Better core allocations: Feasible modern portfolio theory via risk parity
By Roberto Croce and Lee Partridge | January 3, 2013 12:01 am
Too many investors are missing out on the only free lunch in finance: diversification.
Systematically rebalancing to a diversified portfolio typically leads to a reduction in drawdowns during periods of financial stress relative to a concentrated, equity-heavy portfolio. The notion that correlations go to one, or converge, during such periods ignores safe-haven assets for which correlations approach minus one, or inverse relationships. Simultaneously, portfolio risk is too often treated as a consequence of asset allocation rather than as a critical input.
Risk parity refers to a portfolio where each asset contributes equally to portfolio risk. Risk parity is generally implemented using asset classes that behave differently across a variety of economic environments. Risk-parity managers also generally target a specific level of annual portfolio volatility, properly treating risk as an input when determining their asset allocations.
The investment management industry has paid a lot of attention recently to risk parity, and we believe this a step toward implementing a feasible version of modern portfolio theory and its extensions, first introduced by Harry Markowitz in 1952 and later incorporating the ideas of William Sharpe in the 1960s. MPT guides investors to hold the single, “optimal” portfolio that offers the highest level of risk-adjusted return available given their opportunity set, and is determined as the point of tangency between the risk-free rate and the efficient frontier. An investor practicing MPT and desiring a level of risk lower than that of the optimal portfolio would hold a mix of cash and the efficient portfolio. Similarly, investors desiring higher levels of risk and return would borrow funds and allocate more to the optimal portfolio than their capital alone could finance. The central tenet of MPT is that increasing or decreasing exposure to the optimal portfolio, in the manner described above, is preferable to changing allocations along the efficient frontier. Financial innovation in the form of futures and swaps offers the opportunity to make the practice of MPT more efficient, if only we could reliably identify the optimal portfolio.
We believe risk parity offers a feasible alternative. Financial theory tells us that a risk-parity portfolio formed with assets that have equal long-run, risk-adjusted returns and moderate correlations to one another offers risk-adjusted returns similar to those of the optimal portfolio. Because volatilities — and to a lesser degree correlations — are more persistent than returns over time, one can build portfolios that target equal risk contribution from underlying components with reasonable success. Salient's internal research shows that between 1950 and 2011, a na´ve risk-parity strategy allocating to global equity, interest rates and commodity futures performed on par with the long-run optimal asset allocation, which was determined with perfect foresight. Because risk parity can lead to a decent proxy for the optimal long-run asset allocation, many investors should consider viewing it as the starting point for a feasible implementation of MPT.
With interest rates near historic lows, critics often point to the large interest rate exposures common in many risk-parity portfolios as a rationale for avoiding the strategy. Our experience suggests, however, that large tactical tilts in core portfolios — even when they make a lot of sense intuitively — are the most common source of investor misery.
A second criticism of risk parity is that managers generally use the economic leverage embedded in futures contracts and swaps to amplify the return profile of their risk-parity strategies (alternatively, managers could borrow money in order to invest more than 100% of their capital in individual cash securities or hold a mix of derivatives and cash securities).
Our response to this is twofold. First, risk parity need not employ leverage. We find that unleveraged risk-parity strategies generally compounded returns around 7% per year since 1990 with annualized volatility around 4%. It is our experience, however, that investors generally desire a higher level of expected return and volatility than this from their core portfolio.
Further, it is our view that investors are typically better off increasing their holdings of a well-diversified portfolio — even if that means employing leverage — than concentrating them in assets with higher volatility and expected returns. In other words, we believe the modest and prudent use of leverage is often times preferable to pinning our ability to meet future obligations on hopes that equity markets will rally indefinitely. We think this holds true regardless of whether investors are targeting annual returns of 8% or 18%. In practice, we find that a risk-parity strategy targeting 10% volatility generally requires notional exposures that fluctuate between 200% and 400%, with median exposure around 300%. Consistent with these views, we believe core portfolios should be based on a systematic implementation of risk parity that is free of tactical tilts. Even without such discretionary tilts, risk-parity allocations offer the potential to be quite dynamic if practitioners use a sensitive risk model to determine their allocations.
If there is any art in building a risk-parity strategy, it is in choosing which assets ought to be included. Using assets that do not offer attractive long-run risk-adjusted returns will typically produce portfolios that underperform, while failure to include asset classes that diversify one another will frequently lead to the same concentration problems risk parity is supposed to alleviate.
Risk parity is no panacea, but it serves as a starting point for managers looking to build diversified core portfolios that are designed to perform across macroeconomic environments.
Roberto Croce is director of quantitative research and Lee Partridge is chief investment officer of Salient Partners LP, Houston.