By Peter Huber, Investment Writer & Daniel Leveau, VP Investor Solutions
Quantitative investment strategies utilize a range of mathematical techniques to identify and exploit opportunities within the financial markets. The quantitative approach can be contrasted with the traditional discretionary approach; where asset price movements are often intuited via more subjective analyses.
Quantification is revolutionizing investment, with the rapid uptake and utilization of new technologies and datasets. Between 2009 and 2017, the number of quant funds within the asset management industry more than doubled. In the five years since, their numbers have continued to grow. Writing in the Financial Times, Robin Wigglesworth declared, ‘to varying degrees we are all quants now.’ Yet, despite the industry anticipating the intensification of this trend (1), attitudes towards quantitative strategies and methods remain mixed. Indeed, the growth of quant strategies within the industry is counterbalanced by a reluctance in some quarters to embrace quantitative approaches. Commentators have observed a disproportionate scrutiny of systematic investing. In this series of three blog posts we respond to the concerns most often raised within the industry. They are:
- Historical data is an unreliable predictor of future events
Markets do not follow mathematical rules
- Quantitative investment strategies underperform their discretionary counterparts
This blog post reflects on the first of these assertions and offers a response in an effort to address the concerns investors may have about the relative strengths and weaknesses of quant strategies. In subsequent blog posts in this series, the second and third arguments are considered.
The predictive power of historical data
A defining feature of systematic trading strategies is the backtest; data analysis which seeks to determine the profitability of a set of trading rules over a specified historical period. The centrality of backtesting within the quantitative approach is a popular site for criticism. Criticism generally manifests in two forms:
- Overfitting can make any strategy look profitable
The past and future are distinct, with the former unable to provide insight into the latter
Many of the risks of overfitting arise as a consequence of data insufficiencies, with respect to both the quality and volume of historical data points. Furthermore, the breadth of factors one can introduce to explain movements in the financial markets is almost unlimited. In general it will not be difficult to find factors that perfectly explain the past. As such, factor selection is often determined by the recent performance of specific factors. Oftentimes not enough emphasis is placed on understanding the underlying rationale for why a factor should work. This may result in an overfitted strategy appearing to be profitable within a research environment but failing to perform in a live trading portfolio. However, a prudent and astute quant will seek to leaven any over reliance on historical data with a forward-looking perspective.
The constantly evolving dynamics underpinning the financial markets has led many to render historical data – and the patterns found in it – useless. Yet, it seems implausible to claim that one’s understanding of the world and the patterns one recognizes within it are not – even if involuntarily – influenced by the past. Even if we accept the premise that history doesn’t repeat itself, history is still, at the very least, ‘worthy of reference’ (2) and capable of providing valuable insight.
It is worth recognizing that discretionary investment is not without its own pitfalls, arising in the form of cognitive biases. These can be avoided with a greater sensitivity to historical patterns and the adoption of a more systematic approach to making investment decisions. In the absence of a fully pre-defined investment process, psychological research conducted by acclaimed professors Daniel Kahneman and Amos Tversky suggests that individuals are likely to, among other things, overestimate their powers of prophecy, succumb to hasty generalizations based on insufficient sampling, and idealize future outcomes. Seen from this perspective, the backtest is a way of limiting the distortionary effects of misinterpreted historical echoes and the unreliable impulses of the human mind.
(1) SigTech (2022) Hedge Fund Research Report; 95% of survey respondents believed that discretionary managers would increase their use of quantitative analytics. 83% predicted a rise in asset allocation to those strategies.
(2) Aspect Capital (2017) Advantages of Systematic Investing
This content is not, and should not be construed as financial advice or an invitation to purchase financial products. It is provided for information purposes only and is subject to the terms and conditions of our disclaimer which can be accessed here.