Nobel laureate Harry Markowitz famously said that diversification is the only true free lunch in investing. Yet it’s not always free – in fact, it may be getting more expensive.
With yields near multi-decade lows on cash, U.S. Treasuries, developed market sovereign bonds and other traditional diversifiers, a number of investors have concluded that compensation for holding these assets simply isn’t adequate. Some also worry that these assets have limited ability to mitigate portfolio downside: Yields are so low, there’s little room to fall further (see Figure 1).
In recent years some have pared their allocations, while others have meaningfully reduced portfolio duration – only to see performance suffer as yields fell. Now, many are left wondering when to get back in.
FIGURE 1: THE HIGH OPPORTUNITY COST OF TRADITIONAL DIVERSIFIERS
Source: PIMCO and Bloomberg as of 30 September 2015
Hypothetical example for illustrative purposes only. Cash is represented by the three-month USD Libor Index; U.S. core bonds is represented by the Barclays U.S. Aggregate Index; global sovereign bonds is represented by the JP Morgan GBI Global ex-US USD Hedged Index; and long-term Treasuries is represented by the Barclays Long-Term Treasury Index. Current yields represent the yield to maturity for each index as of 30 September 2015. PIMCO's 10-year annualized returns forecast is an estimate of what investments may earn on average over a 10-year period. Return assumptions are not a prediction or a projection of return. Actual returns may be higher or lower than those shown and may vary substantially over shorter time periods. Figures are provided for illustrative purposes and are not indicative of the past or future performance of any PIMCO product.
The recent spike in market volatility underscores the urgency of this issue. While there is no silver bullet, we suggest a two-pronged approach to enhance the diversification and risk-mitigation potential of policy portfolios:
- Get the most “bang for the buck” from traditional diversifiers.
- Consider non-traditional diversifiers with higher expected returns and/or the ability to provide significantly more downside mitigation.
As with most things in investing, long-horizon investors have a clear edge in making these decisions.
Getting the most out of traditional diversifiers
Consider a move from cash to enhanced cash
Cash is a multi-faceted asset which serves three key roles in multi-asset portfolios: liquidity provision, offensive dry powder, and arguably with reduced efficacy, an anchor to windward. All these benefits, however, come with a trade-off: low yield.
To address this, investors can look beyond money market instruments to short-term bond strategies that aim to provide higher yields by investing in assets with slightly longer duration or slightly higher credit risk (see Figure 2). The ongoing reforms in the money markets are creating several opportunities to add further value through active management (see Jerome M. Schneider’s recent article, “The B-Side of Capital Preservation”).
FIGURE 2: YIELDS OF MONEY MARKET INSTRUMENTS VS. OTHER CASH EQUIVALENTS
Source: Bloomberg, ICAP and BofA Merrill Lynch as of 30 September 2015 *397 days refers to the SEC rule stipulating that a security must have less than 397 days until maturity to be eligible for classification as a “money market security.” Agencies: Agency discount note (composite of discount-offered levels received from brokers and dealers for U.S. Agency discount notes); Treasuries (U.S. on-the-run government bill/ note/ and bond indexes); commercial paper, certificate of deposit and corporate are composite curves provided by Bloomberg; general repo (consolidated data provided by ICAP); Agency repo: consolidated data provided by ICAP; AA, A and BBB corporate curves represented by their respective generic Bloomberg USD US Corporate Curves.
Maintain portfolio duration (or be cautious when reducing)
A key question facing investors in a low-yield environment is whether lowering portfolio duration is a wise long-term investment strategy. We believe the answer is no for investors with long-term liabilities and those with meaningful exposure to risky and illiquid assets. Over time, bond investors are compensated for being long duration – as long as yield curves are upward-sloping, which has nearly always been the case (barring the very late stages of economic expansions). As such, a structurally lower duration position is not advisable for the multi-asset investor with a moderate to long horizon, for the most part regardless of the path of interest rates. Even for investors who lack long-term liabilities or are underweight risk assets, the decision to reduce duration should be made with great care and caution. A reduction in duration is likely to lead to diminished diversification at the policy-portfolio level, something investors could consider offsetting by broadening the menu as we discuss below.
FIGURE 3: MIRROR IMAGE: IF 30-YEAR TREASURY YIELDS REPLICATED THEIR TREND IN REVERSE
Source: PIMCO and Bloomberg. Data from September 2000 – September 2015
FIGURE 4: SIMULATED RETURNS OF 10-YEAR AND 30-YEAR U.S. TREASURIES FROM THE MIRRORING EXERCISE (BASED ON TREND IN FIGURE 3)
Source: Bloomberg and PIMCO. Data from September 2000 – September 2015
Hypothetical example for illustrative purposes only. PIMCO simulation based on Bloomberg generic U.S. Treasury yield data from September 2000 – September 2015. Simulation assumes a 10-year or 30-year Treasury is held for one month, then sold and the proceeds reinvested in another 10-year or 30-year Treasury security (respectively). Figures are provided for illustrative purposes and are not indicative of the past or future performance of any PIMCO product.
As discussed in “The Case for Duration in the Long Run,” even if interest rates were to mirror the path they have taken over the past 15 years – i.e., they shoot higher over the next 15 years – strategies with longer duration can still post positive returns. This is due to the magic of compounding yields: As rates grind higher over time, elevated carry may overcome principal losses, providing the potential for net positive returns (see Figures 3 and 4).
So how much duration is appropriate given the prospect for rising rates? The answer, of course, depends on an investor’s objectives, risk tolerance, horizon and existing exposure to risk assets, among other factors. This is a topic of frequent discussion between PIMCO’s solutions group and our clients.
Also, when the Fed eventually decides to hike interest rates, front-end yields are most likely going to rise, but the impact further out the yield curve is a bit more ambiguous if history is any guide. Nevertheless, the liftoff will present active managers with attractive opportunities to add value through duration and yield curve strategies.
Broaden the menu of diversifying strategies
Another way to address the diversification challenge is to broaden the menu. Academic research and modern financial tools have spurred a new class of strategies designed to capture alternative sources of returns with low correlations to traditional assets. Some are beginning to go mainstream.
Time-series momentum (managed futures)
Among the most promising is time-series momentum. This captures the phenomenon whereby an asset’s return history tends to indicate its future performance: Recent winners are expected to continue to win and recent losers should continue to lose – they exhibit momentum.
Time-series momentum is a simple trend-following strategy that goes long in markets that have experienced positive excess return over a defined look-back horizon, and vice versa. It is among the most well-documented and extensively researched market anomalies, shown to work across most major asset classes and countries. Studies show it has the potential to generally deliver long-run returns with negative correlation to risk assets in down markets and positive correlation in up markets. In other words, time-series momentum strategies have the potential to help investors reduce downside capture and add upside capture (see Figure 5).
FIGURE 5: MANAGED FUTURES HAVE OUTPERFORMED IN STRESSED MARKETS
Source: PIMCO and Bloomberg
Hypothetical example for illustrative purposes only.
The Time-Series Momentum Model trades 20 markets: five each in equity index, bond, currency and commodity futures. The strategy trades once per week, taking a long position if the current futures price is above the one-year moving average price, and taking a short position if it is below. Each position is scaled inversely to the recent three-month daily realized volatility of the contract, and the overall strategy is scaled to target 10% volatility using trailing 10-year windows to estimate volatility of the strategy. Over short periods, risk can be skewed to some asset classes. Fixed transaction costs ranging between 1 basis point and 10 basis points were estimated from available market data for each futures market and subtracted from returns. No fees have been applied to model data; performance would be lower if applied. Figures are provided for illustrative purposes and are not indicative of the past or future performance of any PIMCO product.
Even though time-series momentum has gained mainstream acceptance, no one really knows why it works. Popular explanations include initial underreaction and delayed overreaction by investors due to behavioral biases such as anchoring, confirmation, disposition effect and herding – factors unlikely to disappear anytime soon.
In the past, investors could access time-series momentum only through commodity trading advisors (CTA) hedge funds. These were generally opaque and typically charged high fees. Now, these strategies can be accessed at much lower cost via mutual funds that provide daily liquidity at NAV. Regardless of vehicle, manager selection is critical, as there is no single “right” implementation strategy to capture the momentum factor. There is large divergence in risks and returns across managers in this category.
Alternative risk premia
Another approach with the potential to deliver high returns with low to negative correlation to traditional asset classes harnesses alternative risk premia (also known as style premia or structural alpha). This approach involves allocating to a basket of systematic strategies that seek to harvest well-known and persistent sources of excess return potential within and across asset classes (see our In Depth article “Carry and Trend in Lots of Places”). Studies show that a large portion of hedge fund returns can be attributed to systematic exposures to various style premia (see Figure 6).
FIGURE 6: THE ALTERNATIVE RISK PREMIA MODEL HAS PERFORMED HISTORICALLY
Source: PIMCO and Bloomberg
Hypothetical example for illustrative purposes only. The 60/40 portfolio is based on the S&P 500 and the Barclays U.S. Aggregate Index. The Alternative Risk Premia Model is a rules-based, defined opportunity set diversified portfolio of six systematic strategies that seek to capture time-series momentum, value, carry, and volatility risk premia across global equity, interest rates, G-20 currency, and commodity markets. The risk contribution from each systematic strategy is sized to maintain balance across asset classes and styles. The model targets an ex-ante volatility of 10%. For this article we assumed a static weighting for the historical period. The 60/40 + 10% Alternative Risk Premia Model comprises 54% S&P 500 Index, 36% Barclays U.S. Aggregate Index and 10% Alternative Risk Premia Model. The 60/40 + 20% Alternative Risk Premia Model comprises 48% S&P 500 Index, 32% Barclays U.S. Aggregate Index and 20% Alternative Risk Premia Model. The 60/40 + 30% Alternative Risk Premia Model comprises 42% S&P 500 Index, 28% Barclays U.S. Aggregate Index and 30% Alternative Risk Premia Model. No fees have been applied to model data; performance would be lower if applied. Figures are provided for illustrative purposes and are not indicative of the past or future performance of any PIMCO product.
Alternative indexation strategies, also known as “smart beta,” are closely related to alternative risk premia; they take exposure to some of the same factors. Yet there is an important distinction. Smart beta strategies are long-only and therefore have a high degree of market exposure. Alternative risk premia strategies, in contrast, are primarily long/short and therefore have little to no market exposure.
While alternative risk premia can be (partially) accessed by investing in smart beta indexes, a market-neutral, multi-asset approach can be compelling because it can isolate the diversification prowess of these strategies by largely stripping out market beta. Also, combining various alternative risk premia across assets has its own diversification benefits because some of these premia – value and momentum, for instance – are negatively correlated with each other.
Historically, as Figure 6 shows, adding these strategies to a conventional 60/40 portfolio not only improved the risk/reward characteristics, but also provided meaningful diversification benefits. The Alternative Risk Premia Model exhibited close to zero beta to the equity market over the full sample, and negative equity beta during large drawdowns, thus providing diversification when it mattered the most.
The devil is in the details, however, when it comes to investing in alternative risk premia. Questions that frequently arise include which premia to take exposure to, why they continue to persist, whether they can be traded, how much risk to allocate to each, whether allocations should be static or dynamic, and the total market capacity for such strategies. As portfolio risk exposures may vary, customization may make sense for large institutional investors. In any event, it’s critical to select managers with sound answers to these questions and trading and risk management systems to construct these strategies and efficiently and appropriately trade them.
Tail risk hedging
Tail risk hedging is the third arrow in the quest to achieve diversification and downside mitigation with limited opportunity cost. Tail risk hedging is an option-based strategy that seeks to shield portfolios during sudden market shocks by investing in put options or option-like securities. In our view, it is best performed at the policy-portfolio level in a holistic fashion. PIMCO has extensively researched and written about our approach to tail risk hedging, including the book, “Tail Risk Hedging: Creating Robust Portfolios for Volatile Markets.”
The key difference between the strategies outlined above and tail risk hedging is the latter’s ability to address “gap risk.” Tail risk hedging is designed to reduce portfolio drawdowns during black swan events such as Black Monday, 9/11 and the 2010 Flash Crash. Furthermore, the return pattern of tail risk hedging is highly convex, i.e., potential returns increase exponentially with the size of the market sell-off.
While several institutional investors have adopted portfolio-level tail risk hedging and view it as another asset class, others have hesitated to do so as they can’t stomach having to regularly write checks to invest in hedges. An alternate approach is to aim to make the tail risk hedging allocations self-funded by investing in an income-producing strategy that seeks to generate enough yield on a consistent basis to offset the periodic costs of acquiring the hedges.
There are two trade-offs in adopting this approach: higher capital commitment and somewhat less loss mitigation during tail events (as the income-producing strategy in that scenario will likely suffer losses, offsetting some of the gains generated by the tail risk hedging component of the portfolio).
In sum, Markowitz’ free lunch still exists but the pickings on the traditional buffet have gotten slimmer. The solution to rock-bottom yields on core diversifying assets is to expand the menu. Find the right mix of enhancing traditional diversifiers and add non-traditional diversifiers with the potential to enhance returns and allay the downside.
There are still viable options for long-term, risk-conscious investors willing to fight inertia and expand their horizons.
The authors would like to thank Portfolio Managers Josh Davis, Graham Rennison and Matt Dorsten for their contributions to this article.