jeff100383 / equity-risk-model Goto Github PK

This is a fundamentally-based equity risk model that performs factor risk decomposition and single stock risk attribution. The model assumes that only three risk factors—Size, Value, Momentum—are relevant and that only ten securities—AAPL, BA, CAT, DIS, EBAY, F, GOOGL, HOG, IBM, JPM—are in our universe.

Note: This model is a drastic oversimplification of equity risk and is intended purely for instructional purposes only. For the United States market, a more realistic model would typically include several more risk factors and thousands of securities.

The three factors in this model are:

1. Size, which is defined as the natural log of the market capitalization
2. Value, which is defined as the trailing 12-month earnings yield
3. Momentum, which is defined as price change from 12 months ago to 1 month ago

1. Enter the weights of the portfolio and benchmark in their respective CSV files
2. Run the equity risk model
3. Assign the portfolio, benchmark, and active portfolios as RiskModel objects, p, b, and a, respectively:

p = RiskModel(portfolio)

b = RiskModel(benchmark)

a = RiskModel(active)

4. Calculate a risk measure by executing its method. For example, to calculate the total factor risk of the active portfolio, a, execute the following line:

a.factor_risk_total()

For the portfolio, benchmark, and active portfolios, the total risk is defined as the square root of the sum of the squared factor risk and squared specific risk. Risk is further decomposed at both the stock-level and factor-level. And, for each level, we calculate the following three risk measures:

• Marginal contribution to risk
• Contribution to risk
• Percent contribution to risk

These measures are available as the following methods:

Portfolio-level

Stock-level

Factor-level

This project is licensed under the MIT License.