Skip to content

Dispersion – CTA dominates over 3 years

NilssonHedge measures dispersion as the difference between the upper and lower quartiles. As a continuation of our prior work, we are now adding calculations for the cumulative performance of managers. To explore style-specific dispersion follow this link. This results in interesting observations. CTA, with the exception of a 2% median return for Risk Premia, is essentially the only strategy group that you have been able to allocate to and make money in, without selection skill.

This helps you to add additional colors when benchmarking. Not only can you claim that you are a top quartile manager at the end of the period, but you can also track your development over time. Are in the top quartile continuously, or only at specific measurement times.

The VAMI charts should be read the following way:

  • Over one year, the median manager, at the end of September has generated 7% (which is over of course a lot of non-Trend exposure). In April, the median manager (not necessarily the same manager as in September) generated approximately 5%.
  • Top quartile managers have generated more than 20% year-to-date. There are some notable exceptions that have generated much more, but those are “outliers” and managers in the top quartile.
  • The bottom quartile manager, conditional upon surviving, has generated a negative 1% return.
  • The same goes for the three-year chart, without any “selection skill” the median CTA has done close to 20% since the start of 2020. Selecting managers in the upper quartile has generated close to 40%.

The selection of style has been paramount. YTD, a median-performing CTA has outperformed managers in other styles that are not in the top quartile in that specific style.

Looking at other styles, it gets progressively more difficult to generate returns without selection skills. Event Driven and Equity LS have barely generated positive returns over the last three years.

Crypto managers are the strangest strategy. Selection mattered significantly over the last three years with results ranging from +311% to -49% returns. But this year, selection has not been possible and the quartile range is rather small.

These measures are impacted by survivorship and reporting biases. So, any new manager probably suffers from this as well. There are ways to adjust for the reporting bias, but generally not for the survivorship bias.

Leave a Reply