popular solution to the equity premium puzzle is to say that if equity returns are higher than makes sense in a rational market, then it must be because investors are irrational. In the last few decades the entire field of behavioral economics has risen in attempt to explain how phenomena like loss aversion and cognitive biases might scientifically explain irrational investor behavior, but this is something that every financial advisor has firsthand experience with and has been remarked upon for generations. Stock market losses loom large in investors’ minds, and there is always something bad happening in the news that seems to justify getting out or staying out of the stock market until it’s “safe again.” To many, probably most investors, the stock market is scary, and financial advisors often feel they must drag their clients, practically kicking and screaming, towards a higher level of stock exposure. On average, investors seem to be biased against stocks, and that seems to be the main explanation for why they are such perennially good investments.

Houses, on the other hand, have essentially the opposite emotional valence. Home-ownership is a basic part of the American dream (or whatever its analog in other nations), and many people simply consider it an essential component of being a responsible adult. Thus, the majority of houses in the US and most other advanced economies are owner-occupied, meaning far more people “participate” in the housing market than in the stock market. Most homeowners either downplay or don’t even consider the risks associated with housing, and instead tend to focus on resentment over the idea of “throwing money away” on rent. In other words, people seem to be biased in favor of housing, in stark contrast to the stock market. Where financial advisors struggle to encourage their clients to put more money in stocks, they usually struggle in the opposite direction to discourage their clients from spending too much on housing. This makes the finding that housing earns such higher risk-adjusted returns than stocks all the more confounding.

Given how surprising this research is in light of our prior understanding of markets, we should take it with a healthy dose of skepticism. As anybody who’s been following the replication crisis in social science over the last several years knows, most published research findings are false, failing to hold up to independent confirmation, including in finance. And the more surprising a finding is, the less likely it is to hold up over time. This paper is novel research that relies crucially on the tabulation of the history of rents going back a century or more, taken from the unpublished PhD thesis of one of the authors, hardly an infallible source of truth. Not having access to JST’s source material myself, we’ll have to wait for independent replication to verify or disprove the credibility of their data. In the meantime, there are two more general criticisms I’ll make of JST’s findings. 1) It’s probably too good to be true, and 2) to the extent it is true, it’s not really all that relevant to most investors.

 

Why it’s Probably False

On the first point, these findings fly in the face of what we know already about real estate that is publicly traded. Real Estate Investment Trusts (REITs), are a special type of corporation that derive most of their earnings from real estate income. Their shares often trade on stock exchanges and are usually already included in stock indexes and portfolios. As an investment structure they’ve been around for a while. Here’s how US REITs, as represented by the FTSE Nareit All REIT index have done compared to the US stock market in general (MSCI USA).

 

Stocks vs REIT
Stock vs REIT

 

Since 1972, REITs have had about the same returns as the overall stock market, with a little bit more risk. REITs notably lost considerably more than the broad market during the 2008 financial crisis, as well as during the early ’70s oil shock recession. REITs also tend to have moderately high correlation to stocks, moving in the same direction together more often than not. This means REITs are generally not very good diversifiers because adding them to a portfolio doesn’t reduce risk any. So while JST claim that private housing wealth has low volatility and low correlation, readily verifiable information on publicly traded real estate indicates it has high volatility and high correlation.

Is there something about REITs that makes them riskier than other real estate investments? Well, REITs do tend to use lots of debt, and this leverage can make the shares more volatile than their underlying assets, though it should proportionately increase return as well. There’s also the obvious fact that REITs, as publicly traded shares, change hands every day, whereas houses are usually owned by the same person for several years at a time.

The NCREIF Property Index (NPI) addresses both these concerns. It’s the go-to index for privately-held commercial real estate: a nationally representative composite of commercial properties held by institutional investors, reported quarterly on an unleveraged basis. Here’s how it’s done since inception in 1978, with REITs and broad US stocks for comparison.

 

NPI vs Stocks
NPI vs Stocks

 

Okay, so the NPI’s returns are a little bit lower than REITs, suggesting leverage is an important component of the latter’s return. But the volatility has gone WAY down, as has the correlation to the market, consistent with JST’s findings in housing. Could real estate’s mere lack of liquidity be the secret to it’s superior risk-adjusted returns?

No, this is probably just an illusion. A long literature in finance points to the fact that returns on illiquid assets such as real estate, private equity, and hedge funds are artificially smooth. Because these kinds of investments are often held for years at a time without any trading, the managers reporting their valuation at any given time often have considerable discretion in how they price the assets, and every incentive to make it look safer than it may actually be. The NPI, for instance, is based on the value of appraisals, a process with an inherintly high degree of subjectivity. Cannon and Cole (2011) show, for example, that appraisals are off by 12% on average from transaction prices, and even just looking at transaction prices is biased by the fact that investors will usually only sell a property when they believe the price is favorable to them, giving us an incomplete view of the real estate market where we see far more of the winners than the losers. Appraisals are also a much slower process than trading on centralized exchanges, so using them introduces lags in the data, resulting in returns that aren’t aligned with those of the public markets in time. This causes private real estate to look less correlated to stocks and therefore a better diversifier than it actually is. This is clearly visible just looking at the graph above during the time of the financial crisis. Both the NPI and the public markets experience a significant downturn followed by a brisk recovery in the 2007-2010 period, but the NPI seems to perform it on a delay. This would have made it look like an investor with a portfolio of publicly-traded stocks and privately-held real estate would have been cushioned from the depths of the crisis and had a steadier recovery, for an overall smoother, less volatile experience. But if that investor were to actually try to sell her investment property and realize that lower volatility in, say, January 2009, she would almost certainly have discovered that her portfolio wasn’t so protected from losses after all.

A number of researchers, including Pagliari et al (2005) and Ang et al (2013), have tried to correct for these biases and disentangle the risks and returns of private commercial real estate using more sophisticated statistical techniques, and have found that they are basically about the same as their public counterparts, a result consistent with findings in other illiquid asset classes like hedge funds and private equity. We should expect the same to hold true for the broader class of private residential real estate.

In fact, we should probably expect private residential real estate to do worse. While JST find that housing prices modestly outpace inflation (most of their return comes from rent, whether actual or imputed), another recent research paper released by Credit Suisse found that after accounting for capital expenditures and quality adjustments, houses have experienced an average annual real 2.1% net capital loss over the long run. The fact that researchers can’t agree whether housing prices have gone up or down over the last century speaks to the difficulty of getting a fair and complete view of the data on such a heterogeneous asset class, and we may never reach a consensus view on the matter, but that more pessimistic estimate at least is consistent with what virtually every homeowner and landlord has told me, “Owning a house is way more expensive than I originally expected.”

 

Why it’s Probably Irrelevant

There is another major way that stocks and houses differ as an asset class. For stocks and other publicly traded assets, it’s straightforward to “buy the market.” Investors these days can easily purchase an index fund or ETF and own a representative slice of virtually all the world’s financial securities and pay no more than a tiny fraction of a percent to do so. In public markets, investors can achieve nearly perfect diversification. In private real estate markets, however, true diversification is not only imperfect, it’s impossible. Most houses are only owned by one or two people: their occupants; and nobody securitizes their home and sells it on an exchange. Even professional landlords and commercial real estate investors typically only hold a small portfolio of properties with limited geographic diversification and that represent only a drop in the vast ocean of global real estate wealth.

This is sort of an awkward fact for financial economists because many of our theories are based on these ideas of “representative agents,” investors who diversify perfectly across all global assets. And for such an imperfect market, real estate is… kind of a big deal. Estimates for these things vary, but here’s one helpful breakdown of the size of the real estate market in comparison to other assets.

 

Global Asset Distribtuion

Real estate in total is worth more than all the world’s stocks, bonds, and gold combined. Most of it is residential housing, and most of that is privately held and non-investable. This may sound surprising at first, but makes sense when you consider the net worth of the typical person. For most homeowners, their house is by far their most valuable asset, dwarfing everything else in their portfolio. In fact, it’s only in about the top 1% of the wealth distribution that you begin to more commonly see individuals whose liquid investment assets are worth more than their home.

From a financial theory perspective, the important upshot of these facts is this: most investors who do not own a home are massively under-exposed to the real estate market, and most investors who do own a home are massively under-diversified. The point being, even if the real estate market has, as JST claim, earned similar returns as stocks with lower volatility, the statement is sort of meaningless because nobody actually owns the real estate market. Whereas statements about the aggregate returns and risks of stocks correspond to the real investment experience of a wide and diverse set of actual living stockholders the world over, there is no such comparison to make with real estate. Most homeowners own only one house, whose fortunes are tied to the local economy and neighborhood development and whose value will be much more volatile than any national or global composite real estate index (JST suggest individual housing prices are twice as volatile than at the index level, and we’ve already established that they’re probably underestimating volatility). People also generally own houses in the area they work, making them doubly exposed to the local job market, further concentrating the risk of a homeowner’s portfolio. It’s analogous to the situation often found in the stock market where sometimes investors will hold large positions in the shares of their own employer. As great as investing in the stock market is, if the only way you could access it was by buying the stock of the company you worked for, it would no longer seem nearly so appealing. This can sometimes work out great, of course, as in the case of anybody who bought real estate in the Bay Area twenty or thirty years ago and earned millions in windfall profits along with excellent career growth. But it can also go terribly wrong, as with rust belt homeowners whose home equity has declined at the same time as jobs have moved elsewhere.

It doesn’t have to be this way. While you can’t really buy fractional shares of a portfolio of houses in Portland and Chicago and Topeka and Albuquerque, you can buy fractional shares in their mortgages. Securitized mortgages didn’t use to be a thing, but since the 1980s they have become a major asset class, providing much greater liquidity to a formerly fragmented market and lowering borrowing costs for home-buyers. If you own any bond mutual funds or ETFs, you almost certainly own some yourself. There’s no reason, at least in principle, that we couldn’t do a similar thing with home equity itself. Doing so would require not just financial innovation, but probably also legal reform and a cultural shift in norms about home-ownership itself, but would allow individuals to diversify the risk of their most valuable asset.

In the meantime, we are beginning to see some solutions come to market that are offering individual investors access to diversified real estate portfolios that were previously only available to institutional investors. Fintech firms like Fundrise are offering crowd-funded real estate investment accounts, and asset managers like Bluerock are starting to use a legal structure known as a continuously offered closed-end interval fund to offer access to private real estate in a similar wrapper as the familiar mutual fund. These solutions allow investors to tap a very large and previously mostly-inaccessible investment category, with less hassle and more liquidity than directly buying properties outright, and so may be worth considering as part of a diversified investment portfolio. Their main drawbacks for now are A) high fees, at least in comparison to traditional stock and bond investments; and B) high expectations. On their front page Fundrise implies you can earn “8.7% – 12.4%” returns while diversifying your portfolio. Bluerock boldly declares private real estate offers “Equity-like Returns with Bond-like Volatility.” The charitable interpretation of these claims is that they are naively optimistic. As we’ve seen, real estate is much riskier than its deceptively smooth returns might lead you to believe, with higher fees, less liquidity, and less transparency than public markets. It still may be worth including in a portfolio, but investors should curb their enthusiasm before buying. As always, if it sounds too good to be true, it probably is.

This has all been about housing as an investment; of course, most people buy a house in order to have a place to live. In my next post I’ll look at housing from this personal finance perspective and explore when it makes more sense to rent vs buy a house.