Sometimes even Cassandra catches a break. Two years ago we warned of the dangers of ‘risk parity’, an investment approach balancing portfolio allocations by risk before leveraging up the result, and with recent market stress it has been back in the headlines (including here and here) as investors start to question the strategy’s core rationale. Indeed, annualized performance for AQR’s Risk Parity Fund (one of the industry’s mainstays) has been just 1.8% since we wrote our post, as compared with 10.0% for the S&P 500 equity index. With annualized volatilities of 7.2% and 13.0%, respectively, this implies risk-adjusted returns (Sharpe ratios) have been more than three times lower in Risk Parity than equities.
At its heart there is nothing radical or new about risk parity – it is not far from Finance 101’s ‘optimal portfolio theory’: take a basket of asset classes, weight them such that they contribute equally to portfolio risk, and leverage up to your desired return. Consistency of asset-class behavior is key to this being successful, including having visibility of future risk and correlation (something that is, putting it mildly, far from guaranteed).
A major problem lies in how time-varying correlations can be, especially between the core components of equities and bonds. In the US, monthly correlation between stocks and bonds has veered between -0.9 to +0.4 between January 2000 and August 2015, with no clear path in-between. Then there is the perennial elephant in the room – liquidity. You simply cannot include assets that aren’t readily convertible into cash in a risk parity portfolio at the constant mercy of margin calls, and this excludes sources of real value for long-term investors who have a competitive advantage as liquidity providers. We would never advocate that long-term investors go without truly differentiated, illiquid assets, capable of delivering returns over time – but this is incompatible with a risk parity approach.
But setting aside these shortcomings, our real issue lies with risk parity’s overriding message: investment diligence can be avoided, fundamental analysis is unnecessary, and portfolios can be built mechanically for the long term. It is true that risk parity performs in back-tests, so this is a tempting enough viewpoint. But unfortunately reality is forward-looking, and levering up assets prone to bouts of positive correlation is always going to be a hard sell, especially today with bonds looking anything but cheap.
The fact is that bond yields haven’t been this low since before Napoleon, and no models, no back-tested simulations, can accurately portray the risk of leveraged duration bets in this environment. Risk parity certainly has not ‘failed’, but it is our view that there are no formulaic approaches to investing that work over the longer term, that there are no shortcuts to the diligence and analysis required in portfolio construction.
Source: Gatemore and Bloomberg, October 1, 2010 to October 19, 2015; original post October 17, 2013