Why Not 100% Equities

(Or "I Can't Believe We Are Doing This One Again")

Recently, a new paper has been making a big splash in our small pond of academic/quant investing (it was recently the #1 most downloaded finance paper on SSRN). By "new," I just mean "recently written," as much of it ain't new. The paper comes to the rather startling game-changing conclusion that long-term investors should be 100% in equities, not in a more diversified portfolio. We have been down this road before with this idea refuted with alacrity and panache at least as far back as the mid-1990s.

Basically, both the new and old version of this argument for 100% equities come down to the rather trivial observation that the asset with (as supported by both theory and long-term empirics) a higher expected return (stocks) has, on average, a higher realized return. And thus, a long-term investor should own the higher expected return asset. Staggering stuff.

I will be concise (relative to my norm), as this is well-trod ground:

• Equities winning long term vs. bonds isn't a surprise result, it's exactly what is supposed to happen and is entirely consistent with very long-established theory (which holds up pretty darn well, BTW). It just ain't interesting to show the higher expected return asset has generally a higher realized return with the probability of winning (by some margin) increasing with your time horizon. It's finance 101. It's actually just math 101. Yet every few years someone writes a paper and gets a lot of attention by showing the higher expected return asset has, wait for it, a higher average realized return.
• I don't know anyone who thinks a portfolio of stocks and bonds (60%/40% to keep things simple, though that ubiquitous choice is pretty arbitrary) has a higher unconditional expected return than 100% stocks. We (academics, practitioners, anyone who's taken a cursory look at modern finance) prefer a diversified portfolio because we believe it has a higher return for the risk taken, not a higher expected return.
• In finance 101 we are taught that in general we should separate the choice of 1) what is the best return-for-risk portfolio?, and 2) what risk we should take? This new paper, and many like it, confuse the two. If the best return-for-risk portfolio doesn't have enough expected return for you, then you lever it (within reason). If it has too much risk for you, you de-lever it with cash. Remarkably this has been shown to work.
• The above is enough to make 100% equity a silly argument. But it gets worse. This sample period is likely biased quite high. Rising valuations have made equities themselves (and especially the USA vs. ex-USA differential) a very likely overestimation of the future. The new paper that kicked off this blog actually looks at both a 100% domestic portfolio and a 50%/50% domestic international version, which is better than simply defaulting to 100% domestic – but many, perhaps most, boil this argument down to "just own the S&P 500."
• In particular, this only-a-little-new new paper makes one statement that is just an indefensible whopper. They state, "Given the sheer magnitude of US retirement savings, we estimate that Americans could realize trillions of dollars in welfare gains by adopting the all-equity strategy." This is very poor economic reasoning. It's a violation of the same principle supporting my long-time rant that "there are no sidelines"! Equities are already 100% owned. If some investors read this "new" paper and decide to buy more equities, they have to buy those equities from other investors. This can force the price up, and the expected future return down, but everyone can't suddenly have double the normal amount of equity dollar return out of thin air. Claiming there are trillions being left on the table is really just non-economic hype.
• I've couched everything so far in terms of bonds and stocks, but most of it also applies to liquid alternatives (alts). Liquid alts, even if diversifying and attractive on their own, can be hard to use to improve a portfolio's top-line expected return (not its risk-adjusted return, which is much easier to do) if they take very low volatility. Low volatility is a problem unless one a) can lever the new better portfolio containing liquid alts or b) can change other weights in the portfolio (e.g., funding from bonds, which are also low volatility, can make low volatility alts useful, if still not as useful as if they were high volatility).
• Of course, why alts cannot be tolerated at more standalone aggressive levels has vexed me since at least 1998. Many effectively say, "investors shouldn't invest in alternative assets that are marked-to-market even if they believe in them because they can't take the drawdowns—even if these drawdowns (and recoveries) are not related to, or even act as a hedge for, their portfolio-wide large stock market exposure."
• We've found most can't acknowledge that this rationale is the same as saying "this would improve an investor's portfolio, but very occasionally it could make them look unconventionally bad relative to others, so they might give up on it. Therefore investors should just stand pat with the less-effective portfolio." Citing that many can't do it is not a valid justification for not trying to change things for the better. It is not our collective job to tell investors "keep doing what will likely earn you less for the same risk, as it's just oh so comfortable for you." Rather, it's our job to collectively convince, cajole, and clamor for what we believe is right, and then help investors stick to it. Too often the observation that doing what's right is hard becomes a self-fulfilling prophesy and an excuse to continue doing what's wrong. Well, one day, we will figure out how to "volatility launder" high-volatility alternatives by not marking them to market – then the problem will be solved.

Back to the paper. There is a piece to be written that would be more useful, which:

1. Starts with a realistic guess at the equity risk premium that doesn't only examine a period of rising valuations often focusing on the country that ex-post won.
2. Explains why the most basic finding of modern finance—that, within reason, you should lever the best portfolio and not put it all in the highest returning asset—is impossible to implement (as opposed to "it's totally possible and really not unsafe, but we're scared we will panic").
3. Explains why some subset of investors should take much more risk than they are currently (despite somehow not being able to tolerate a smidge of leverage, again an odd combination).

I would likely disagree with such a paper, but it would also likely be more reasonable than "just put it all, but not more than all (that's leverage!), in the highest expected return asset," and contain some useful analysis. In turn, simply looking at historical results and urging investors to "buy the thing that's gone up the most over the long term" is not financial analysis, it is finger painting.

The bottom line is diversification works, theory works (eventually), owning one asset is suboptimal, extrapolating the winning country over a period of valuation increases is dangerous, finance 101 is actually helpful – and we'll likely have to do this again after the next bull market.

Article by Cliff Asness, AQR