Gearing-up on bonds looks remarkably like LDI, notes Gwion Moore. So what does this tell us about the suitability of risk parity for an investor whose starting point is a significant short position in long-dated interest rates?
For a pension scheme, the low risk/high return performance of risk parity managers is an attractive contrast to the enormous swings in funding ratio experienced by most pension schemes. As an investor for defined benefit pension schemes, I immediately thought, "Why can't I do that?"
So I investigated and came to some surprising conclusions. For the average asset and liability investor, applying risk parity principles is not so different from following the more familiar principals of liability-driven investment (LDI) and return diversification, but with a focus on longer-term investing. The real question that risk parity poses to pension scheme concerns the role of leverage in the asset and liability portfolio.
The basic principle of risk parity is that many investors' exposures are overwhelmingly dominated by equity market risk. Part of the problem is that equity markets are so much more volatile than most other asset markets. As much as 90% of the risk for a portfolio split 50%/50% between equity and bonds comes from the equity allocation. Reducing exposure to equity and increasing exposure to bonds reduces the level of return the investor can achieve, but the answer to this is to take a better mix of risks and lever up the exposures until the desired level of return is achieved.
In practice this means levering the exposure to interest rate risk, via bond futures, repo, swaps and so on - effectively borrowing cash to buy bonds. The expectation based on both financial theory and a fair amount of historical evidence is that bonds will on average out-perform cash by about 1% per year - the interest rate risk premium. The usual explanation is that the lender of money will, on average, demand more compensation to provide security of funding for longer periods. Alternatively, the lender will need to be compensated for the interest rate duration risk for lending longer. Likewise, equity is expected to out-perform cash (the equity risk premium) and credit is expected to out-perform government bonds (the credit risk premium). The principle of risk parity is that by bringing the different market risks into roughly equivalent sizes is a robust portfolio construction rule-of-thumb. However, it is difficult to argue from theory that any two risk factors should be equal in size - instead, a better approach is to generalise the concept to mean the use of leverage to improve investment performance.
How does this principle apply to pension schemes and how does it contrast with the approach taken by more absolute return-focused risk parity managers? Starting simply, we can probably decompose 80-90% of the market risk exposures to the funding ratio of a pension scheme into four broad categories: nominal and real interest rate risk, credit risk and equity risk. Of course we could include property, currency, volatility, illiquidity, perhaps mortality risk - but these four are a good start. It is the balance and directional exposure to these risks that define the investment strategy of a pension scheme. Figure 1 shows a stylised comparison of the investment strategy of a risk parity investor versus a pension scheme investor.
The risk premia exposures of the risk parity investor are positive and in rough balance, whereas the pension scheme investor has large negative exposures through its liabilities to real and nominal interest rate risk. The pension scheme investor also has a large exposure to equity and a comparatively small exposure to credit and other risk premia. To adopt an investment strategy for a pension scheme that is employing risk parity principles would involve:
• Hedging more liability risk;
• Increasing credit and other-alternative risk exposures;
• Decreasing equity risk exposure.
These principles are very much part of the investment trend of pension schemes over the past decade towards LDI and greater diversification.
In practice, the special brew of any particular risk parity provider is in the balance in which these risk exposures are taken, the degree to which classes of risk premia can themselves be better diversified, and the inclusion of more exotic risk premia. However, for the pension scheme investor the most interesting part comes when we look at the difference in the treatment of interest rate and inflation risk between pension and risk parity investors.
Typically, the use of leverage by risk parity providers is significantly focused on increasing the exposure of the funds to nominal and real interest rate risk premia. The duration component often provides counter-cyclical returns and significant diversification benefit, while the real exposure provides protection against out breaks of inflation. In addition to the difference in directional exposure to real and nominal interest rate risk premia another key difference between absolute return-focused risk parity investors and pension schemes is the duration of the interest rate exposures.
For an absolute return investor the trade-offs between carry and duration exposure to interest rate risk premia can be quite different from those of a pension scheme investor. For an absolute return investor the focus will be on short/medium-term duration exposures. These maturities provide the best balance between carry and the counter-cyclical impact of duration driven returns. However, pension schemes normally have exposure to very long dated liabilities, and so long-durations hedges are used, leading to a more duration and less carry-focused exposure.
So, pension schemes use leverage to partially close a very large negative duration exposure, while risk parity providers use leverage to take a positive duration exposure with a volatility comparable to that of the equity exposure. If we assume that the pension scheme has diversified its growth portfolio and has the same leverage constraints as a risk parity fund, then we can compare in a stylised fashion the different approaches to investing in the form of the traditional risk-return trade-off graph. Figure 2 has two graphs, one showing risk and return from a funding-ratio perspective and the other from an absolute-return perspective. On each graph is plotted the risk and return positions of the risk parity provider and a pension scheme investor. The risk and return axis of the two graphs do not have the same scale.
From the perspective of an ALM pension investor, risk parity is sub-optimal, and vice versa. In practice the funding-ratio information ratio of most growth focused pension schemes is much lower than the absolute-return information ratio of risk parity investors. Very few pension schemes have fully hedged their liabilities, leaving most pension schemes with a negative exposure to the interest rate risk premia. If a pension scheme were to try and make its funding ratio behave like the returns of a risk parity provider, it would have to hedge all of its liabilities and then take on some more interest rate risk. This would imply a degree of leverage that would make both pension schemes and risk parity investors uncomfortable.
In his recent article on risk parity, Cliff Asness argues that leverage aversion changes the predictions of modern portfolio theory - "It implies that safer assets must offer higher risk-adjusted returns than riskier assets" - and explains the historic out-performance of risk parity strategies.
So, the real question risk parity poses to defined benefit pension schemes is, how should they best deploy leverage and derivatives to meet multi decade-long commitments. The answer to this is a complicated one as it as much about pension schemes implementation and risk management capabilities, internal expertise and cost as it is about innovation in investment techniques.
Gwion Moore is head of investment strategy UK at MN