Two Risks That Ruin Long-Run Investing

The first risk of investing is the Drawdown Risk - the loss from the peak. The second risk of investing is the Low Return Risk - the under-performance vs. expectations over a stretched period of time.

First, a few words about drawdown. Quants measure risk in many ways like Volatility, Skew, Variance, Beta, Tracking Error etc - but, in my experience, clients care the most about the drawdown. Not only is it conveniently measured in the same units as returns, it also captures the cognitive threshold of pain that clients are willing to bare before calling it quits. Saying that realized volatility was 15% vs expected 8% is very different than a loss from peak of 45% vs 24%. The worst witnessed drawdown for the U.S. Large Cap Equities was -84% on June 1932, when the 5-year standard deviation was 32%. Down 84% is clear; standard deviation of 32% is not. In 2018, S&P was down -4%, but its monthly drawdown was -13%. The calendar year return looks harmless, but by the year end, investors were close to panicking, because of the drawdown number. So while drawdown doesn’t fit nicely into the mean-variance math and typical benchmarking, it is the biggest risk that ruins long-run investing.

Second, it’s the Low Return Risk. Investors rarely put low returns into the ‘risk’ category, and often juxtapose risks vs returns. I believe that low returns are a risk because an extended period of under-performance causes investors to abandon their previously selected strategies (asset allocation, sub-asset classes, and individual managers) and switch out at a very bad time. This locks in accumulated under-performance and switches to a typically over-valued alternative solution, creating permanent damage to the portfolios - and, hence, I define it as a risk.

Quants have a term for it - it’s called “Model Risk”. This is when your model stops generating a positive return.

During my first year as a quant on Wall Street, in order to focus my creativity on the Model Risk, my boss, Magali, used to say something that stuck with me: “if we loose 1 basis points a day relative to our benchmark, we will have a perfect tracking error of 0, no factor biases, no sector and stock-level risks, perfect portfolio construction and yet we will be out of business in less than a year.”

Many long-time quants still remember vividly August of 2007, the month of the famous quant meltdown. During the second week of August, many popular quantitative factors had outsized losses and quant strategies experienced massive drawdowns. Because of the drawdown risk, investors naturally panicked and some started to withdraw their assets. However, the returns stabilized quickly and not much long-term harm was done by that drawdown alone. It was only once the second Risk showed up, the traditional quant industry faced the lasting outflows - it was the Low Return risk. The graph below helps visualize what happened. During my talks, I have given it many names: “The Inverted Hockey Stick”; “Quant’s Most Feared Graph”; “Model Risk Manifested”, “It’s the First Moment that matters most” and “The Low Return Risk”.

Kenneth French Data Library

Kenneth French Data Library

In the total portfolio context, we see the Low Return risk show up recently in Over-Diversified portfolios. During the bull market that started at the bottom in 2009, many robust diversification solutions were engineered to solve the 2008 drawdown risk, but many of them have under-estimated the low return risk. This resulted in investors asking for ‘more risk’ near the end of the bull market, selling some of the low-return diversifies in favor of equity risk. Phrases like ‘stretching for yield’, ‘long-only is better than long-short’, ‘70/30 meets our expected returns better than our current 60/40 allocation’, ‘hedge funds don’t work’ - have all become common client preferences around 2017. Low Return risk shows up by causing investors to change their allocations without a previously specified plan, in a reactive and value destructive way.

Over the years, my definition of risk has evolved from a very narrow one (tracking error) to a broader one that is more representative of client’s experience: I now define Risk as anything that causes investors to take actions that destroy value of their portfolios. Larger than expected drawdowns and lower than expected returns destroy value. First, because they drag down performance. And second, because they make investors abandon their ‘long-run’ strategies, miss out on the potential bounce back, and jump on new things that outperformed recently but are about mean-revert. All of this is not new and well documented and it creates a vicious cycle of actual returns significantly lagging the average returns of the approaches that investors are trying to implement.

When I sit back and think of all the examples when someone (including myself) was tempted to change an approach, it was most always a result of either drawdown or the prolonged low return.