privateassets 0.1.0

Quantitative analytics for private-asset returns: multi-factor money-weighted PME (MATF), unsmoothing, and risk-adjusted performance estimation.

privateassets

Quantitative analytics for private-asset returns — private equity, private credit, and other private-market strategies.

Core method: MATF — a multi-factor, money-weighted PME that estimates risk-adjusted alpha and systematic factor exposures (β) directly from private-asset cash flows, generalising Direct Alpha / KS-PME / GPME from a single benchmark to a multi-factor deflator. Additional analytics (unsmoothing, volatility/beta estimation) currently live within matf and may be promoted to top-level submodules as they generalise.

Built on qis and factorlasso; it does not depend on optimalportfolios (sibling, not child).

Install

pip install privateassets        # once published
# or, from source:
pip install -e .

import privateassets
from privateassets.matf import ...   # estimator API

Layout

privateassets/                 # MATF estimator engine (deflator, PME, panel-MLE
│   ├── __init__.py            #   unsmoothing, rolling covar, shrinkage, pipeline)
│   └── matf/
│       └── illustrations/     # 14 figure scripts (python -m privateassets.matf <id>)
scripts/                       # reproduction runners
paper_code/                    # publication tracks: jfqa_matf_pme, faj_application (unwritten)
data/                          # git-ignored, private: licensed inputs (see DATA_README.md)
outputs/                       # git-ignored: generated artifacts
projects/                      # git-ignored, private: mandate engagements (not part of the package)

Papers

paper_code/ holds the clean, public publication tracks. The JFQA paper (estimator + theory, MSCI cohort data) and the FAJ paper (fund-level application, Preqin data) are not yet written. The precursor Oaktree manuscript runs on proprietary LP data and is kept in the private projects/ tree, out of this public package.

Data

No data is stored in this repository. All vendor / LP data is obtained under licence and placed in data/ (git-ignored). See DATA_README.md for the per-source licensing rules before adding or committing anything.