In investing, and in life, you do not always get compensated for taking risk. Investors sometimes forget that not every risk in a portfolio comes with a corresponding return. This is true even if “markets are efficient”.
The purpose of this post is to point out that enumerating the risks in your portfolio is just step 1. Step 2, arguably the more important step, is to determine which of those risks you are getting paid to take and which you are taking without any compensation. This is a key step because it is not true that all risks come hand in hand with excess returns.
We’ll start with a stylized example and then move into more complex asset allocation issues. With NFL football season in full swing, consider the following scenario:
Louis does two things with his savings. One, he invests in equities, and two, every week he bets on NFL football games -- specifically he bets the “over” in all games (meaning he bets for both teams to score more total points than expected). Those are his two “investments”.
Let’s analyze Louis’s portfolio. What are the main forces (often called “factors”) that are driving the day-to-day movements of his portfolio? To do this we might run what is called a “factor analysis”, and if we did, we would very likely see two risk factors:
- Equity markets, and
- Total points scored in the NFL.
Those are the two main “risk factors” in Louis’s portfolio. In other words, if you could ask any two questions about the world to get a handle on how Louis’s portfolio is doing, you’d ask “how are equity markets doing?” and “how many points are NFL teams scoring?”.
Step one is complete: we have successfully mapped the risks in Louis’s portfolio. That’s the easy part. But now we say “which of those risks is Louis getting compensated to take?” In other words: which of those risks carries a positive return premium? What we’d likely find is that equities carry a positive return premium (described in more detail below), but NFL points do not. The latter is uncompensated risk in Louis’s portfolio (unless he has predictive alpha in that space, meaning he can systematically out-predict the market, but let’s assume he doesn’t). This is true even if overs are perfectly priced by the NFL betting market.
The reader may find it very obvious that Louis is taking uncompensated risk in betting on NFL overs, but this is a good example for actual dilemmas investors face when it comes to asset allocation decisions. Replace NFL overs with commodities, and things become more tricky.
Many investors have a commodity allocation in their portfolio -- a fully intentional allocation that seems to indicate that most investors believe that taking on commodity risk brings a return premium. But is commodity risk more like equity risk? Or more like betting NFL overs risk?
A good starting point is to assume that a risk factor does not have a return premium, unless there is a strong reason why it should. Let’s go through the examples from NFL overs to commodities to illustrate what we mean:
- Equity Risk: Equities have strong theoretical support for carrying a positive return premium, in that they wouldn’t exist unless they had one. When businesses sell equity, they want cash in exchange. But from the investors’ point of view, equities are a pure investment product. There’s no real reason for anybody buy equities, to give businesses that cash, unless the businesses price that equity to return more than the risk-free rate. Thus, equities almost by virtue of their existence as pure investment products are very likely to bear a positive expected return. Incidentally, the same goes for bonds. Nobody is taking on credit (or term) risk without getting compensated for it. If markets are efficient to any degree at all, both equities and bonds have positive return premia.
- NFL Overs Risk: Theoretically there should be no return premium associated with betting NFL overs. There is nobody that needs to “sell” this risk to investors in the way that companies need to sell equity. There is also no correlation between total NFL points scored and the broader economy, so NFL betters are not likely to shy away from NFL overs for fear that they will crash at the same time that they lose their jobs, which is a very salient concern with equities. Perhaps an empirical analysis would show that betting NFL overs pays well over time, and maybe that analysis would be convincing on a pure empirical level.1 But absent a convincing empirical analysis, assigning “zero premium” to the NFL overs risk factor is probably a good idea.
- Short Equity Risk: Before moving to commodities let’s take a short interlude for short equity risk. Going short the equity market is risky! But, if we believe going long equities has a positive return premium, then going short equities must have a negative return premium. So this is an intuitive and quick example of a risk that is not rewarded by the market over the long term. Investors nevertheless sometimes take short risk for the possibility of short term gain, but it is, for good reason, not commonly part of long term asset allocation.
- Commodities Risk: Commodities are different than equities in that they are not pure investment products, and they do not exist solely to generate returns for their investors. In fact, finance professionals might be interested to hear that this is not even their primary purpose. Commodities are useful goods. There is no real reason to expect that holding a commodity would entitle you to any sort of premium, any more than I should expect a return premium for buying and holding a box of plastic spoons. The spoons are there to use, not to invest in. I shouldn’t expect my box of spoons to go up in price on any theoretical grounds. I certainly shouldn’t invest in millions of boxes of plastic spoons and make it part of my asset allocation. Isn’t oil, a useful product, more like a box of plastic spoons than a pure investment product?
So the case for commodities is tough to justify. Yet many, if not most, investors seem to think commodities have an expected return. Hopefully this comes from careful analysis of the sort that Two Sigma does in the “Forecasting Factor Returns Whitepaper”, and not from some assumption that where there is risk, there must be a return, or that because something is labelled an “asset”, it must have a positive expected return. Plastic spoons are an asset too.
In case the reader is interested, a quick summary of Two Sigma’s take on commodities is that to land in a place where we’d actually expect positive returns from commodities, we need to be investing via futures (not spot commodities): holding futures instead of the commodities directly can potentially provide 'insurance' to commodity producers, by giving them a market to sell their future production at a fixed price. However, this effect does not appear universal, and seems to apply to only pro-cyclical commodities such as energies and industrial metals. More details are in the aforementioned white paper, but suffice it to say, these conclusions are not obvious or trivial, and even the empirical evidence for commodity returns is weak if one accounts for their correlation to equity risk. So it surprises us to see investors so confidently taking general commodities risk as if the case for a positive return premia were as strong as that for equities.
While we’re at it, let’s explore a few more common risks in portfolios:
- Developed Market Currency Risk: Currencies have real uses like commodities, so they are not purely investable assets. That should make us wary that they carry a return premium. Currencies also run aground on symmetry arguments: a US based investor is taking price risk by holding Euros, but a Euro based investor is taking risk by holding dollars. They can’t both be getting a positive expected return 2. So, at least with developed market currencies (emerging markets are another story), the case for a return premium is very weak, and we think it’s best to assume it is zero.
- Real Estate Risk: Real estate is an interesting asset to consider, and is somewhat inconclusive. Should investors expect a return premium from holding real estate? On the one hand, it’s not a purely investable asset -- it has real uses, and to a first approximation everybody needs a home, so why would a home carry a risk premium? Fundamentally, it’s hard to imagine that early humans, living in self-made huts, were sitting on an asset with a compensated return premium because they were bearing price risk. Everybody just needed a hut and the “price risk” was something they were forced to take on if they wanted a roof over their head. There’s no reason to expect they were getting compensated for it.
On the other hand, one can think of the value of a property as the present value of a stream of rent payments. Since those rent payments are not riskless, they should offer a return premium over the risk-free rate to the owner. So in that sense real estate is like a bond. This probably leads to the conclusion that real estate has a risk premium for investors who are renting out homes, but not for homeowners (or hut owners!). This would mean that the premium would be earned through rent payments, and not through actual price appreciation of the property. Indeed, for all the hype, real estate real price appreciation in the United States since 1890 has been a surprisingly low 0.4% per year according to Case-Shiller data (remember, homes get old so they depreciate too!). Higher returns from real estate investments mostly come from including rental yields in the numbers.4
The real estate discussion may have been a little much, but the point is that the assessment of whether any given risk brings return is completely non-trivial, and requires a lot of thought, theory, and empirical work to get right. It’s a far cry from the assumption that any “risky asset” automatically has a return premium.
So where does that leave the investor? What we recommend is taking a two step approach:
- Understand what you own by identifying the risks in your portfolio using statistical methods (Venn is a great platform to help you accomplish this task)
- Go through your set of risks and determine (perhaps after reading Two Sigma’s white paper and/or other research) which risks are compensating you, and which ones aren’t.
Step 1 is pretty clean. As far as step 2, we’ll just say expect a lot of debates.
1 As it turns out, there is actually an interesting strand of finance literature related to the topic of sports betting. We thought we’d share a few examples in case you’re interested in reading:
Avery, Chevalier (1999). Identifying Investor Sentiment from Price Paths: The Case of Football Betting. Journal of Business.
Gandar, Zuber, O'Brien, Russo (1988). Testing Rationality in the Point Spread Betting Market. Journal of Finance.
Moskowitz (2017). Asset Pricing and Sports Betting. Working Paper.
2 Although Siegel’s paradox suggests that all investors may have positive excess return expectations from holding static foreign currency exposure due to Jensen’s inequality, Campbell et al. (2010) show that the expected premium from this mathematical curiosity is negligible and we will not all get rich by trading currency exposures with one another.
4 Source: Jordá, Knoll, Kuvshinov, Schularick, Taylor (2017). The Rate of Return on Everything, 1870-2015. Working Paper. See Table 7 for a breakdown of nominal housing returns in the U.S. across price appreciation and rent payments. Not including rent payments in the calculation of housing’s total return makes the asset class look less attractive on a real basis.
This article is not an endorsement by Two Sigma Investor Solutions, LP or any of its affiliates (collectively, “Two Sigma”) of the topics discussed. The views expressed above reflect those of the authors and are not necessarily the views of Two Sigma. This article (i) is only for informational and educational purposes, (ii) is not intended to provide, and should not be relied upon, for investment, accounting, legal or tax advice, and (iii) is not a recommendation as to any portfolio, allocation, strategy or investment. This article is not an offer to sell or the solicitation of an offer to buy any securities or other instruments. This article is current as of the date of issuance (or any earlier date as referenced herein) and is subject to change without notice. The analytics or other services available on Venn change frequently and the content of this article should be expected to become outdated and less accurate over time. Two Sigma has no obligation to update the article nor does Two Sigma make any express or implied warranties or representations as to its completeness or accuracy. This material uses some trademarks owned by entities other than Two Sigma purely for identification and comment as fair nominative use. That use does not imply any association with or endorsement of the other company by Two Sigma, or vice versa. Click here for other important disclaimers and disclosures.