As we enter the Global, Digital, Responsible age, the alignment between the most technology enabled strategy, being quantitative (quant) trading and responsible investing is being examined.
The adoption of environmental, social and governance (ESG) principles into quantitative models has come from either a desire to meet investor needs for responsible investing, or as an alternative data set to be used to generate alpha. The use of these ESG principles and other alternative data sets in a quantitative strategy has some interesting challenges for clients, but also significant opportunities which we will examine further.
Time has been spent understanding alternative data challenges faced and there has been substantial consideration on the next generation of opportunities in the alternative investments space, including quantamental strategies.
Traditional quantitative strategies have relied on market related data sets (exchange rates, listed equity prices, bond yields). This should be defined as merely data or market data in the investment universe.
The use of other non-market related data has long been a documented source of alpha for top performing hedge funds (think about the known use of satellite images of retail parking lots to predict retailers’ quarterly results) and now the pace of data generation is accelerating.
Studies predict that there will be 26 times more connected ‘things’ in the Internet of Things than people this year.
With this expected growth in data, it will be a significant growth, either in terms of creating new investment strategies or for data service providers to mine data for sale.
Unlike traditional market data, alternative data is not frequently available and criticism has been levied that large funds have been able to purchase data sets at an unfair advantage.
It will be interesting to see where the use of alternative data goes in the future given latest privacy trends and calls to better regulate Big Tech.
Alternative data in quantitative strategies
As relates to alternative data sets, the main issues faced include:
- Lack of uniform standards or KPIs
- Frequency of reporting
- Short history of data sets
These issues are even more challenging for those wishing to implement a quantitative strategy as they are unable to successfully back-test their investment thesis given the short history of these alternative data sets currently available.
Accordingly, in quantitative strategies, alternative data sets (including ESG factors) are being used as a negative screen with few funds making buy or sell decisions based solely on such data.
As data improves, quant strategy investment thesis will be proven and the use of alternative data enabled quant strategies will become more mainstream. It is also likely that the global proliferation of common ESG reporting standards will lead to better alternative data sets.
Until such time though, there will be a need for human intervention in any strategies using alternative data. This, among other factors, has given rise to the quantamentalist.
Progress towards Artificial Intelligence (AI) led trading has been slower than initially anticipated - as widely reported deep neural network (DNN) shortcomings persist with current AI systems struggling to cope with unfamiliar territory.
This fact, along with the difficulty in implementing quant strategies using alternative data mean there is still a need and in fact a big opportunity for humans to interpret alternative data to generate alpha.
Quantamentalism is the use of quant strategies and/or machine learning to analyze vast amounts of data and present a machine determined buy or sell decision for approval by a human.
For the time being, quantamentalism represents the current best-case interaction between humans and machines. Improvements in alternative data are to be expected and will be driven by acceptance of common ESG reporting standards and a significant increase in more true machine lead quant strategies to follow.
by Justin Thomas, Director