- Below-average market volatility is typically associated with above-average returns. Given a choice, therefore, most investors would prefer low volatility to high.
- For active managers, however, the choice is less obvious: lower market volatility is associated with lower correlation and lower dispersion, both of which make active management harder to justify.
- Active portfolios are typically more volatile than their benchmarks; how much more volatile depends in part on correlations. Active managers pay an implicit cost of concentration, which rises when correlations decline.
- Low dispersion makes it harder for active managers to add value, and reduces the incremental return of those who do.
- These perspectives highlight the conflict between the goals of absolute and relative return generation.
A SIMPLE QUESTION
Should an active manager prefer to operate in a low volatility environment or a high volatility environment? What factors should influence this decision?
At first glance, the choice seems fairly easy. Exhibit 1 reminds us that volatility and returns are inversely related. Rising volatility typically penalizes results and vice versa.
We can see this more directly in Exhibit 2. Here, we separated the months in our database by intra-month volatility and examined return data in each set of months.
These exhibits make the manager’s choice look obvious: if volatility is high, returns tend to be negative; if volatility is low, average returns are substantially positive. Positive returns mean that the manager’s clients are making money, which they usually appreciate, and that the manager’s fees (if asset-based) are also rising. Attracting new assets is easier in a rising market, whereas “investors do not reward outperformance in down markets with higher subsequent flows.”
Lower volatility means that managers and clients alike enjoy a smoother return path with fewer surprises. The manager should obviously wish for low volatility, both for its own sake and because of its connection to higher returns. What could go wrong?