Interview: William Kelly, CEO, CAIA
What are you most looking forward to at the AI & Data Science in Trading conference?
The pace of change in the financial services industry will accelerate in the period ahead primarily due to the amount of alternative data that has accumulated in the last decade or so. How it is to be analysed and used in a cogent investment process remains to be seen, but the early adopters seem to be coming from the quant and systematic shops that have long been consumers of large amounts of data. The ADST conference series is a tremendous platform to raise awareness of some of the threats and opportunities, and all professionals will be best served by learning more about this developing theme likely to define the future of our industry.
What do you think are the biggest challenges facing data scientists/AI experts/quantitative investors in 2019/2020? Why are they important?
Data scientists are highly trained specialists who have been typically charged with solving some of the most complex problems in our world. As they migrate toward the asset management industry and participate in an investment process, they clearly bring a very new set of uncorrelated skills. We must embrace this but equally recognize that they are not trained analysts and do not have the same domain knowledge and communication skills to articulate a sound investment thesis. Much of the alternative data (and how it is used) is very new and the extraction and analysis of data sets could quite easily result in overfitting or other unintended outcomes. Today’s analyst must think about additive and complementary skills beyond a CAIA or CFA designation, with a heavier emphasis and understanding of (Python) programming, statistics and performance evaluation to ensure that they have the ability, skill, and courage to challenge the investment theses of their new colleagues.
Considering alt data, some news articles have suggested that US based data sets are performing badly for alpha as they are over-mined. Do you agree with this, and if so, where are better sources to be found?
A lot of the early discussion on alternative data sets seems to have been focused on the more obvious satellite images, drones, and credit card transactions. This data is widely available in the market and most algorithms can very quickly draw similar conclusions but, on a standalone basis, is likely less interesting and no longer a path toward the sustainable production of alpha. There is a tremendous amount of exhaust data or other forms of unstructured data, some of which might not be directly in the public domain. On the surface, this might be a better source of alpha, but it is imperative that the user understands the ownership of said data, related privacy issues, and any consequences that might violate regulations around insider trading or front running. There is very little regulation in place that specifically addresses a more digitized investment world. As a result, it is incumbent upon all investment professionals to self-police our industry to better out comes via greater transparency, and always putting the client first.
What is your biggest professional achievement to date?
The Chartered Alternative Investment Analyst (“CAIA”) Association is a professional body focused exclusively on alternative investment education, with an emphasis on a greater demonstrated degree of professionalism that will yield better outcomes for the investment community whom we serve. The hallmark for this achievement is the CAIA Charter which has become the global standard for alternative investment education and is held by 11,000 investment professionals in over 95 countries. The advent of the growing use of alternative data, the greater use of computer programming and coding, and a wider deployment of machine learning and artificial intelligence were largely seen first by the alternative managers using systematic and highly-quantitative investment processes. These changes resulted in the hiring of a very different set of more tech-enabled professionals and specifically, data scientists. CAIA saw an increasing need to fill the widening skills gap between the trained analyst and the data scientist who essentially work together on (better) outcomes for the same end investor. As a consequence, we formed the Financial Data Professional (“FDP”) Institute and the related FDP credential as a new learning pathway for the analyst who now needs to be better equipped with coding skills, familiarity with data mining and machine learning, and other concepts such as the use of predictive and other time series models.