Nicolas Mirjolet - Multivariate Trend Following (S7E8)
Corey Hoffstein chats with Nicolas Mirjolet about scaling statistical arbitrage strategies and the advantages of larger players. They explore Quantica's managed futures program, comparing multivariate and univariate trend-following methods, and the role of correlation in signal generation. The episode covers market applicability, historical outperformance, and innovations in trend following. Nicolas discusses adding markets, estimating correlation matrices, and achieving risk diversification. They also analyze recent trend-following environments and offer insights on investment diversification and the trade-offs between diversification and convexity in trend programs.
Key Points
- Nicolas Mirjolet emphasizes that while Quantica Capital employs a multivariate approach to trend following, the firm focuses on maintaining a homogeneous liquidity profile by trading only the most liquid markets, ensuring consistent and predictable returns.
- The effective number of risk factors driving trend-following performance can significantly vary year to year, with fewer risk factors often correlating with higher performance, as observed in the contrasting years of 2022 and 2023.
- Investment universe diversification is crucial for trend-following strategies, especially in adverse market environments where macro factors are less dominant, helping to capture idiosyncratic trends that contribute to more stable returns.
In this episode, I speak with Nicolas Mirjolet, CEO and Co-Head of Research at Quantica Capital.
We begin with Nicolas’s experience operating a statistical arbitrage fund, where he provides his thoughts as to what makes a strategy easier or harder to scale a business on. Nicolas also provides some context for his somewhat counter-intuitive view that the larger players had a bigger edge in this capital constrained space.
We then transition to Quantica’s flagship managed futures program. Nicolas explains that while Quantica is a price-based trend follower, they apply a multivariate approach to their signal analysis. We discuss how the approach works and how it contrasts against a standard univariate approach. Specifically, Nicolas shares his thoughts on how the multivariate approach impacts the portfolio return profile and why you may want more or fewer variables in your signal universe than your tradable market universe.
We end the conversation with Quantica’s most recent quarterly research paper, which provides quantitative insight into the convexity versus robustness tradeoff trend managers make when they add more markets to their portfolio.
Please enjoy my conversation with Nicolas Mirjolet.
Chapters
| 0:00 | |
| 3:45 | |
| 10:40 | |
| 19:24 | |
| 24:45 | |
| 30:01 | |
| 34:22 | |
| 40:14 | |
| 47:46 | |
| 53:35 | |
| 1:00:37 | |
| 1:03:06 |
Transcript
Loading transcript...