Thu, 27 June 2019, 18:00 – 21:00 BST
Lulu, 125b George St, Edinburgh, EH2 4JN, United Kingdom
Can equity analysts be replaced by machine learning or is a combined approach optimal?
Presented by: Alistair Haig
Time: doors opening at 6pm with the presentation starting at 6.15pm Sharpe
Cost: Free (drinks and canapés sponsored by Northfield and Axioma)
Machine learning (ML) is now being used to perform investment analysis. We evaluate the relative effectiveness of valuations made by analysts and a machine learning algorithm. Analysts in our global sample are independent of brokerage incentives and collectively make unbiased valuations using a standardised discounted cash flow model. Neither analyst not ML valuations serve as viable predictors of subsequent returns when used alone. In combination they offer credible predictive power which can be increased by adding a price momentum factor. Analyst and ML valuations could be improved by placing greater emphasis on past returns.
Alistair Haig – University of Edinburgh Business School
Alistair started his career at an investment technology firm in the 1990s then moved to buy-side and sell-side investment research. He now teaches practice-based courses in finance and investment at the University of Edinburgh Business School.
His research is driven by questions from practice with a focus on regulatory, institutional and technological change in investment analysis. His current projects examine the use of machine learning in investment analysis. Alistair’s paper can be found at: http://ssrn.com/author=2062896