sf_quant.research.run_ff_regression#

sf_quant.research.run_ff_regression(portfolio_returns: PortfolioRetSchema) → DataFrame#

Run a Fama-French 5-factor regression on a single portfolio’s returns.

Regresses portfolio excess returns (return minus risk-free rate) against the Fama-French 5 factors using OLS. Factors are shifted by -1 day so that today’s return is explained by yesterday’s factor realizations.

Parameters#

portfolio_returnsPortfolioRetSchema

Portfolio returns containing:

  • date (date): The observation date.

  • return (float): Daily portfolio return.

Returns#

pl.DataFrame

Regression summary with columns:

  • variable (str): Factor name (Intercept, mkt_rf, smb, hml, rmw, cma).

  • coefficient (float): OLS coefficient estimate.

  • tstat (float): T-statistic for the coefficient.

Notes#

  • Factors are loaded automatically from the date range in portfolio_returns.

  • Returns and factors are scaled to daily percent (×100) before regression.

  • Factor values are lagged by one day (shift(-1)) prior to joining.

Examples#

>>> import polars as pl
>>> import sf_quant.research as sfr
>>> import datetime as dt
>>> ports = pl.DataFrame({
...     'date': [dt.date(2024, 1, i) for i in range(2, 101)],
...     'return': [0.001] * 99,
... })
>>> sfr.run_ff_regression(ports)
shape: (6, 3)
┌───────────┬─────────────┬─────────┐
│ variable  ┆ coefficient ┆ tstat   │
│ ---       ┆ ---         ┆ ---     │
│ str       ┆ f64         ┆ f64     │
╞═══════════╪═════════════╪═════════╡
│ Intercept ┆ 0.0521      ┆ 2.134   │
│ mkt_rf    ┆ 0.9876      ┆ 15.432  │
│ smb       ┆ 0.1234      ┆ 2.341   │
│ hml       ┆ -0.0987     ┆ -1.876  │
│ rmw       ┆ 0.0543      ┆ 1.023   │
│ cma       ┆ -0.0321     ┆ -0.612  │
└───────────┴─────────────┴─────────┘
On this page