About ten years ago I interviewed Cliff Asness at an Institutional Investor event – as with his investment and research philosophy, it was honest and provocative. AQR went from $12 billion to $225 billion in a decade and a half, used data science and open source research from day one and has pioneered a number of investment concepts and AI/ML in asset management, successfully.
In short, a good case study for us to examine a few key concepts around AI and data science, but let’s start from the beginning:
Twenty years ago, a team of researchers that met during their PhD work at the University of Chicago left Goldman Sachs’ quant group to set up AQR, putting the QUANT front and center.
Working mostly with institutional investors, they decided to share their research and investment process in-depth from day one, including publishing it in academic journals – no black boxes.
Risky, but it worked.
With a focus on diversification, factor investing and alternative investment (uhm, AI) strategies, AQR jumped into risk parity, managed futures, liquid alts and other concepts early on. They added academics and high-level researchers to their bench, with 84 PhDs among their 1000+ employees.
Focusing on the effective use of data science, AI and ML to help with investments, AQR:
– three months ago hired a Head of ML, along with his team and a $13 billion book of business from Guggenheim.
– opened a Tokyo office and an engineering center in Bangalore, India in 2018.
– started the QUANTA academy, a learning and development center for employees and clients (tech skills, leadership/management and personal enrichment).
– hand out the AQR Insight Award ($100k) for important research endeavors. A bit similar to Two Sigma’s big data competitions we wrote about recently.
– started the AQR Asset Management Institute in partnership with the London School of Economics (LSE, grants/prizes)
– sponsors the AQR Top Finance Graduate Award at Copenhagen Business Schoool
– offers datasets to download on their website (betting against beta, commodities for the long run, credit risk premium, et al).
– have “the curious investors” podcast (recent episodes include an active/passive discussion between Cliff Asness and John Bogle, and one on “the rise of the machines”, with John Liew on the early days of quant, and AQR’s CTO).
– Cliff’s perspectives: “Sometimes we just want to make a point, or start a conversation. These timely posts offer opinions about everything from quantitative finance to baseball stats.”
– And, of course, for the 20th anniversary: 20 for 20, a collection of AQR’s most influential research papers, 14 of which written by Cliff Asness. Very unlike Ray Dalio’s principles, yet also very similar.
For the complete case study and more details on the use of AI for investments, distribution and research, please click here.