When markets fell sharply in late August, a number of traditional managers pointed their fingers at a relatively new breed of asset managers pursuing a relatively new strategy and blamed them for the spike in market volatility.
The new strategy in question is called ‘risk parity,’ which was developed in the 1990s by Ray Dalio of Bridgewater & Associates, the largest hedge fund in the world with nearly $175 billion in assets under management.
The basic idea behind risk parity is to allocate your portfolio by the amount of risk that you are taking with each asset class.
Traditionally, managers (including us) split the assets by the total dollars in the portfolio: when we say we’re 50/50 stock/bond portfolio, we mean that 50 percent of the assets are in stocks and 50 percent are in bonds.
The risk parity manager says that this is a suboptimal approach because stocks are much more volatile than bonds and, as a result, too much of the portfolio’s total risk comes from stocks. They are right that the volatility of a 50/50 stock bond portfolio is dominated by stock volatility.
Using historical data for the S&P 500 and five-year US Treasury notes back to 1926, I calculated how much risk in a 50/50 stock/bond portfolio comes from the stock portfolio and it’s 95 percent (in technical terms, this is the contribution to variance).
Assuming that the historical return, volatility and correlation estimates hold true in the future, the expected return on this portfolio is 4.3 percent (net of inflation) and an expected Sharpe ratio on a 50/50 stock bond portfolio is 0.41.
That’s right, 95 percent of the total risk of a 50/50 stock/bond portfolio can be explained by the volatility of stocks – it swamps the volatility of bonds.
The risk parity manager argues that the portfolio should be balanced in terms of risk for optimal results – 50 percent of the volatility should come from stocks and 50 percent should come from bonds.
In order to achieve this, you have to dial the capital allocation to stocks way back. I came up with capital weights of 17 percent stock and 83 percent stock to get a portfolio that is evenly balanced in terms of risk.
With such a small stock allocation, the expected return on the risk parity portfolio is 2.5 percent (again, net of inflation) – much lower than the traditional portfolio assumptions from above. The risk parity manager likes it, though, because it has a higher Sharpe ratio of 0.51 – more return per unit of risk.
Now here’s the rub. In order to get higher returns, the risk parity managers use leverage to magnify their returns. In my example, the risk parity manager doubling their return by borrowing $1 for every $1 in the fund and get higher returns than the capital-allocation manager (assuming no borrowing costs).
The volatility also increases along with that leverage, but it increases proportionally so that the Sharpe ratio is still the same. Now the risk parity has a higher expected return and a higher Sharpe ratio than the capital allocation – what’s otherwise known as Nirvana.
Critics say that the leverage makes the portfolio risky, but that’s not exactly true. The starting portfolio is really conservative (17 percent stocks and 83 percent bonds in this example) and the leverage makes it as volatile as a 50/50 stock bond portfolio, not more so.
Most risk parity managers target an overall level of volatility and use leverage to amp up the previously described conservative portfolio. Traditional managers just accept whatever volatility comes along, whether it’s high or low.
Risk parity managers who target volatility have to adjust their overall risk levels up and down constantly. They do this by buying securities when volatility is low and selling them when volatility is high. So in sharp selloffs, risk parity managers that target a certain level of volatility are big sellers and that’s what the traditional managers are carping about.
Personally, I don’t buy this story because the risk parity community is very small compared to the overall market, even when you factor in their leverage.
Volatility targeting is common among risk parity and traditional managers. Because the combined groups are much larger, I think that this may be to blame, but it shouldn’t be conflated with risk parity alone.
Let me be clear: we don’t do any risk parity in our portfolios; or volatility targeting for that matter. I happen to think that both concepts are cool strategies and have some merit, especially risk parity as an allocation in an alternative allocation in a conventional portfolio because the risk, return and correlation characteristics are compelling.
But I really don’t have a dog in the hunt, I just like watching the show to see what happens next.