Three Approaches to Better Outcomes in

Better Beta: Thinking Differently Rethinking Alpha addresses the Taking the Sting out of the Tail: About Core Fixed Income goes back challenge facing many fixed-income Hedging Against Extreme Events to the building blocks of portfolio investors today—generating alpha in summarizes various approaches across construction to answer the question today’s environment of low rates—by the capital markets to buffer portfolios “Is it possible to construct a portfolio exploring the benefits created when in periods of market stress. that delivers solid performance the definition of alpha is expanded irrespective of fluctuations in the to reflect new market realities. macroeconomic cycle?” 2 Richard Brink Senior Portfolio Manager

Michael Mon Portfolio Manager Seeking a more stable footing in Dimitri Silva changing fixed-income markets Portfolio Manager

yield on the Barclays US Aggregate Index Better Beta: Thinking Differently came from US Treasuries (Display 1). In the years that followed, US Treasuries About Core Fixed Income accounted for the bulk of the market yield and generated huge capital gains as interest rates fell. Today, by contrast, a Is it possible to construct a portfolio that delivers solid perform- 100% US Treasury weighting would be far ance irrespective of fluctuations in the macroeconomic cycle? from optimal. Nongovernment securities— now a much more important driver of This paper explores how investors can get closer to that goal. market yield—accounted for almost half of market yield at the end of 2013.

Investors expect their core fixed-income credit risk premiums offered by corporate Indices Can Be Dangerous portfolio to do three things for them: bond spreads were not attractive. In In lieu of 100% government bonds, most diversify their equity exposure, generate September 1981,16.1% of the 16.5% investors today have adopted a broad returns consistent with those of the broad bond market and display low . But today’s low yields, together with concerns about the prospect of Display 1: Impact of in a Core Portfolio Has Changed Radically rising interest rates, are making investors Index Yields question whether the traditional core approach—using a broad market index, which can bring exposure to unintended Change Between Sep 30, 1981, and Dec 31, 2013 20 risks—meets their needs. Barclays US Aggregate Index Yield 16.5% 2.5%

In this paper, we discuss an alternative Contribution from US Treasuries 16.1% 1.4% 15 approach to core, one that has largely Contribution from Non-US Treasuries 0.4% 1.1% the same objectives as the traditional approach but that seeks to create a 10 Barclays US Aggregate Index Yield

balance of exposures that is less sensitive Percent to changes in the macro environment. 5 The Evolving Concept of Core Thirty years ago, developed-country Barclays US Treasury Index Yield investors seeking core fixed-income 0 exposure might have been satisfied with a 81 91 01 11 simple 100% allocation to government debt. For example, in the US bond market Historical analysis does not guarantee future results. in 1981, US Treasuries were on the brink For the period January 1, 1981, to December 31, 2013 of a bull run that played out over three Yields are measured as yield to worst for the Barclays US Aggregate and the Barclays US Treasury. Source: Barclays and AllianceBernstein decades. Much of the nongovernment market was still in its infancy, and the

1 market index as their starting point. But Then, the same thing happened in of a more thoughtful, balanced set of the problem with nearly all bond indices is reverse. Securitized issuance dried up and exposures, or “better beta,” began with that they are issuance weighted. So, US government stimulus sent the US this fundamental question. companies or governments that issue Treasury supply soaring. Non-agency more debt acquire a bigger role in the securitizations staged their biggest rally in Begin with a Multisector Approach index—and in the portfolios of index- history, but index-focused investors There are good reasons for including all benchmarked investors.* captured less of the return than they major fixed-income sectors in a core might have because their proportional portfolio: government bonds and This can have unforeseen consequences. exposure to that sector was falling and inflation-linked government securities, For example, as the US pursued the their exposure to US Treasuries was rising. along with other sectors such as mortgage-led borrowing spree that corporate and hard- emerging- caused the 2008–2009 credit crisis, the An Investor-Designed Index market (EM) debt. All these sectors non-agency securitized segment’s Rather than catering to investors’ needs, have historically generated solid weighting in the index more than a market index is in effect an issuer’s returns, allowing for a broad opportu- doubled, from about 3% to 7%, by index. But we believe that investors nity set. Diversification is another 2007. So, investors had significant should be the ones to control which obvious benefit: over the past 10 years, exposure to the ensuing sell-off as, for risks—and how much risk—they take. high-yield corporate bonds have had a example, commercial mortgage-backed negative correlation—about (0.2)— security (CMBS) spreads soared from an What would a core fixed-income index with Treasuries. average of less than 70 basis points (b.p.) look like if it were designed by investors over Treasuries, to more than 1,600 b.p. rather than driven by issuers? Our pursuit But perhaps the strongest argument for having a cross section of exposures is the fact that the major bond sectors have added value at different times, in Display 2: No Fixed-Income Sector Wins All the Time–Different Sectors Add Value different environments. For example, in in Different Environments the turbulent environment of 2008, US January 1, 2008–December 31, 2013 (Percent) Treasuries outperformed high-yield credit by 40.5% (Display 2). The following year, 2008 2009 2010 2011 2012 2013 as risk appetite returned, high-yield US High Yield High Yield US TIPS EMD USD High Yield Best Treasuries credit outperformed US Treasuries by 13.8 58.2 15.1 13.6 17.9 7.4 more than 60%. In 2012, hard-currency US TIPS EMD USD EMD USD US High Yield Inv.-Grade Treasuries Corps. EM debt outperformed US Treasuries by (2.4) 28.2 12.0 15.8 9.8 (1.5) 15.9%. These highly varied return Inv.-Grade Currency US TIPS EMD USD Currency Currency patterns demonstrate the value of Corps. 8.5 (4.9) 22.0 6.3 10.5 (1.9) combining sectors. The question is: EMD USD Inv.-Grade Inv.-Grade Inv.-Grade Inv.-Grade US What’s the most efficient way to do it? Corps. Corps. Corps. Corps. Treasuries (10.9) 18.7 9.0 8.2 9.8 (2.0) High Yield US TIPS US High Yield US TIPS EMD USD Taking Turns over Time Treasuries (26.2) 11.4 5.9 5.0 7.0 (6.6) Although the sectors in Display 2 all behaved in different ways, they were Currency US Currency Currency US US TIPS Worst Treasuries Treasuries polarized into two groupings that tended (26.7) (3.6) 1.5 1.5 2.0 (8.6) to take turns outperforming each other. Difference 40.5 61.8 13.6 12.1 15.9 16.1 One was interest-rate-related assets—

Past performance does not guarantee future results. These returns are for illustrative purposes only Treasuries and Treasury Inflation-Protected and do not reflect the performance of any fund. Diversification does not eliminate the risk of loss. Securities (TIPS)—which primarily provide US Treasuries are represented by Barclays US Treasury, US TIPS by Barclays US TIPS, high yield by Barclays US Corporate High-Yield, investment-grade corporates by Barclays US Investment Grade, emerging-market debt investors with interest-rate exposure. The (USD) by J.P. Morgan EMBI-Global and currency by Deutsche Bank G10 Currency Future Harvest plus cash return. other was growth-related assets (such as Source: Barclays, Deutsche Bank, J.P. Morgan and AllianceBernstein credit and EM spreads), which provide credit or currency exposure. These two

*For more on the limitations of the benchmarked approach, see our June 2011 research paper The Tyranny of Bond Benchmarks.

2 groupings are the building blocks of corporate bonds, a growth-related asset. no exception: US Treasuries beat high yield what’s known as a credit barbell strategy. Historically, each end of the barbell has by 18% during the 1990 US Savings & come to investors’ aid at different times. Loan crisis, then high yield bounced back, Credit Barbell Hints at Better Beta As the display shows, in 1984 Treasuries by a of 31% in 1991. The asset A credit barbell strategy concentrates on outperformed high yield by 5%, but in class’s risk characteristics in a particular high-credit-quality, interest-rate-sensitive 1985 high yield beat Treasuries by 4%. environment determined returns. assets on one hand, and on lower-credit- quality, growth-related assets on the other What’s also apparent is that returns Over the last 30 years, investors seeking a hand. These two extremes are sometimes generated by the two halves of the barbell portfolio that was more effective over viewed as risk reducing versus return diverged significantly—by an annual time would have wanted to own both US seeking. Display 3 compares the relative average of about 12%. And, the diver- Treasuries and high-yield bonds. This performance of five-year US Treasuries, an gence was greater at times of turmoil and combination outperformed a core interest-rate-related asset, and high-yield during market rallies. The credit crisis was portfolio of only US Treasuries with almost the same volatility. In other words, it had a significantly better excess return per unit of risk. Display 3: A 50/50* Barbell Strategy Delivers More Return per Unit of Risk January 1, 1984–December 31, 2013 Risk Premiums: The Building Blocks Despite the barbell’s effectiveness, it’s still a Years When Years When simple 50/50 construction. Could we US Treasuries Outperformed High-Yield Bonds Outperformed improve the results by adding other sectors 1984 or changing the mix? To answer this, we first need to delve a little deeper into why 1988 fixed-income sectors behave as they do.

1992 In the same way that each person represents a unique combination of 1996 genetic building blocks, each fixed- 2000 income asset class represents a unique combination of premiums for various 2004 kinds of risk. We believe that the key to better beta is to focus not on sectors, 2008 but on the underlying risk premiums.

2012 For example, within the two sectors in our credit barbell, we can identify four 40 20 0 20 40 60 separate risk premiums.* Size of Outperformance (Percent) US Treasuries High Yield 50/50 Treasuries, which tend to be viewed as purely interest-rate related, have two main Average Return 7.3% 10.8% 8.7% Annualized Vol. 4.9% 8.6% 5.0% risk premiums: Sharpe Ratio** 0.6 0.7 0.9 n Real-interest-rate risk premium: the Past performance does not guarantee future results. These returns are for illustrative purposes only portion of a government bond coupon and do not reflect the performance of any fund. Diversification does not eliminate the risk of loss. that compensates investors for lending *50/50 is an equal blend of US Treasuries (represented by Barclays 5-Year Bellwether US Treasury) and high-yield bonds (represented by Barclays US Corporate High-Yield). money to the government **Annualized return in excess of cash, divided by annualized volatility Source: Barclays and AllianceBernstein n Inflation compensation risk premium: the portion of a government bond

*We’ve simplified somewhat. There are other risk premiums. For example, there is an inflation uncertainty premium and there is optionality inherent in the generally callable high-yield market.

3 coupon that protects the purchasing sectors, but on the underlying risk At the overall sector level, the net effect power of the principal. premiums. The combination of these of higher growth expectations is reactions results in the distinctive return generally negative for government yields consist of the yield patterns of various sectors. bonds, but net positive for high yield. on comparable-maturity government bonds, plus a , which can be Factor Sensitivities Drive Returns Here’s the crucial point: government and broken down into further risk premiums. This dynamic is analogous to a system of high-yield bonds don’t move in opposite So, in addition to the two risk premiums levers and forces. The risk premiums are directions because they are driven by above, high-yield corporates present: the levers, upon which macro factors different macro factors—they do so exert force, resulting in changes in bond because they respond differently to the n Credit and default risk premium: the returns. Display 4 shows a macro same dominant macro factor. A credit portion of the credit spread that factor—rising gross domestic product barbell smooths volatility well because reflects the risk of a change in credit (GDP) at work—and illustrates how a the sum of the two parts greatly reduces quality or that a company will fail to credit barbell can be a stable strategy in the impact of changes in growth service its debt or go bankrupt a rising-growth environment. expectations on the returns of the whole. n Liquidity risk premium: the portion of Rising growth tends to increase the Factors and Forces Vary in Strength the credit spread that reflects the demand for, and cost of, money. This Economic growth expectations are a high-yield market’s tendency to sends real yields higher, which negatively consistently important factor in bond become illiquid from time to time. affects the real-interest-rate component returns, but there are others, such as (Illiquidity increases the risk of loss if, of government bond yields. inflation expectations and real yields (or for example, wide bid-ask spreads central bank policy). And the relative make it hard for investors to sell at a Likewise, if growth causes inflation to strength of the force these factors exert favorable price.) rise, that risk premium would contribute can vary over time. negatively. By contrast, economic growth Each premium fluctuates in value over tends to result in lower risk of credit For example, in the US, inflation became time, reacting with a different sensitivity downgrades, defaults and liquidity the dominant factor in 1974 following the to changing macroeconomic factors. So, problems, so it would be positive for the 1973 oil crisis. At the time, both Treasury the key to better beta is to focus not on growth-related risk premiums. and high-yield returns fell—which was unusual given the historically negative correlation between those sectors. But it makes intuitive sense given that the Display 4: Factor Sensitivities Help Explain Sector Return Patterns inflation-compensation ”lever” affects Rising GDP Drives Rate- and Growth-Related Risk Premiums in Opposite Directions them both in the same way. A different factor, yield expectations, became High-Yield Corporate Bonds dominant in 1994, when massive rate Liquidity hikes by the Fed triggered losses in both Treasuries and high yield. In this case, the real yield component was the main lever Credit/Default being activated. Government Bonds Adding Additional Levers Adding other sectors to the portfolio Inflation Inflation introduces other levers, or risk premiums, Real Yields Real Yields such as the EM risk premium. TIPS have a component that adjusts a bond’s Impact of Rising-Growth Expectations on Returns principal value in line with inflation. This lever is potentially very useful because, For illustrative purposes only Source: AllianceBernstein when inflation expectations are the driving factor, we would expect the TIPS

4 Using 21st-Century Financial Market Tools to Create Better Beta

More Efficient Building Blocks: Isolating Risk Premiums Fine-Tuning Volatility Bonds consist of bundles of risk premiums that are sensitive to An investor may have the optimal blend of sectors, in the certain macroeconomic drivers. Display 4 shows that duration- optimal proportions for current market conditions, but for and credit-related premiums tend to react differently in a various reasons, he or she may want to take more or less risk. rising-growth environment, making both desirable in a barbell This can be done without changing the mix, but by adjusting portfolio. But the challenge with using cash bonds is that overall portfolio volatility. corporate bonds also contain duration. In fact, of the risk premiums in high yield, the ones we really want are credit risk Volatility adjustments lend themselves to more tactical, - and default risk. term decisions. For example, if investors are worried about the near-term outlook and want to reduce risk, they might want to The table below shows how using credit default swaps in place dial back their exposure. By contrast, volatility can be increased of cash bonds can create a cleaner barbell portfolio that isolates as a way of adding octane if the market is powering into a rally. the credit and default components of high yield. This would have been hard to do 30 years ago, but today it’s relatively simple. Volatility adjustments can be useful in bringing a portfolio closer into line with investors’ objectives. For example, if they want For example, investors can get their credit exposure by using a more yield than their mix is currently generating, they can define credit default index to take a position in a basket of a range of acceptable yields and then manage risk exposure up companies. To get the duration, they could assemble a or down to keep portfolio yield within that range. portfolio of cash government bonds of the desired maturities or employ interest-rate futures. Compared with cash bonds, How is volatility adjusted? One benefit of a strategy that can derivatives can often be a cheaper, easier and more liquid use synthetic risk premiums is that it is often relatively easy and solution. And many of the derivatives in question are standard- cheap to adjust volatility. With a cash-bond portfolio, many ized or exchange-traded rather than custom over-the-counter transactions may be needed to recalibrate risk exposures, and swaps. For example, (CDX) indices can be the transaction costs are likely to be higher than for derivatives. used to gain exposure to baskets of investment-grade corpo- Derivatives can be used to de-risk by increasing the proportion rates, high-yield corporates or EM sovereigns. of cash in the portfolio. Or, to amplify exposure to the mix, a manager could reduce the cash proportion, possibly making use of some leverage. Being Able to Isolate Components Creates a Cleaner Barbell When Things Fall Apart: Tail-Risk Hedging Inflation Real Liquidity Credit/ The systematic approaches that serve investors well under normal Instrument Exposure Yields Risk Default Risk market conditions often fall short in periods of extreme market stress. The challenge is to protect against tail risk at times of crisis. Treasuries We believe that an effective tail-risk hedging strategy can help investors by buffering their portfolios when extreme events occur High-Yield and can increase the robustness of a better beta approach. Corporate Bonds There are numerous ways to manage the cost of tail risk by High-Yield Credit constructing strategies that allow investors to be long volatility Default Swaps in a cost-efficient manner. For a discussion of this technique, see

For illustrative purposes only Taking the Sting out of the Tail: Hedging Against Extreme Source: AllianceBernstein Events, AllianceBernstein, January 2013.

5 inflation component to work in the Putting Better Beta into Practice the many possible ways to combine risk opposite direction from the inflation The next step would be to use an premiums, we solved for the mix that has premium embedded in Treasuries. understanding of this system of forces the lowest sensitivity to the primary and levers to help create a macro-insensi- macro factors. While the list of levers can be extensive, tive, better beta portfolio. The goal was the list of primary forces is short. Our to assemble a portfolio of carefully What does the winning combination research suggests that risk premiums calibrated risk exposures that together look like in practice? Not surprisingly, it across the fixed-income landscape are would minimize sensitivity to all the main still resembles a barbell, with exposure to driven, to a great extent, by a small market factors. both interest-rate-related and growth- number of shared macro factors. This is related assets. However, the difference is important because it implies that What makes this possible is the ability to that it includes securities from multiple accounting for a handful of macro factors isolate individual risk premiums using sectors and countries. The interest-rate can help stabilize a portfolio across many, instruments (see explanation in (more defensive) side of the portfolio if not most, economic scenarios. sidebar, page 5). The task then boiled would typically have exposure to the down to a mathematical calibration: of government bond yield curves of

Display 5: Better Beta Compared with Barbell and Market Portfolios Total Return: December 31, 1999–December 31, 2013

Better Beta Barbell Market Portfolio Composition Cumulative Return 185% 153% 117% December 31, 2013 Annualized Return 8.4% 7.4% 6.1% Annualized Volatility 5.8% 5.1% 3.6% 300 Sharpe Ratio 1.0 0.9 1.0 Better Beta Better Beta Max Drawdown (Peak to Trough) (12.5)% (13.8)% (3.8)% 260

Barbell Barbell 220

Index* Market 180 Market

140 Primarily interest-rate driven 100 Primarily growth driven 00 01 02 03 04 05 06 07 08 09 10 11 12 13

For illustrative purposes only. Historical analysis does not guarantee future results. Better beta: Simulated performance of multisector portfolio composed of Markit CDX High Yield Credit, Markit CDX Investment Grade Credit, Markit CDX Emerging Market Credit and Deutsche Bank G10 Currency Future Harvest (all primarily growth driven), and Barclays US Government Inflation-Linked Bond and Barclays US Aggregate Treasury (both primarily interest-rate driven) Barbell: Simulated performance of portfolio composed of 50% Barclays US Corporate High-Yield and 50% Barclays 5-Year Bellwether US Treasury Market: Historical performance of Barclays US Aggregate Treasury composed of US Treasuries, US agencies/government-related bonds, mortgage-backed securities, commercial mortgage-backed securities and asset-backed securities (all primarily interest-rate driven), and investment-grade corporates (primarily growth driven) Simulated performance results have certain inherent limitations. Unlike an actual performance record, these results do not represent actual trading. Also, because these trades have not actually been executed, these results may have under- or overcompensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated or hypothetical trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to these being shown. *100 = 1999 Source: Barclays, Deutsche Bank, Markit and AllianceBernstein

6 multiple countries and have both Adding Additional Enhancements this approach would produce attractive nominal- and real-interest-rate exposure, Is it possible to improve even further on risk-adjusted returns. And we would i.e., exposure to both inflation-linked and the better beta results? One recommen- expect it to make an increasingly useful ordinary government bonds. The growth- dation is to introduce tail-risk hedging. contribution as we move into a rising-rate related side of the portfolio would Macro-insensitive portfolio construction environment. typically consist of investment-grade and can improve outcomes over long time high-yield credit in multiple countries as horizons, but over shorter time horizons Summary: Take Control of Your Risk well as exposure to EM government debt (such as the 2008–2009 credit crisis), Thirty years ago, investors didn’t need a and currency carry. (This example is from when asset performance is heavily surgical approach to bond exposure— a US perspective, but we believe that the influenced by additional financial market government bonds allowed them to meet principles are broadly similar for investors factors, we would advocate using all their goals. In today’s bond markets, in other developed countries.) tail-risk hedging to try to cushion the meeting bond investors’ objectives is downside. more of a challenge. But while the task is Display 5 uses historical data to simulate more complex, so is the fixed-income tool the performance of a better beta On an ongoing basis, we also recom- set at our disposal. Investors are able to portfolio from 2000 through 2013. The mend a dynamic approach that seeks to do things that were previously impossible. simulation compares the performance of enhance risk-adjusted returns by There are more ways to brace a portfolio a simple 50/50 barbell and a better beta modifying the balance according to against macroeconomic shocks, and more portfolio that includes multiple countries, prevailing financial market conditions. In ways of being nimble and responsive to yield curves and credit sectors. Since the other words, while we think better beta financial market shocks. The key distinc- beginning of 2000, the barbell has is a good investor’s benchmark, we tion to make in exploiting this tool set is outperformed the market by a cumula- don’t advocate a passive set-it-and-for- to think of risk premiums, rather than tive 36 percentage points. Better beta get-it approach. sectors, as the building blocks. We believe added another 32 percentage points on that isolating out and blending those top of that. While better beta has outperformed the exposures in a complementary way can index over the past 13 years, we believe unlock significantly greater portfolio In this period, the broad market index that it has even greater potential to efficiency. By taking a more controlled and did generate roughly the same risk- outperform in the coming years. The systematic approach to beta, investors can adjusted return as the barbell and the growing popularity of unconstrained improve their chances of adapting and better beta portfolio—mainly because its approaches today reflects the conviction thriving in today’s, and tomorrow’s, performance was dominated by falling that interest rates in the developed world fixed-income markets. n Treasury yields. But that bull run is are near a long-term bottom. We believe unlikely to be repeated. that, in a stable-interest-rate environment,

7 This page intentionally left blank.

8 Alison Martier Senior Portfolio Manager

Michael Mon Portfolio Manager Searching for investments that Dimitri Silva offer more upside than downside Portfolio Manager

investors stand a very good chance of Rethinking Alpha maintaining, and perhaps even improv- ing, alpha if they are willing to explore Alpha, or the value created from active investment manage- new ways to capture it. ment, is traditionally associated with the weighting of securi- We believe, based on our research and ties against an index. We believe that this view of alpha investment experience, that the way to needs to be expanded to reflect new market realities. preserve or improve alpha and risk-ad- justed returns in this environment lies not just in capturing upside, but also in limiting the potential downside of an After a 30-year bull run in bond markets, (and so limit the opportunities for investment opportunity. Such an fixed-income investors find themselves in investment) and put upward pressure on approach consists of generating an environment that is increasingly transaction costs. For many investors, asymmetric returns, or what the challenging their ability to maintain this adds up to a bleak prospect of fixed-income world calls positive alpha. For example, nominal yields are diminishing returns and sources of alpha. convexity. The concept of convexity can low and market liquidity has declined involve some complex mathematics, but (Display 1)—conditions which, respec- We do not accept that such an outcome here we use it in its simplest sense, to tively, lower the dispersion of returns is inevitable, however. In our view, describe the phenomenon where risk

Display 1: Challenges Abound for Fixed-Income Investors

Yields Are Low… …and Liquidity Has Declined US and German 10-Year Bond Yields Dealer Balance Sheets* as Percentage of Credit Outstanding 6.0 14 More US Treasuries Liquidity 5.0 4.0 3.0 7 l d (%) i e Percent Y 2.0 German Bunds 1.0 Less 0.0 0 Liquidity Jan 01 Jan 04 Jan 07 Jan 10 Jan 13 01 04 07 10 13

Historical analysis does not guarantee future results *Primary dealer balance sheets for positions greater than one year Treasury yields through November 29, 2013; primary dealer balance sheets through May 31, 2013 Source: Bloomberg, Haver Analytics, Investment Company Institute, Morningstar and AllianceBernstein

For institutional and financial professional use only. Not for inspection by, distribution or quotation to, the general public.

9 and reward are asymmetrical. By positive multisector investment universe that Expanding the Tool Kit convexity, we refer to an investment with spans credit ratings and capital struc- As part of this evolution, investors have more upside than downside, and by tures, and to investing in both physical gradually adopted broader investment negative convexity, we mean that the bonds and derivatives markets. As we universes—through global mandates, downside is greater than the upside. shall see, alpha opportunities exist across for example, or “core plus” strategies market sectors and geographical regions. that permit investments in international Convexity is particularly important for bond markets and securities with absolute-return investors who are Pricing anomalies have always existed in below-investment-grade ratings. But typically measured against a cash fixed-income markets. One example might derivatives are still regarded with a benchmark and who need to be able to be when a borrower issues new bonds healthy dose of skepticism, with many produce positive returns through the and investor demand causes the price of investors preferring to limit or restrict investment cycle. They aim to achieve the new bonds to exceed the value of the their use in portfolios. attractive returns with an eye on risk. issuer’s preexisting bonds (a phenomenon known in the US Treasury market as Some of this reluctance stems from The big question, of course, is: how can “on-the-run” vs. “off-the-run”). widely publicized financial scandals investors achieve such results? involving the use of derivatives by In credit markets, a similar opportunity “greedy” managers or rogue traders. It’s Below, we attempt to answer this can arise when a company loses its important to note, however, that question by looking at some of the investment-grade credit rating and derivatives are not bad in themselves; enhanced investment techniques that becomes a “fallen angel,” its bonds they are only as good or as bad as the can generate alpha with asymmetric relegated to high-yield status. Investors uses to which they are put. It’s true that return profiles. able to buy such securities can do so at a they can be used to leverage exposures steep discount to the bonds’ historical in a portfolio and so amplify the returns Global Opportunities—with Risk Control valuation, potentially making outsize gains of those investment ideas. If those ideas Alpha, traditionally defined as the value as selling pressure on the bonds eases and are flawed and if they have been created from active investment manage- the outlook for the company’s credit rating leveraged with derivatives without due ment, is usually associated with security improves. Between January 1997 and regard to the underlying risks, then the selection—that is, the overweighting and March 2011, for example, fallen angels in consequences can be damaging. But we underweighting of securities against an the US delivered total returns of 10.3% believe that ruling out, or severely index. But we believe that this way of compared with 6.5% for both investment- limiting, the responsible use of deriva- thinking about alpha needs to be grade bonds and bonds that were high tives in an investment strategy can be a expanded to reflect new market realities. yield at issue; volatility, respectively, was significant handicap, not just to captur- In our view, it’s necessary to think of alpha 11.4%, 5.7% and 9.8%.¹ ing better returns, but to embracing as having two equally important comp- better risk-control strategies. onents. The first is to identify a suitably Given the evolution that has taken place promising investment opportunity. The in global financial markets during the last This is because derivatives can be used as second is to approach the opportunity in 20 years—allowing investors to express a a to isolate a specific opportunity. such a way as to achieve asymmetric view in multiple ways (through physical For example, an investor might identify returns or positive convexity. bonds, futures or other derivatives corporate bonds as attractive but not markets)—one might assume such want to take on the interest-rate risk Identifying the Opportunity opportunities to have been arbitraged associated with those bonds. The The scope for identifying alpha opportu- away. In fact, the opposite is true. The investor could purchase the bonds and nities grows with the size of the confluence of new regulations, a then sell a to hedge the investment universe and with the range changing liquidity landscape and interest-rate risk, remaining exposed to of tools and instruments that investors individual investor preferences leave the just the movement in credit spreads. can use to implement an idea. The door wide open for those willing to Alternatively, if an investor is positive potential opportunities are greatest for research the opportunities. about the outlook for corporate bonds, investors who are open to a global, he or she can express that view by simply

1Source: Ridgemont Equity Partners

10 buying a call on an underlying Display 2: Convexity Is Valuable to Bond Investors asset, for example, an investor would gain the right to buy the asset at a predetermined . This would Bond maintain most of any upside gains while, Higher on the downside, restricting the maximum loss to the size of the premium paid for the option (Display 3). Price Putting the Theory to Work The ever-changing nature of the invest- ment landscape often presents oppor- Non-asymmetric Return Profile Lower tunities to construct investments with Lower Higher attractive, asymmetric return profiles. Level of Interest Rates These opportunities exist in environments of abundant or scarce liquidity, and they Highly simplified example for illustrative purposes only can be structured in such a way as to Source: AllianceBernstein make money irrespective of which way the market moves—thereby creating a greater universe of opportunities to executing an index credit default swap One way to improve upon a win-or-lose capture alpha. (CDS)² to gain exposure to a basket of proposition is to preserve most of the corporate issuers. upside potential of an investment when Let us examine some of these strategies the market rises, but limit the downside in detail. Creating Positive Convexity risk when the market falls. Alpha opportunities with positive Yield-Curve Opportunities: Taking a convexity can help mitigate the risks of An example of such a payout profile can Direction-Neutral View on Rates the linear return profile typically accepted be found in the use of options. By The natural shape of the yield curve—the by traditional long-only investors. A linear, or non-asymmetric, return profile is essentially a win-or-lose proposition. Display 3: Options Can Provide Downside Protection The purchase of a US$100 stock stands as much chance of making a large gain Profit/Loss as a large loss. An investor would lose Typical Long Position 100% of his investment if the price fell Long to zero, or double his money if the price rose to US$200. In the fixed-income Maximum Loss = markets, return profiles are less linear. Option Premium Indeed, they amount to negative Price asymmetry, given that the most a bondholder can hope for is a return of capital plus interest while the downside risk in the event of borrower default could be as much as 100%. Hence, the idea of positive convexity carries a particular appeal for fixed-income Highly simplified example for illustrative purposes only investors (Display 2). Source: AllianceBernstein

2Credit default swaps are a form of insurance for bond investors against the risk of losses on bonds they own.

11 positive-convexity characteristics. The Display 4: Yield Curves Revert to the Mean fact that the trade carries less risk than Difference Between Five-Year US Treasury Yields and Average of Two- and 10-Year Yields an outright duration bet also makes it attractive from a risk-adjusted perspec- 0.4 tive. A similar strategy could be con- structed using options. 0.3 0.2 Sector Opportunities: How a Barbell 0.1 Can Lift Performance 0.0 Another way of generating positive convexity can lie in constructing barbell Percent (0.1) strategies (which combine offsetting (0.2) exposures to different fixed-income sectors) (0.3) where they offer better risk-adjusted (0.4) returns than a single-sector exposure. For 1990 1993 1996 2000 2003 2006 2010 2013 example, the financial crisis of 2009 produced some notable opportunities in Historical analysis does not guarantee future results this respect. After a massive sell-off, Through September 30, 2013 corporate spreads had widened by Source: Bloomberg hundreds of basis points, and markets were pricing in default rates consistent with doomsday scenarios that even the line that describes the relationship in five-year yields relative to two- and most pessimistic of analysts saw as between bond yields of different 10-year yields, while movements below unlikely. Investors who bought corporate maturities—tends to be concave, the line reflect the opposite. bonds during the panic were rewarded irrespective of either the absolute level of handsomely as markets recovered, but market interest rates or where long-term One way to exploit this information is there was a more effective way to take yields stand in relation to short-term. through a combination of long and short advantage of the opportunity. From time to time, however, the yield positions known as a trade, in curve can become relatively linear in which a pair of trades on the two- and At the time, the extreme cheapness of shape. History shows that the US yield 10-year parts of the curve (the “wings”) corporate bonds was mirrored by the high curve tends to be strongly mean-revert- inversely complement a trade on the valuations of government bonds. Yields ing (that is, it has a strong tendency to five-year part (the “body”). When the on government debt were at all-time lows revert to a normal concave shape). This data series is below the median (that is, and, barring a doomsday scenario, it was can create an opportunity to profit from when the yield differential is negative difficult to see how a portfolio of changes in the shape of the yield curve, because five-year yields are lower than investment-grade credit could possibly at relatively little risk. the average of two- and 10-year yields), underperform government bonds. Many the time is right to take a short position investors saw significant return potential Display 4 shows a change in the in the five-year part of the curve and a in taking well-diversified positions in relationship between the nominal yield long position in the two- and 10-year investment-grade corporate debt without on the five-year US Treasury over time parts (in the expectation that yields will assuming what seemed to be undue risk. relative to its two- and 10-year counter- mean-revert and five-year yields will rise Not surprisingly, within the investment- parts (the line represents the difference relative to the two- and 10-year yields). grade corporate sector, financials were between the five-year yield and the A butterfly trade in the opposite particularly cheap. And within the average of the sum of the two- and configuration would be used when the financials sector, Tier 1 bonds—which are 10-year yields). Given the strongly yield differential is positive—that is, at the riskier end of a bank’s debt capital mean-reverting nature of the data series, five-year yields are higher than the structure—were trading at distressed any extreme move on either side of the average of two- and 10-year yields. prices close to 30 cents on the dollar. Such zero line has a reasonable probability of bonds were shunned by most investors at returning to a position closer to the line. The mean-reverting tendency of the the time, as they were seen as very risky, Movements above the line reflect a rise yield-curve shape gives the trade strong given the ongoing financial turmoil.

12 Thus, by bending the return profile in their Display 5: A 60/40 Portfolio Outperforms favor, investors could, in this case, have Projected Returns generated significantly higher returns without any meaningful increase in risk. 100% Corporates 60% Treasuries/40% Tier 1 150 Basis Trades: Exploiting Inefficiencies Between Bond and Derivatives Markets +50% Credit risk should, in theory, be priced 100 +22% identically in the physical bond market (17)% (19)% and the CDS market. But this is not

Percent always the case. For a variety of reasons, 50 including the inability to finance positions and unwillingness by investors 0 to realize losses, some physical bonds Crash Mar 09 Recovery Crash Mar 09 Recovery have limited liquidity compared with the Corporates Treasuries Tier 1 Banks CDS that references the same corporate entity. This could make it difficult to sell

As of March 31, 2010 the physical bond when, for example, Historical analysis is not a guarantee of future results. news headlines about the borrower raise Corporates refers to the corporate component of the Barclays Global Aggregate Index; Treasuries refers to US bonds of 25 years or more; Tier 1 Banks refers to Baa/BBB-rated Tier 1 bonds. concerns about its creditworthiness. In Source: Barclays, Bloomberg and AllianceBernstein such cases, the CDS market can become the sole means of hedging out credit risk. CDS spreads then tend to gap wider Instead of avoiding such an exposure that many investors had been expecting. than the spreads on physical bonds, altogether, however, an investor could Investment-grade corporates did well, creating a “positive basis.” use a barbell strategy, allocating 40% of returning 22%. But the 60/40 portfolio in the trade to these distressed bonds and our analysis returned 50%—even though Conversely, when an issuer raises more the remaining 60% to US Treasury US Treasuries, which represented more money in the bond market, supply and bonds. Our analysis showed that this than half the portfolio, suffered losses. demand can cause its bond spreads to combination would generate a more attractive risk/return profile than a diversified allocation to investment-grade Display 6: Changes in Basis Spreads Create Opportunities credit. In a scenario of economic J.P. Morgan High-Grade Credit Default Swaps/Physical Bonds Basis History recovery, our analysis indicated that the Tier 1 bonds would more than double in price, easily compensating for the 100 CDS Cheap modest decline in US Treasury prices that to Physical Bonds 50 would occur. In a “crash” scenario, assuming greater financial turmoil, the 0 Tier 1 debt would default and return (50) only an estimated 10 cents on the dollar, (100) Physical Bonds but US Treasuries would likely gain Cheap to CDS (150)

strongly in a flight to quality, mitigating Basis Spread total portfolio losses. In other words, the (200) trade would have more upside potential (250) Increased Volatility than a straight investment-grade credit (300) exposure, with broadly similar downside 05 06 07 08 09 10 risk (Display 5).

For illustrative purposes only As the financials sector stabilized, the Source: J.P. Morgan market delivered the recovery scenario

13 widen relative to CDS spreads, causing a “negative basis.” A negative basis can Display 7: Short Foreign Currency Trade vs. Options Put Spread also arise when financing becomes Payout Profile scarce, causing some investors to liquidate their physical positions—an 20 event that would drive volatility (Display 6, previous page). When such pricing 10 discrepancies occur, investors with access Strike Price to funding can capture the spread 0 (Bought Put) between the two markets by buying the Strike Price physical bond and buying CDS protec- (10) (Sold Put) tion (or vice versa) on the expectation Profit/Loss (Percent) that the basis will return to its normal (20) range. This is a classic arbitrage oppor- Lower Higher tunity in which the investor is not Price exposed to credit risk because, in the Short Foreign Currency Trade Options Put Spread case of default, the bond (long credit risk) and the CDS position (short credit Highly simplified example for illustrative purposes only risk) offset each other. Source: AllianceBernstein

Foreign Currency Opportunities: Improving on the Zero Sum alone, the foreign exchange markets are Take the example of a currency that has From time to time, economic turmoil or highly liquid compared with other asset deteriorating fundamentals. In this case, widespread deleveraging can lead to classes. Currency options markets are an investor could potentially generate significant volatility in the foreign also highly liquid, which facilitates the attractive returns by shorting it. But in exchange markets. With some US$1.5 use of derivative strategies to mitigate periods of high volatility, this strategy is trillion traded daily in spot markets risk and enhance returns. risky because a sudden reversal in the

Addressing Counterparty Risk

When using derivatives that are not cleared through an involves monitoring through a special-purpose committee using exchange, one important consideration is counterparty risk: the a process that incorporates fundamental credit analysis, including risk that a counterparty will fail to fulfill its contractual obliga- input from internal and external credit-rating resources. We also tions. Reform efforts in the US and around the world are driving recommend regular monitoring of market-based credit measures, the markets for derivatives toward central clearing. These efforts such as credit default spreads. should significantly mitigate counterparty risk associated with trading over-the-counter (OTC) derivatives, as well as increasing The posting of collateral is another powerful way in which transparency. While we welcome these developments, we counterparty credit risk is addressed in the swaps market. For believe that they should be regarded as enhancements to example, in standard single-name CDS contracts, AllianceBern- internal third-party counterparty credit-monitoring processes. stein requires the daily posting of collateral (typically cash or US Treasury bills) to offset changes in the mark-to-market value of When selecting counterparties, it is crucial that they are in good the CDS. In other words, as the CDS spread widens—indicating financial health with the ability to meet their commitments. that the reference entity is getting progressively closer to With this in mind, investors and their risk managers should default—the protection seller is required to post more collateral regularly monitor the counterparties with which they transact. to compensate for the increased risk of default by the reference Ideally this should be done through a risk-based monitoring entity. The process works in reverse when the spread narrows, approach that focuses more frequent monitoring on the largest with collateral being paid back to the protection seller. and most important counterparties. Best practice, in our view,

14 currency’s downtrend could trigger purchased. The investor still faces a approach this goal; indeed, there is more large losses. potential loss if the currency strengthens, than one way of thinking about alpha but the size of the loss would be reduced itself. We believe that the traditional way An alternative strategy would be to by the premium received. of thinking about alpha—simply as value purchase a , which would limit created by overweighting and under- the maximum possible loss on the trade In the event of the spot weighting securities against an index—is to the premium paid for the option. rising, the “plain vanilla” short position no longer adequate. would generate linear losses; with the The put option has an attractive put option, however, the loss would be Investors, in our view, should think about asymmetric structure in which the limited to the cost of the option. alpha more broadly as a dual process maximum potential gain is far greater consisting, firstly, of a search for opportu- than the maximum potential loss. In the event of the spot level falling (as nities across the fixed-income universe Capturing that benefit comes at a cost, the investor expects), then both the and, secondly, in terms of constructing however, and the option price could be outright short and the put option would strategies to take advantage of opportuni- high if the market expects volatility to be be profitable. By selling a put option ties in such a way as to achieve asymmet- high. It is possible, however, to reduce with the same but at a much ric returns or positive convexity, in which the net cost by creating an option lower strike price, we can mitigate some the upside potential is greater than the strategy known as a “put spread.” of the cost of the purchase of the put. downside risk. The key is to remain flexible While this also caps the upside, the and take advantage of as many sectors This involves the investor buying the same overall risk-adjusted return is attractive and investment methods as possible. put option while also selling a put option (Display 7). on the same underlying currency with a The ever-changing nature of today’s strike price further out of the money. In Conclusion fixed-income markets means that, with this strategy, the premium received for All active investment strategies share a the right analysis and risk manage- selling the out-of-the-money put would common purpose—to generate alpha. ment, there is always another new help offset the premium paid for the put But there is more than one strategy to opportunity ahead. n

15 This page intentionally left blank.

16 Seeking cost-effective tail-risk hedging opportunities across multiple asset classes

through July 2011. In a normal distribu- Taking the Sting out of the Tail: tion, a negative tail event should have occurred on about 28 days since 1928. Hedging Against Extreme Events In reality, extreme losses occurred about seven times as often, on 197 days.

Investors are still smarting from losses that—according to Portfolios are traditionally structured to widely held statistical assumptions—should have been cope with “normal” levels of volatility, usually in line with a long-run historical almost impossible. And they want protection against a repeat average, so portfolio construction and of such shocks in the future. We believe that the right tail-risk sources of diversification often fall short of expectations in periods of extreme hedging strategy can offer a solution. market stress. The average 60/40 portfolio lost more than 29% in the year ending February 28, 2009.1 Compound- ing the pain was the challenge of In our experience, many investors next page). A tail-risk event is defined as winning back lost ground: when a $100 remember 2008 as the year when the probability of an investment’s return investment declines to $70 (a loss of diversification failed. The global financial moving more than three standard 30%), investors need to generate a 43% crisis caused historical relationships deviations above or below the average in return just to break even. between many asset classes to break a given period. Assuming a normal down and correlations to rise, so that distribution of returns, the likelihood of a In recent months, more of our clients even carefully diversified portfolios three-standard-deviation event is roughly have been asking about protection suffered. Since then, recurring volatility 0.3%, or three times in every 1,000 strategies—whether they own traditional has driven investors to seek strategies occurrences. And the probability of a stock-and-bond portfolios or more that provide better protection, particular- negative tail event is half that number. complex asset classes such as hedge ly in periods of extreme market stress or funds. We think that when diversification “tail risk.” The idea of buying protection against falls short, tail hedging can be a solution. such a rare occurrence seems counter- There are many ways to approach tail The notion of tail risk refers to the “tail” intuitive, but history shows that real- hedging, ranging from a simple purchase associated with a normal (“bell curve”) world returns have not always behaved of “insurance” via put options to a distribution of returns. Most returns are like a normal distribution. This is portfolio of volatility-focused strategies clustered around the average, but some illustrated by the pattern of daily returns that spans multiple asset classes. fall well beyond the average (Display 1, for the S&P 500 from January 1928

160% S&P 500 and 40% Barclays Capital US Aggregate Index, rebalanced monthly, from March 1, 2008 to February 28, 2009. Source: Barclays Capital, S&P and AllianceBernstein

17 Equity Put Options Are Straight- purchase equity put options. These give helping to put a floor under potential forward but Can Be Costly the owner of the contract the right to losses if stock prices fall significantly. A One popular way of hedging tail risk is to sell at a specified price—effectively common approach is a “buy and hold” strategy with out-of-the-money put options (the strike price is below current equity prices). Display 1: Tail Risk in Theory: A Normal Distribution The value of a put option is driven by price movements and volatility. A fall in Lowest 0.15% Highest 0.15% equity prices below the strike price or a of Observations of Observations rise in volatility would increase the value of the option.

99.7% of Put options can be an expensive way to Observations hedge because the seller requires a risk premium. The option also has a time- value component, becoming less valuable as it approaches expiration (this makes intuitive sense: as time passes, the probability of stocks making a major Tail Tail move decreases). This decay accelerates as the contract gets closer to expiration.

Source: AllianceBernstein Because of this decay effect, a buy and hold strategy that entails rolling from one contract into the next on expiration has not always yielded the results an Display 2: Put Options: A Naive Buy and Hold Strategy Can Be Expensive investor might expect. Display 2 shows Value of $1,000 in Three-Month S&P Put Options the value of a $1,000 investment in two out-of-the-money put option strategies. From February 2006 through February 1,100 Strategy Was Profitable 2008, a buy and hold strategy would have detracted from portfolio perfor- 1,000 mance. While falling share prices and rising volatility made the option position 900 more valuable during the depths of the SPX D12.5 crisis, a “return to normal” saw those US Dollars 800 gains given back by the first quarter of 2009. So this was a relatively inefficient 700 Strategy Lost Money way of achieving a few months’ worth of downside protection. SPX D25 600 A More Nuanced Approach: 2005 2006 2007 2008 2009 2010 2011 2012 Identifying Shared Drivers of Risk In some ways, returns are about differ- February 2006–December 2012 SPX D25 represents a passive strategy where an investor buys and holds a three-month 25 Delta put option on ence while risk is about similarity. For the S&P 500, holding until maturity and enters a new transaction on each consecutive expiration date. SPX return-focused managers, success often D12.5 represents a passive strategy where an investor buys and holds a three-month 12.5 Delta put option on the S&P 500, holding until maturity and enters a new transaction on each consecutive expiration date. lies in identifying differentiating factors Source: Bloomberg, S&P and AllianceBernstein between investments. Traditional long-only investors look at distinct

18 opportunities in and between different asset classes.2 And they focus on different Display 3: Volatilities Are Highly Correlated: fundamentals: for example, equity investors might look at earnings growth 6 Equity Volatility (VIX) while corporate bond investors are more 5 Australian Dollar Volatility interested in the ability to repay debt. High-Grade Corporate Bond Spread 4 But for managers with a risk focus, the Emerging-Market Debt Spread 3 answers often lie in the similarities. For example, as recent events illustrate, 2 volatility is a common link between asset Z-Score classes (Display 3), and the returns on 1 different asset classes can be highly 0 correlated in times of crisis. (1) Under normal circumstances, different (2) asset classes are driven by different sets of factors, which is what makes them 1996 1998 2000 2002 2004 2006 2008 2010 2012 good diversifiers. But there is usually an overlap; for example, the prices of most Through December 31, 2012 assets are influenced by macroeconomic Historical analysis is not a guarantee of future results. factors such as interest rates and GDP All time series are shown in Z-score units, standardized for the period January 1996–August 2011. A Z-score is a measure of the distance from the mean of a distribution normalized by the standard deviation of the growth. In a crisis, the usual drivers of distribution. performance may be superseded by new Z-scores help quantify how different from normal a recorded value is. Equity volatility is represented by the CBOE Volatility Index (VIX), Australian dollar volatility by one-month factors that affect all asset classes in a on AUD/USD, and debt volatilities by option-adjusted spreads on the Barclays Capital US similar way. Corporate Index and Barclays Capital Emerging Markets Index. Source: Bank of America Merrill Lynch, Barclays Capital, Bloomberg and Chicago Board Options Exchange

For example, changes in market liquidity can affect multiple asset classes at once. And given that many assets exploit some kind of risk premium—such as the value The goal is to identify investments and put options. For example, unnecessary premium in equities or the interest-rate instruments that are highly price sensitive costs can often be avoided by taking a differential (“carry”) between curren- and can be used to balance these more active approach, in which the cies—when risk aversion spikes, the exposures. In this way, when volatility manager continuously monitors returns from most active strategies suffer. spikes, gains from tail-risk hedging should volatility curves and rolls out of one help to offset losses that accrue in a position into another well before the As a result of these relationships, we portfolio during periods of market expiration date. turbulence. We believe that the best believe one of the keys to success is to n Call options on equity volatility (such be volatility centric—in other words, to results can be achieved by seeking hedging strategies across multiple asset as options on the Chicago Board identify cost-effective ways in which an Options Exchange VIX index) benefit investor can be “long volatility.” classes, dynamically allocating funds across these opportunities in a way that when volatility rises. Ways to Go “Long Volatility” seeks to minimize the cost of implementa- n In , options can be used to When investors own assets that carry a tion. Examples of such strategies include: construct strategies. One example is risk premium, such as emerging-market n In equities, put options benefit when an “anti-carry” trade that takes a long currencies or high-yield bonds, we think equity prices fall and volatility rises. As position in low-yielding currencies like of them as being “short volatility” discussed earlier, a pure buy and hold the Japanese yen and a short position because when volatility rises, they are strategy can be expensive. But there in a high-yielding currency like the likely to lose money. are more cost-effective ways to use Australian dollar. The idea is that

19 strategies can vary greatly. So it is a Display 4: Implied and Actual Volatility Can Diverge Significantly: Realized (Trailing) question of identifying where protection Volatility and Implied Volatility of 10-Year Treasury Yields is most reasonably priced.

100 Taking an example from the interest-rate 90 = Actual Volatility > Implied Volatility Actual options market, Display 4 shows the Risk “Underpriced”/Protection “Cheap” 80 Implied actual volatility of 10-year Treasury yields since 2000, compared with the level of 70 volatility implied by option pricing 60 beforehand. At many points in time, 50 actual volatility turned out to be higher than implied volatility. This suggests that Percent 40 risk was underpriced and protection was 30 relatively cheap. At other times, implied 20 volatility was higher than actual volatility, 10 suggesting that protection was expensive and that, effectively, the risk premium 0 was too high. 2000 2002 2004 2006 2008 2010 2012 This ebb and flow of costs highlights April 2000–December 2012 how important it is to seek opportunities Historical analysis is not a guarantee of future results. Implied volatility represented by three-month options on 10-year swap rates. Source: Bloomberg and across multiple asset classes, rather than AllianceBernstein relying on a single strategy.

Display 5 arranges the same data in a different way, showing the percentage of higher-yielding currencies typically basket of liquid securities. Insurance time that actual volatility turned out to underperform safe-haven currencies against the first five defaults costs more be much lower (the left side of the curve) when market participants get more than insurance against defaults by the or higher (the right side) than implied risk averse. next five and so on, because the volatility. The distribution of the data is likelihood of many companies failing is n In fixed income, interest-rate swaps skewed, with more extreme values falling lower. In a stable economy, it can be very can be used to take interest-rate to the right. In other words, especially at cost-effective to buy insurance against exposure (duration) so that the times of turbulence, the interest-rate an extreme scenario. portfolio benefits when yields fall, for options market has tended to underesti- example when panic triggers a flight mate how volatile yields are going to be, into Treasuries. Protection at the Right Price and “tail options” (insurance against Like any other asset class, some of the extreme losses) have been priced n Interest-rate swaps can be used to instruments that provide protection are too cheaply. construct a yield-curve flattener—for more expensive than others, and example, a short position in two-year in-depth research can identify pricing The skewness also implies that because yields and an equal long position in anomalies and opportunities. the market is not perfectly efficient in 10-year yields. Yield curves tend to pricing risk, investors can use strong flatten in an economic crisis, with While different asset classes tend to research to gain an edge. longer maturities outperforming as experience similar levels of volatility at growth and inflation expectations fall. times of extreme market stress, there can Identifying Cost-Effective Strategies n In credit, credit default swaps (which be a wide dispersion under more normal We have researched and implemented function like put options) benefit if market conditions. This implies that ways to manage the cost of tail risk corporate bond values fall. Credit default while all forms of tail-risk protection by constructing strategies that allow swaps can also be purchased to insure become expensive in a crisis, for the rest investors to be “long volatility” in a against default in different tranches of a of the time the cost of tail hedging very cost-efficient manner.

20 Display 5: Distribution of Differences Shows That Tail Risk Can be Cheap: Realized (Trailing) Volatility Less Implied Volatility of 10-Year Treasury Yields

Actual Volatility < Implied Volatility Actual Volatility > Implied Volatility Risk “Overpriced” Risk “Underpriced” Protection “Expensive” Protection “Cheap” 800

600

400 Occurrences

200

0 (36) (33) (30) (27) (24) (21) (18) (15) (12) (9) (6) (3) 0 3 6 9 12 15 18 21 24 27 30 33 36

Current analysis is not a guarantee of future results. April 2000–December 2013 Implied volatility as represented by three-month options on 10-year swap rates Source: Bloomberg and AllianceBernstein

One way to do this is making an to default even in normal conditions. protect against tail risk at times of crisis, exchange, in which one can trade off Buying protection on the least risky while actively managing any negative (sell) insurance against higher-frequency/ tranche is cheaper because the market carry costs associated with the protection. low-impact events in order to fund the doesn’t expect many defaults unless a purchase of insurance against the much extreme scenario unfolds. Put options on We believe that an effective tail-risk more damaging lower-frequency/high- these “super senior“ tranches become hedging strategy can help investors by impact events. This opportunity presents more valuable at times of market stress buffering their portfolios when extreme itself in many different markets. because people start worrying about the events occur. We recommend maximiz- potential for high numbers of correlated ing the opportunity set by researching This could be compared to an automobile or systematic defaults. strategies across asset classes. insurance premium with a high deduct- ible. In insurance, the higher the deduct- There are many other ways to hedge in a For best results, we advocate an agile, ible, the lower the annual premium cost-efficient manner, including exploit- active approach that dynamically charged by the insurance company. If you ing term structures in rates markets. For allocates funds between strategies have a high deductible, you are willing to example, earlier we referred to a depending on current market pricing and cover the costs of dents, scratches and yield-curve flattener. If yield curves are prevailing conditions.n other minor incidents (high frequency, low particularly steep, this position can pay impact) in order to have a lower premium for itself because long maturities for collision and other major accidents compensate investors so well in the form (low frequency, high impact). of yield and roll return.

Another example would be the credit In Summary tranche strategy discussed previously. While portfolio diversification has served Buying protection on riskier tranches is investors well under “normal” market expensive (and selling it is lucrative) conditions, it may fall short in periods of because the market expects a few names extreme market stress. The challenge is to

21 AllianceBernstein L.P. AllianceBernstein Canada, Inc. AllianceBernstein Hong Kong Limited 1345 Avenue of the Americas Brookfield Place, 161 Bay Street, 27th Floor Suite 3401, 34/F New York, NY 10105 Toronto, Ontario M5J 2S1 One International Finance Centre 212.969.1000 416.572.2534 1 Harbour View Street, Central, Hong Kong +852 2918 7888 AllianceBernstein Limited AllianceBernstein Japan Ltd. 50 Berkeley Street, London W1J 8HA Marunouchi Trust Tower Main 17F AllianceBernstein (Singapore) Ltd. United Kingdom 1-8-3, Marunouchi, Chiyoda-ku 30 Cecil Street, #28-08, Prudential Tower +44 20 7470 0100 Tokyo 100-0005, Japan Singapore 049712 +81 3 5962 9000 +65 6230 4600 AllianceBernstein Australia Limited Level 37, Chifley Tower, 2 Chifley Square AllianceBernstein Investments, Inc. Sanford C. Bernstein & Co., LLC Sydney NSW 2000, Australia Tokyo Branch 1345 Avenue of the Americas +61 2 9255 1200 Marunouchi Trust Tower Main 17F New York, NY 10105 1-8-3, Marunouchi, Chiyoda-ku 212.969.1000 Tokyo 100-0005, Japan +81 3 5962 9700

© 2014 AllianceBernstein L.P. This document is confidential and proprietary information and should not be disclosed to any other person without the prior written consent of AllianceBernstein Limited.The contents are being provided to you solely for informational purposes and as an example of our internal research. It is not intended to be an offer or solicitation, or the basis for any contract to purchase or sell any security, product or other instrument, or for AllianceBernstein to enter into or arrange any type of transaction as a consequence of any information contained herein. The views and opinions expressed in this document are based on AllianceBernstein’s internal forecasts and should not be relied upon as an indication of future market performance.

A Word About Risk Market Risk: All investments involve risk. The market values of a portfolio’s holdings rise and fall from day to day, investments may lose value. Derivatives Risk: Derivative instruments such as options, futures, forwards or swaps can be riskier than traditional investments, and may be more volatile, especially in a down market.

Note to All Readers: The information contained here reflects the views of AllianceBernstein L.P. or its affiliates and sources it believes are reliable as of the date of this publication. AllianceBernstein L.P. makes no representations or warranties concerning the accuracy of any data. There is no guarantee that any projection, forecast or opinion in this material will be realized. Past performance does not guarantee future results. The views expressed here may change at any time after the date of this publication. This document is for informational purposes only and does not constitute investment advice. AllianceBernstein L.P. does not provide tax, legal or accounting advice. It does not take an investor’s personal investment objectives or financial situation into account; investors should discuss their individual circumstances with appropriate professionals before making any decisions. This information should not be construed as sales or marketing material or an offer or solicitation for the purchase or sale of any , product or service sponsored by AllianceBernstein or its affiliates.

Note to Canadian Readers: This publication has been provided by AllianceBernstein Canada, Inc. or Sanford C. Bernstein & Co., LLC and is for general information purposes only. It should not be construed as advice as to the investing in or the buying or selling of securities, or as an activity in furtherance of a trade in securities. Neither AllianceBernstein Institutional Investments nor AllianceBernstein L.P. provides investment advice or deals in securities in Canada.

Note to UK Readers: This information is issued by AllianceBernstein Limited, a company registered in England under company number 2551144. AllianceBernstein Limited is authorised and regulated in the UK by the Financial Conduct Authority (FCA – Reference Number 147956). This information is directed at Professional Clients only.

Note to Japanese Institutional Readers: This document has been provided by AllianceBernstein Japan Ltd. AllianceBernstein Japan Ltd. is a registered investment-manage- ment company (registration number: Kanto Local Financial Bureau no. 303). The firm is also a member of Japan Investment Advisers Association and the Investment Trusts Association, Japan.

Note to Australian Readers: This document has been issued by AllianceBernstein Australia Limited (ABN 53 095 022 718 and AFSL 230698). Information in this document is intended only for persons who qualify as “wholesale clients,” as defined in the Corporations Act 2001 (Cth of Australia), and should not be construed as advice.

Note to New Zealand Readers: This document has been issued by AllianceBernstein New Zealand Limited (AK 980088, FSP17141). Information in this document is intended only for persons who qualify as “wholesale clients,” as defined by the Financial Advisers Act 2008 (New Zealand), and should not be construed as advice.

Note to Singapore Readers: This document has been issued by AllianceBernstein (Singapore) Ltd. (Company Registration No. 199703364C). The Company is a holder of a Capital Markets Services Licence issued by the Monetary Authority of Singapore to conduct regulated activity in fund management.

Note to Hong Kong Readers: This document is issued in Hong Kong by AllianceBernstein Hong Kong Limited, a licensed entity regulated by the Hong Kong Securities and Futures Commission. This document has not been reviewed by the Hong Kong Securities and Futures Commission.

Note to Readers in Vietnam, the Philippines, Brunei, Thailand, Indonesia, China, Taiwan and India: This document is provided solely for the informational purposes of institutional investors and is not investment advice, nor is it intended to be an offer or solicitation, and does not pertain to the specific investment objectives, financial situation or particular needs of any person to whom it is sent. This document is not an advertisement and is not intended for public use or additional distribution. AllianceBernstein is not licensed to, and does not purport to, conduct any business or offer any services in any of the above countries.

Note to Readers in Malaysia: Nothing in this document should be construed as an invitation or offer to subscribe to or purchase any securities, nor is it an offering of fund-management services, advice, analysis or a report concerning securities. AllianceBernstein is not licensed to, and does not purport to, conduct any business or offer any services in Malaysia. Without prejudice to the generality of the foregoing, AllianceBernstein does not hold a capital-markets services license under the Capital Markets & Services Act 2007 of Malaysia, and does not, nor does it purport to, deal in securities, trade in futures contracts, manage funds, offer corporate finance or investment advice, or provide financial-planning services in Malaysia.

MSCI Note: MSCI makes no express or implied warranties or representations and shall have no liability whatsoever with respect to any MSCI data contained herein. The MSCI data may not be further redistributed or used as a basis for other indices or any securities or financial products. This report is not approved, reviewed or produced by MSCI.

13–3074 INS–7076–0314 www.alliancebernstein.com