The Risk Prism: Master the Architecture of Modern Factor Models

Overview

Master the architecture of modern factor models. Transition from asset-class silos to a surgical, multidimensional understanding of risk drivers. From the Factor Zoo to Generative AI Factors, explore the systematic decomposition of market risk.

1. The Asset Class Illusion

Traditional finance allocates capital by asset class (e.g., 60% equities, 40% bonds). However, asset classes are just wrappers. A corporate bond, for example, is not a pure "bond." It is a combination of interest rate risk, credit risk, and equity risk.

True diversification requires looking through the wrapper to the underlying factors.

2. The Factor Zoo

Since Fama and French proposed their 3-factor model in 1992, academics have discovered hundreds of new "factors" (the so-called "Factor Zoo").

• Macro Factors: Growth, Inflation, Real Rates, Liquidity.
• Style Factors: Value, Momentum, Quality, Size, Low Volatility.
• Statistical Factors: Principal Components extracted purely from the covariance matrix without economic intuition.

3. Generative AI Factors

The bleeding edge of factor modeling involves using Large Language Models to extract unstructured data factors.

• Supply Chain Contagion: Using LLMs to map global supply chains and create a "supply chain risk" factor.
• Regulatory Sentiment: Quantifying the tone of regulatory filings to build a "regulatory burden" factor.
• CEO Overconfidence: Analyzing earnings calls to construct a behavioral factor.