Q: At the start of each year, we like to begin with a discussion of your outlook across all global asset classes. As you look at the broad opportunity set for 2017, what are your return expectations, and where are you seeing opportunities for attractive returns?

Arnott: Before we look to the future, let’s reflect briefly on 2016, a year that seems to have marked the end of a nine-year drought for value stocks and a three-year drought for most “Third Pillar” asset classes (inflation-linked assets, credit and emerging markets). In 2016, as in 2012, the PIMCO All Asset funds delivered attractive returns, with very light equity holdings. But as the saying goes, “Don’t mistake a bull market for genius.” In 2016, as in 2012, the Third Pillar fared well,¹ with an average real return of 5.9% for unhedged U.S. dollar investors, representing a healthy bounce back from the depressed levels at the start of the year (see Figure 1, which shows the trailing one- and three-year returns for passive exposure to a host of global asset classes sorted by 2016 performance).

The recent performance may cause some investors to ask whether the upswing in many Third Pillar markets has run its course. We think not. Consider emerging market (EM) equities and local debt, for instance: Even after returns of 9.0% and 6.3% through November 2016, respectively, three-year past returns for these asset classes remain negative, leaving room for further appreciation. (We recently published a white paper, “The Emerging Markets Hat Trick,” that shows how compelling EM investments have become as a consequence of their grinding bear markets.) Today, the combination of attractive valuations, depressed currencies and positive price and economic momentum increases our confidence that these and other Third Pillar markets still have substantial runway as we enter 2017.

Digging a bit deeper, we see a current dividend yield of 2.6% for EM equities, representing a spread of 50 basis points (bps) over the S&P 500. Similarly, the J.P. Morgan GBI-EM local bond index has a nominal yield of 7.7%, 490 bps above U.S. corporates (as measured by the Barclays U.S. Intermediate Corporate Index) and 110 bps over U.S. high yield bonds (as measured by the Barclays U.S. Corporate High Yield Index). EM yields exceeding U.S. high yield? That’s interesting, especially given that half the EM market is investment grade. Yields are often strong indicators of future performance. Still, looking at yields alone can be misleading, as they ignore other important factors that can influence performance, such as credit losses, currency risk, capital gains or losses from shifting yields (often driven by mean reversion), and shifting fundamentals, to name a few.

Research Affiliates’ capital market expectations, which incorporate all of these return drivers using robust modeling, may help provide a more complete picture. (Our real returns web application at www.researchaffiliates.com/en_us/asset-allocation.html offers these forecasting models free of charge.)

In the case of EM, the path suggested by yields – i.e., that EM stocks and bonds are projected to outperform other assets – holds true in our more thorough estimation (see Figure 2). In fact, relative valuations that are symmetrically opposed between EM and the U.S. complement the information derived from starting yields. For example, the current Shiller P/E valuation ratio in the U.S. is 27 (versus a long-term average close to 16), which implies a near 70% overvaluation of the S&P 500, all else being equal; the cap-weighted EM index, on the other hand, is trading at a Shiller P/E of under 12, implying a near 30% undervaluation.

To gain a fuller understanding of what is driving expected returns for each asset class, it’s helpful to look at the relative impact of valuation compared with other key factors (see table). We rank-ordered the asset classes based on the impact of valuation from most positive (EM equity, with a 3.3% boost from valuation) to most negative (U.S. small cap, which takes a 3.7% hit). The more than 7% spread in the impact of valuations across asset classes indicates a substantial tailwind for strategies favoring cheap Third Pillar assets.

Q: While valuation metrics like Shiller P/E are tested ways to identify whether assets are over- or undervalued, they provide little guidance as to when those valuations might revert. With expected annual valuation changes of more than 3% in each direction depending on the asset class, could you be putting too much emphasis on these metrics?

Arnott: What we pay for any given asset is, of course, of the utmost importance, given that the price hugely influences future returns. The “yield plus growth” parts of future returns are easier to model (though admittedly not precisely) because we know the yields of assets we choose to buy and we know their historical (or real) income growth. Future changes in valuation multiples (or yields) – and the length of time over which these changes take place – are far more difficult to predict. Full mean reversion is entirely possible, as are valuation overshoots (i.e., going from overvalued to undervalued and vice versa). A new normal is also possible, in which current valuations face no pressure to revert to a long-term historical norm. If we assume full mean reversion, the valuation “tail” wags the expected return “dog,” utterly dominating return forecasts.

For these reasons, we haircut our expectations for mean reversion. For example, we would need to see a 5% drop in the U.S. Shiller P/E ratio every year for 10 years to revert to historical norms; but maybe today’s valuations are here to stay and represent a new normal. To account for both eventualities, we split the difference: We project a more modest 2.4% drop in Shiller P/E each year for the next decade, which would leave us with a still-expensive Shiller P/E of 21x in the year 2026. Given our haircut, it’s hard to make a compelling case that we’re too pessimistic!

But what happens if current valuations really are here to stay? We can easily run the following thought experiment. By simply setting future valuation changes to zero in our model, we can see the potential impact of a steady state for valuations (see Figure 3). While this assumption is probably unrealistic, it may be useful to see what could be in store if we ignore valuations and focus entirely on yield and expected growth.

Even under this extreme assumption, we see that EM stocks and bonds have higher expected returns than most other assets. The same goes for high yield and REITs – so all in all, a good showing for Third Pillar assets. The notable change in this scenario is the outlook for U.S. stocks, both large and small cap. If current valuations do represent a new normal, these assets begin to look more appealing.²

At the other end of the spectrum – full mean reversion – the changes play out in the opposite direction. In this scenario, we would expect a 12% return advantage for EM over U.S. equities over the next decade, with EM delivering 11% per year and the U.S. losing 1%.

There’s an old saying that a man with a watch knows the time, but a man with two watches can never be sure. By showing two models (for a halfway mean reversion and no mean reversion), have we simply muddied the waters? We don’t think so. Both views show us that assets like EM stocks and bonds, high yield and the like would do reasonably well whether or not valuations revert toward historical norms. Our All Asset strategies would likely only suffer in a world where expensive assets become more expensive and cheap assets become cheaper. That happens too, as it did most recently in 2015. We think the best response under such circumstances is to accept the inevitable psychological and monetary pain and buy more of whatever has become the cheapest! Eventually, years like 2016 will likely come around to reward us for this investment discipline.

While EM stocks and bonds appear positioned to win whether valuations revert to historical norms or not, U.S. stocks would require valuations to remain very high in order to earn even 3% above inflation. Ultimately, we believe it’s more prudent to invest in assets that are expected to enjoy valuation tailwinds but that don’t require them to deliver a nice return over a 10-year horizon.

Q: Thus far we’ve looked at a mean reversion model and a model that assumes no change in valuations, but what if price momentum continues? How would that affect expected returns?

Arnott: Momentum tends to exert its influence over months, not years, before running out of steam and changing direction. At the start of 2016, Third Pillar assets had poor momentum and were very cheap. After a three-year bear market, the tide finally turned on 21 January. Now, Third Pillar assets have great momentum and are a bit less cheap. So if momentum asserts itself in the usual way, we should expect good performance for the least expensive Third Pillar asset classes for some time to come. This is not the momentum scenario that has some investors worried. They’re wondering what happens if high-valuation assets resume their march toward ever-higher valuations and the cheap Third Pillar markets revisit their January 2015 lows (or worse).

Suppose this scenario were to prevail for an extended period, against all odds. Over time, ever-cheaper assets would look increasingly attractive from the perspective of future return expectations while ever-more expensive assets would have lower and lower expected returns. This is the obvious result based on the assumptions in our models. But we should consider a more important, if subtler, point. After the fact, a dollar of realized return is exactly the same whether it comes from yield or capital gains (taxes aside). Not so when we’re modeling future returns. Remember that our confidence in the ability of yield to predict future returns is high; we are less confident in the predictive powers of growth rates and default risk, and our confidence that valuations are predictive, particularly over shorter time horizons, is lower still. Since yield has generally proven dependable in expected return models, it needs no haircut, making returns from yield more reliable for investors considering their future plans and spending needs. On the other hand, since valuation changes may or may not happen, we haircut them.

When we put it all together, yes, declines in prices on already cheap assets can be painful as they happen; however, the future benefit can be substantial, not just because of the future valuation tailwind, but also because yields grow higher! So as you choose your preferred asset allocation mix to meet your future needs, remember this: The components of returns can be as important as the returns themselves. And negative momentum – the type that likely has some investors worried about cheap assets – sets the stage for healthy and reliable yields in the future. This should be comforting to those who fear that the march against a contrarian, value-oriented investment process may resume, despite the more recent and favorable reversal.

The All Asset strategies represent a joint effort between PIMCO and Research Affiliates. PIMCO provides the broad range of underlying strategies – spanning global stocks, global bonds, commodities, real estate and liquid alternative strategies – each actively managed to maximize potential alpha. Research Affiliates, an investment advisory firm founded in 2002 by Rob Arnott and a global leader in asset allocation, serves as the sub-advisor responsible for the asset allocation decisions. Research Affiliates uses their deep research focus to develop a series of value-oriented, contrarian models that determine the appropriate mix of underlying PIMCO strategies in seeking All Asset’s return and risk goals.