Creating a CTA from Scratch- II

What is driving CTA performance?

November 2022. Reading Time: 10 Minutes. Author: Nicolas Rabener.

SUMMARY

  • CTAs pursue trends across asset classes, regardless if long or short
  • 2022 is exceptional as CTAs have more short positions than during the GFC
  • No single long bond index position remains, fixed income is one big short

INTRODUCTION

In our last research article (read Creating a CTA from Scratch – I) we demonstrated that building a CTA strategy is not particularly difficult. All it takes is a broad set of asset classes, a trend following methodology, and a portfolio construction model. We chose 59 indices that can be traded via futures contracts, the 12-month rolling return as a trading signal, and a simple equal-weighting with leverage to achieve a target volatility.

The do-it-yourself (DIY) CTA Strategy had a correlation of 0.6 to the SG CTA Index, which is one of the benchmark indices for the managed futures industry. The entire backtest was completed in Excel and took less than 3 hours, so we could likely achieve a higher correlation with more time and effort, or a different approach like creating a replication index via factor exposure analysis.

However, the DIY CTA strategy is good enough to investigate what has been driving CTA performance. Managed futures products have been labeled as black-box strategies as their portfolios change constantly, which tends to confuse investors. 

In this research article, we will explore the performance drivers of CTAs. 

SG CTA INDEX VS DIY CTA STRATEGY

When we created the DIY CTA Strategy, we ensured that the assumptions were reasonable from an implementation perspective, but ignored costs. In terms of making the backtest more realistic, we assume a management fee of 2% and a performance fee of 20% above a high watermark, as well as 0.10% transaction costs per trade. Rebalancing occurs on a monthly basis.

Contrasting the performance of the SG CTA Index and DIY CTA Strategy highlights a similar return in the period from 2000 to 2022. In some years the performance was almost identical, while somewhat different in others. The differences are likely explained by the universe of tradable assets and portfolio construction.