The Long and Short of Portfolios and Liabilities Matching

Transcription

1 McVite Equity Group Newsletter 3rd Quarter Singapore, July 18, 2017 The Long and Short of Portfolios and Liabilities Matching Rates will not stay at historic lows forever and as global markets enjoy a period of synchronized growth, monetary policy is changing to reflect that. As a consequence government bond yields will inevitably follow suit and this has clear implications for pension fund investors, who could see their capital eroded. Defined Benefit (DB) liabilities are intrinsically linked to government bond yields and, as committed fixed income investors, schemes will need to revisit their investment strategies. One possible solution to the threat of rising yields is to allocate part of your assets away from longer-dated bonds to their shorter-dated counterparts. Long-dated corporate bonds have historically provided a useful match for the liability profile of many DB schemes. However, they are also much more sensitive both to changes in gilt yields and to overall market volatility. Re-orienting towards corporate bonds with a shorter maturity (typically defined as five years or less) will provide investors with not only a lower duration but also a lower spread duration. This means that the bonds will exhibit less sensitivity to changes in government bond yields as well as credit spreads. Short duration bonds can therefore be a useful tool in managing the risks of rising yields and market volatility. Central banks will be withdrawing liquidity and interest rates will be on the rise. More than de-risking Investing in short duration bonds offers sensible de-risking opportunities but can also provide much needed returns. For investors who have been hoarding cash but have been

2 unrewarded for doing so, they might want to re-risk their portfolio by moving into short duration bonds. Essentially this is the next step up from cash on the risk ladder since short duration bonds typically offer a greater reward than cash but minimize interest rate risk and the impact of market volatility. And while less liquid than cash by definition, a well-designed short duration bond strategy will have an attractive liquidity profile. Regular cash flows from maturing bonds and coupon income enables a robust short duration strategy to minimize turnover while still implementing active strategies. This has the advantage of helping to keep transaction costs low. Going Global While attractive opportunities still exist in the short-dated sterling credit market, investors are also increasingly looking overseas to capture the potentially higher returns available from global corporate bonds. Broadening the bond allocation into global short duration provides diversification away from sterling, and for investors looking to take on more risk, including high yield bonds and emerging markets may also offer additional risk-return benefits. Pension funds have already started to embrace short duration strategies as a way to retain their exposure to credit markets while reducing volatility and interest rate risk. As yields rise and volatility returns, more schemes will likely follow suit, keen to protect against the gradual renewed hawkishness from central banks. It is important to recognize that a range of short duration strategies can help pension schemes improve their risk-return profiles Cash-flow Driven Investment (CDI) and Liability Driven Investment (LDI) Cash-flow driven investment (CDI) is a system of portfolio management that accepts and builds upon liability-driven investment (LDI). The idea, in both cases, is to match all future projected liability payments as they fall due, which is by definition the imperative of pension funds and the like. But, CDI is not merely a variant on growth-plus-ldi. Rather it involves managing fixed income assets on a buy and maintain ((B&M)) basis and holding them to maturity, or at least only substituting bonds with the same cash-flow profile. Back in 1974 Robert C. Merton, an American Economist, designed a model for explaining credit risk premiums. He said that they can be understood as the results for a holder of a corporate bond who sells a put option on the value of the underlying enterprise, with the strike to the put serving as the default threshold of the enterprise.

3 One can understand CDI through the same theoretical framework. If a portfolio manager is going to hold a certain 10 year corporate bond to maturity, he will know the absolute maximum return that asset can yield, the amount promised. Likewise, he will know the minimum result (($0)) in the unhappy event of quick default and unsuccessful recovery. If the fund manager holds a number of such bonds, he can add the promised results together and, again, he knows the maximum. The Arithmetic The portfolio manager can also know that the difference between the actual outcome and that maximum will be determined by losses on defaults. For a large portfolio of bonds this is reasonably predictable and can be reserved for relatively easily. It involves netting out the recoveries one will get on some of those defaults but, again, there is nothing mysterious or especially speculative about the arithmetic. Moody's data suggests an expected 40% recovery on defaults. The following graph demonstrates the distribution of cumulative defaults over ten-year rolling periods. In nearly 40% of the periods studied (over the last 90 years) bond defaults have stayed at or below 0.1%. The credit spread, that is, the premium earned by selling the hypothetical put in the Merton model as explained above, will on average exceed the expected loss in the underlying asset, the equities. This is true regardless of whether or not volatility itself is earning a premium and regardless of whether or not there is a skew in the volatility assumption.

4 This feature of the model means that the expected losses on default will be below the premium earned, which is where one wants it. What about those proposed 100-year bonds to finance long liability infrastructure or higher risk income assets? Is the 100-year bond a potential investment for pension funds? What about 50-year bonds? Some institutional investors said 100 years is simply too long a period to hold an investment, while others had concerns the cash flow profile would not match pension liabilities particularly well. Oddly, an irredeemable bond might be a better match. The only reason schemes are invested in bonds at present is a desire to match accounting values of liabilities. This is simply not rational at current yields. A 100-year bond would be too long to match liabilities and would be massively exposed to price falls when interest rates rise. Defined benefit pension schemes are thirsty for duration their liabilities typically have a duration of years. If they invest 50% of their assets in bonds then they ll be wanting a duration of years. Longer-dated bonds are useful, particularly for trustees that don t feel comfortable with the use of derivatives. It depends on the way yields go. If you think rates may go even lower then such a bond, with its long duration, is a great investment. Under such falling rate environments, portfolio optimizing schemes should look at increasing their allocation either cautiously or boldly diversification should be the reason for doing so. Potentially, with safeguards, these 50 t0 100 year bonds could be an attractive asset class for steady and solid investment returns that helps diversify a portfolio. Infrastructure investment potentially falls between two stools: it s not effective as part of the hedging assets of a scheme bonds and swaps are far better, and it probably does not have an expected return high enough to justify inclusion in the growth part of the portfolio either. A far stronger case might exist for active management in high-yield credit than for developed-market equities. Passive management could be viewed as investing in credits that are effectively lending money to everyone who wants to borrow, in the proportion that they want to borrow. Some active management presents a far better opportunity, but at times can struggle to compete with indices. High-yield bond managers can struggle to beat their indices. They focus on avoidance of defaults while maximizing yields, and underperformance often occurs when the riskier CCC rated part of the market outperforms. That said, high yield is a relatively inefficient part of the fixed-income universe, and active management is certainly warranted in that schema.

5 When attempting to match or beat an index, fees will always act in favor of the indices. But, the more important question is; what will the return be on the indices, and to answer that we need to define what time horizon we are considering. So long or short duration fixed income portfolios, risker asset classes, matching or unmatched liabilities, hedging strategies all begs a highly sophisticated analysis of term structure and optimal allocation. One might start by consulting the works of economists Jack Hirshleifer On the Theory of Optimal Investment Decision and Reuben A. Kessel on the Cyclical Behavior of the Term Structure of Interest Rates to understand the complexity of the problem and as well as appreciate these examples of elegance in presentation. We can help you in evaluating optimal portfolio duration and liability matching. Contact Us: info@mcviteequity.group