AI is a powerful new tool for portfolio risk management. The reasons are multiple. Here are a few:
Risk lives in the data, not in the model. Conventional risk models like VaR utilize standard probability distributions to summarize the risk in a security’s price performance. IndicatorLab’s AI approach utilizes only select events in the history of a security to provide risk analysis. The AI approach is forward-looking in risk parameters.
Risk alerts are also different in an AI world. Risk alerts can be customized to include a number or combination of parameters, not just price. Since the AI input includes things like economic data and social indicators, movements in a security’s event space can trigger alerts, not just price.
Idiosyncratic risk on an individual security can be modeled with AI individually since the AI model can be fit for each security. This is not a “market model,” evaluating a security return relative to the market, but an individualized model with company-specific parameters.
For more reach out to IndicatorLab or eBooleant Consulting LLC.