Better statistics – disappointing outcomes?

You are taking a goals-based approach to managing your clients’ investments: you understand their financial goals, you have planned for the investment portfolio to fund their monetary needs, and you report your results in clear, understandable terms. That’s great, right? But you notice that while you have worked to create portfolios with better-than-average performance statistics – such as return and volatility risk – the outcomes from these portfolios show no improvement. What’s going on? Is there something else at work here that is preventing your portfolios from producing better results? Often the culprit is another source of risk that often goes unnoticed, and so it is unmanaged. That risk is known as negative skewness.

What is this thing called “Skewness?”

Traditional investing assumes that investments form a symmetrical, bell-shaped curve, with most of the returns in the middle, and the less-likely returns (large gains and losses) falling at the “tails” of this distribution. That’s a convenient abstraction, but the reality is that these return distributions are often tilted a bit – and sometimes quite a bit – to the left. This increases the likelihood of “disappointing returns” that are lower than expectations. And while it is likely that the “delightful returns” that exceed expectations will make up the difference, the problem arises when we spend money during times of lower market returns. After all, spending money only increases the deficit caused by low returns and losses. This increases the risk that we will lose portfolio value needed to support future withdrawals.

The Case of Negative Skewness in the Investment Markets

We examined over 3,000 mutual funds across all asset classes to find out whether they were symmetrical (normally distributed) or if they showed positive or negative skewness. As the graph of our results shows, the majority of funds were negatively skewed. While a skewness of zero indicates symmetry, a value of -1.0 or greater is where negative skewness becomes a significant problem. As we can see, most financial assets are significantly negatively skewed. More surprising, there does not appear to be any “return premium” attached to this negative skewness risk, Instead, we see no relationship between return and negative skewness.

Negative Skewness in our Investment Portfolios

Many investment portfolios are carefully designed to minimize volatility risk for a given expected return. This helps to provide more stable returns while minimizing short-term losses. But while focusing on minimizing volatility, many investors ignore negative skewness risk, often to the detriment of their results. After all, when an optimizer focuses on minimizing short-term volatility, it ignores other types of risk. As a result, investors often end up simply trading one type of risk for another – in this case, creating a portfolio with low volatility but a high degree of negative skewness. This produces a surprising result: any benefits we expected from lower volatility risk are wiped out by negative skewness risk.

We illustrate this in an example taken from a study of mutual funds over the five years ending April 2017. The client withdraws 5% from the portfolio in monthly installments. The portfolios have the same expected return, but different levels of volatility and negative skewness. And as we now come to realize, the portfolio that has the higher risk-adjusted return (i.e. higher Sharpe ratio) and sits above the “Efficient Frontier” has the same set of outcomes as the portfolio with higher volatility but less negative skewness. Essentially, one risk offsets the other, producing the same outcomes from the “super-efficient” and the “inefficient” portfolios.

Individual investments tend to be negatively skewed

Optimized portfolios relative to the average fund portfolio

Statistics vs. Outcomes

Keys to Managing Negative Skewness Risk

We can learn a few helpful concepts about negative skewness risk from our case study:

1. Many solutions that reduce volatility risk simply replace it with negative skewness risk
2. If you consider negative skewness in your optimization, you can lower its effect on your portfolio.
3. Like volatility, negative skewness can be lowered through diversification.
4. Unlike volatility, if you ignore skewness, its level in the portfolio may increase: the negative skewness in the portfolio can be higher than the negative skewness of its assets in isolation!