Interview
Michael Kearns: Algorithmic Trading and the Role of AI in Investment at Different Time Scales
- Financial exchanges are projected to become largely electronic as algorithms and automation progressively assume tasks where machines outperform humans.
- Automated trading is predicted to first dominate optimization and execution for large institutional trades by leveraging historical and real-time data to schedule transactions across multiple exchanges.
- High-frequency trading is expected to remain a machine-dominated domain focused on identifying temporary mispricings and predicting directional movement using granular exchange data.
- Statistical arbitrage strategies are currently effective within timeframes ranging from a few seconds to a few days for forecasting asset price direction and returns.
- Machine learning models are anticipated to require significant development before they can effectively manage the risks and data sources necessary for long-term investment horizons.
- Humans managing long-term strategies that involve economic cycles, recessions, and geopolitical risks are projected to remain insulated from automation for the foreseeable future.
- The emergence of a "robo Warren Buffett" is not expected in the near future due to the complexities of understanding human nature and the political landscape.