Q: With a significant proportion of the developed world’s sovereign bonds offering zero or negative yields, how have you positioned the All Asset strategies to seek their return benchmark?
Brightman: Central banks across the developed world have forced interest rates far below the inflation rate, and in many cases into negative yields, by purchasing vast quantities of securities from the public. The resulting negative real interest rates on cash and government bonds force us to abandon our industry’s assumption of a positive real risk-free rate of interest, which forms a foundation both for the long-term return potential of our investment portfolios and for finance theory itself. If the world’s central bankers turn to Milton Friedman’s idea of helicopter money (as repeated by Ben Bernanke in 2002) – that is, direct purchases of new government debt for redistribution to the citizenry or other favored recipients – as is increasingly being discussed, then inflation may soar. The time to protect against inflation is when it’s not expected.
Before delving into recent positioning, let’s revisit the overall goals of the All Asset strategies since launch back in 2002 and 2003. We seek to 1) increase returns relative to a traditional mix of mainstream stocks and bonds, 2) diversify away from stock market risk and 3) counter mainstream assets’ vulnerability to inflation. Our approach to achieving these goals has always been to widen the opportunity set by investing in asset classes beyond traditional stocks and bonds, seek alpha within each asset class and actively manage the asset mix.
To achieve these objectives, we invest the All Asset strategies primarily in what we call the Third Pillar:1 asset classes that we believe offer higher yield potential and stronger growth opportunities than mainstream stocks and bonds, low correlation to the U.S. stock market and positive correlation with inflation – in contrast to developed market stocks that form the First Pillar and core bonds that form the Second Pillar of many investors’ portfolios. We can and do invest in mainstream asset classes, but only when they’re cheap and thereby potentially offering higher yields and stronger expected returns. We shun mainstream stocks and bonds when, as today, they trade at high prices and provide correspondingly anemic yields. The bear market in Third Pillar diversifying markets from 2013 until January of this year – driven by tumbling inflation expectations – leaves these markets at deep bargains relative to mainstream stocks and bonds.
We favor higher-yielding asset classes because long-term asset class returns have tended to closely track starting yields. Since 1800, the starting yields of U.S. 10-year Treasury bonds and their subsequent 10-year returns have exhibited a correlation of 96% (see Figure 1). As of 31 July 2016, the real yield on 10-year U.S. Treasuries is 0.0% and for short-term U.S. Treasuries, between -0.7 and -1.0% (for 3-year and 1-year Treasuries, respectively). Similarly, in real terms over the next 10 years, our capital market assumptions suggest core bond indexes such as the Barclays U.S. Aggregate and the Barclays Global Aggregate are priced to give us 0.32% and 0.29%, respectively, as of 31 July 2016.2 In short, these Second Pillar asset classes are currently priced to offer roughly zero returns, or lose ground, net of inflation. And that’s before we pay our taxes.
Let’s now turn to how we’ve positioned our portfolios for this low-yielding environment. Over the last few years, we have tilted away from core bonds and ramped up our exposure to a roster of higher-yielding, faster-growing asset classes, many of which have been pushed into bargain territory. This rotation is a natural outcome of our systematic, value-oriented asset allocation process.
As a result, the All Asset strategy’s allocation to core bonds – namely, investment grade developed market bonds – tumbled from 32% at the end of 2009 to 11% as of 31 July 2016. Over the same period, our exposure to Third Pillar asset classes (which tend to be inflation-sensitive) rose from 64% to 81%. The story that has unfolded recently in these markets has us pleased we made the trade: An equally weighted average of Third Pillar indices outperformed core bonds – as represented by the Barclays U.S. Aggregate Index – by more than 8% year-to-date as of 15 August.
We have also turned to liquid alternatives, in lieu of cash and mainstream bonds, to serve as a source of dry powder to deploy the next time markets are hit hard and offer bargains. As real rates have declined into negative territory, our average exposure to alternative strategies has risen, from an average of 5% in 2012 to 14% as of 31 July 2016. With an absolute return – not relative return – focus, alternative strategies are projected to provide the majority of their return from active positioning instead of relying upon market rates of interest. These liquid alternative strategies are an important part of our toolkit, serving to de-risk the portfolio while still providing the opportunity for positive rates of return and a buffer against inflation.
With yields of government bonds and U.S. stocks severely repressed, we project dismal long-term returns for portfolios comprising these mainstream asset classes. By expanding the opportunity set, seeking alpha and contra-trading with a value-oriented discipline, we have strong conviction that the All Asset strategies can achieve their objectives of higher returns, lower equity beta and greater inflation hedging relative to mainstream stocks and bonds.
Q: Given the All Asset strategies employ a model-driven approach to drive asset allocation decisions, how does manager discretion enter into the investment process?
Arnott: We begin with a model-driven approach for executing a value-oriented discipline and contrarian philosophy in the All Asset strategies. Nearly 40 years of experience have reinforced a critical point about successful long-term investing: Profiting from mean reversion requires the ability to go against human nature and to patiently endure the discomfort of a contrarian positioning. The cliché, “If it ain’t broken, don’t fix it,” comes to mind. Well, if a portfolio has delivered handsome gains, it may be very timely to fix it, and if it’s been disappointing, it is usually helpful to either stay the course or – difficult for most investors – increase our bets. “If it ain’t broken, it’s time to fix it!” To make these trades that are uncomfortable in the moment but that in our view may be profitable ultimately, our investment process in the All Asset strategies is designed to be systematic.
While our holdings are mostly driven by our models, there are three areas where we apply judgment to our investment process.
Firstly, when current events or circumstances are not captured in our models, then a subjective adjustment to the model output may be warranted. Discussions about these adjustments typically occur during monthly collaboration sessions between senior investment professionals at Research Affiliates and PIMCO. Representing a small component of our investment process, these subjective adjustments tend to be expressed as modest offsets to the final allocations produced by the model. For example, if we think something is happening that will benefit Treasury Inflation-Protected Securities (TIPS) that our models can’t capture, we might choose to hold 5% more in TIPS than whatever the model indicates.
We consider subjective adjustments only after they clear three hurdles. First, we ask ourselves: Is the event or issue at hand missing from the model? For example, perhaps our models are not accounting for a recent regulatory or accounting change or a new demand driver for an asset class. Next, we ask: Would the adjustment be performance-chasing or contrarian? Would we want more of something that’s already recently soared? Our own experience has taught us that contrarian adjustments tend to be more effective than trend-following adjustments. This second filter is much harder to clear but it’s critical. Finally, is this issue well-known and likely priced into markets, or are the markets also missing it? Given these hurdles, subjective adjustments are not common and rarely affect more than two or three asset classes; at this moment, we have no subjective adjustments.
Secondly, we subjectively decide how quickly to trade toward the target weights indicated by our models; this process involves collaboration between our colleagues at Research Affiliates and PIMCO’s trading desk. The valuations-based, contrarian nature of our trading means we tend to be a provider, rather than a taker, of liquidity. We do not harbor any illusions that our models will contain useful information about the behavior of markets over spans of days or weeks. Also, we buy assets that others want to sell, and vice versa. As such, our own fund inflows and outflows are actively used as a part of our rebalancing process to help us trade toward our targets. When our allocations move more than a percentage point or two away from our target, we begin to take more direct action, such as selling one fund and buying another.
Finally, an important application of judgment is reflected in our model design, which is guided by our investment beliefs. I’ve often said that what is comfortable is rarely profitable. While unpopular moves don’t always pay off immediately, they ultimately tend to reward, adding value over a fiduciary horizon. We believe that if we buy assets with attractive prices and high yields – and stay true to our conviction to hold these positions patiently until mean reversion kicks in – then we reward our long-term investors.3 We consider ourselves life-long students of the capital markets; as we learn more about the markets, we continually refine our models, while hewing to this central belief.