principal-portfolios 1.0.5

A Python implementation of the Principal Portfolios methodology by Kelly, Malamud, and Pedersen (2023), enabling optimal asset allocation by exploiting cross-predictability among asset returns.

principal_portfolios

A Python package implementing the Principal Portfolios methodology introduced by Kelly, Malamud, and Pedersen (2023), The Journal of Finance.

Overview

This package provides tools for constructing and analyzing Principal Portfolios—linear trading strategies derived from the singular value decomposition (SVD) of the prediction matrix that captures both own-asset and cross-asset predictive signals.

Key components include:

• Construction of the prediction matrix from asset returns and signals
• Decomposition into:
  • Principal Portfolios (PPs): timeable portfolios ordered by predictability
  • Principal Exposure Portfolios (PEPs): factor-exposed strategies (beta)
  • Principal Alpha Portfolios (PAPs): factor-neutral strategies (alpha)

Reference

Kelly, B., Malamud, S., & Pedersen, L. H. (2023). Principal Portfolios. The Journal of Finance, 78(1), 347–392.

Installation

After uploading to PyPI, install via:

pip install principal_portfolios