Author: James LoBuono, Business Development Head, Cognitir
Today’s prominent technologies like machine learning (ML) and artificial intelligence (AI) are having a profound impact on investment evaluation and decision-making frameworks in the context of the major asset classes.
It is imperative for traditional finance professionals to learn the foundations of how to find, source, and interface effectively with data scientists and technical folks.
Paradigm shifting forces like higher computing power, cloud, low-cost data storage, and robust business intelligence tools, all within the context of challenging excess return environments, have sparked significant interest in data science from investment teams across major asset classes. Not only is it relevant for investment managers to be aware of these powerful trends, but it is also vital for them to be able to source and work with technical experts when data science is considered for investment decisions.
As the effectiveness of traditional approaches like classic value investing blurs, especially when applied to heavily followed assets (e.g. big-cap equities), pioneering investment managers are turning towards non-traditional or “alternative” sources of data to drive tactical buy and sell decisions. Moreover, data science is not just being deployed in obvious segments such as liquid equity markets through algorithmic based trading. Complex ML tools are also being utilised by private equity players to improve operational decision making in real-time at the portfolio level.
Stepping back, machine learning trains computers in recognising patterns just like humans can, but exponentially faster, and with datasets that are orders of magnitude larger. The aforementioned fusion of powerful computing, combined with low-cost access to massive cloud-based data stores, renders machine learning more feasible today at scale. These factors also make ML resoundingly effective across all industries, including investment analysis.
According to a recent Deloitte report labeled Artificial Intelligence: the next frontier for investment management firms, one of the most preeminent applications of machine learning in investment management is, and will be, the utilisation of data-driven technologies to generate excess alpha. Select investment teams are drawing upon pools of “alternative data” sources to garner an edge. These data stores include mining predictive patterns in satellite imagery, shopping mall foot traffic, and sentiment towards a company on social platforms like Twitter.
This is all being done at rates that are thousands of times faster than a single capable human.
For context, think briefly about the edge in pure actionable insight that investment managers could gain relative to competition when the latter cohort is still spending time mired in more traditional financial statement and valuation analysis. In fairness, many of the related datasets are costly to say the least and have mainly been utilised by the biggest of funds, including the likes of Point72 and Two Sigma. However, in time, more players will incorporate these alternative sources into their investment processes at higher scale and lower cost. This trend should be similar to the development and proliferation of other general-purpose technologies.
AI and ML are not just limited to short run trading strategies. Large private equity shops including Two Six Capital also use complex business intelligence platforms for operational and cost management purposes at the portfolio company level. A July of 2019 Wharton School post on the blog Knowledge@Wharton discusses Two Six Capital co-founder Sajjad Jaffer’s remarks at a March of 2019 conference about how analysis is moving from old school spreadsheets to big data platforms. Jaffer illuminates how ML transitions company analysis from longer intervals down to near real-time, minute to minute operational updates. According to Two Six Capital, it uses over three dozen machine learning models.
Where do these tech trends leave investment managers? In a position that will make it more necessary than ever to triangulate towards the right mix of technical talent that not only can deploy and run machine learning models, but that can combine these capabilities synergistically within traditional human ingenuity, creativity, and experience. Inside countless firms and organisations, there is a dichotomy between technical personnel and business decision makers. Technical teams must understand business lingo, and conversely, business people must also understand technical lingo. Ways to facilitate the incorporation of this ideal include setting time aside to make innovation a formal part of decision processes, hosting team events focused on technical creativity, inviting speakers from the tech space, and by even investing in basic business intelligence technologies like Sisense and IBM Cognos.
Some places to begin connecting with and discovering technical talent are on sites like AngelList and Crunchbase. Which qualifications should one look for when searching for technical talent? Someone with a great data-driven mindset will also need to have a deep understanding of data science concepts and statistics, experience working with unstructured/complex data, the propensity to communicate findings clearly, and most importantly, the ability to draw conclusions from complex models and make impactful recommendations in the context of capital management. Leaders within the finance world ultimately need to recognise the opportunities and threats that data analytics present, while simultaneously closing the gap between their current platform relative to the capture of strategic insight.