JPMC Evergreen Private Funds Capstone

This repository contains a JPMC capstone project on evergreen private funds, reported private market returns, synthetic index construction, and portfolio allocation under an illiquidity budget.

The project studies whether evergreen private funds improve traditional portfolio construction and where private allocations are most useful across private equity, private credit, and private real estate. The analysis combines public SEC filings, public market anchors, macro factor data, synthetic return modeling, and portfolio optimization.

This was prepared as an academic capstone project. JPMC is referenced only to identify the capstone context. The public repository excludes internal meeting notes, sponsor prompt files, and sponsor provided index data.

Project Question

Evergreen private funds report smoother returns than public assets, but that smoothness can come from both real economic differences and stale or appraisal based marks. The core question is how an allocator should use private assets when both public and private return series have measurement problems.

The project asks:

• How can the U.S. evergreen private fund universe be defined from public filings?
• Can monthly private fund returns be recovered from SEC N-PORT, 10-K, and 10-Q filings?
• How do reported private fund indices compare with public market anchors?
• Can factor based synthetic returns extend the history and adjust for stale marks?
• Where should a limited illiquidity budget be allocated in a traditional portfolio?

What The Project Does

• Builds a taxonomy of U.S. evergreen private funds from SEC filings
• Scrapes and cleans monthly fund return data from N-PORT filings
• Reconstructs additional return histories from NAV and distribution data
• Constructs equal weighted, semi annually rebalanced private fund indices
• Compares private equity, private credit, and private real estate sleeves against public anchors
• Uses factor analysis and regime dependent shrinkage to build synthetic private asset returns
• Tests portfolio allocations for 50/30/20 and 60/40 investor frameworks
• Produces final plots, index outputs, notebooks, and a methodology report

Main Findings

• Private assets can improve portfolio efficiency, but the improvement is selective.
• The largest gains often come from improving the public sleeve before adding large private allocations.
• Reported private returns should not be treated as pure mark to market risk because appraisal smoothing and stale marks matter.
• In the modeled allocation framework, private real estate is the strongest first use of illiquidity, followed by private credit.
• Private equity receives the smallest private allocation in the preferred framework.
• A practical final allocation is closer to moderate private exposure than maximum private exposure, roughly 70 percent public and 30 percent private.

Repository Layout

1. Part 1 - Evergreen Fund Taxonomy/
  Fund universe construction from SEC data

2. Part 2 - Performance Data Collection, Analysis & Cleaning/
  Return scraping, manual return reconstruction, cleaning, and diagnostics

3. Part 3 to 5 - Index Construction, Analysis & Comparison/
  Private fund index construction and public market comparisons

4. Part 6 - Portfolio Construction & Analysis/
  Factor analysis, synthetic returns, portfolio allocation, and final plots

5. Final PPT & Methodology Report/
  Final methodology report in markdown form

METHODOLOGY.md
  Root level copy of the final methodology report

Selected Outputs

Reported private index levels:

Composite private index versus public anchors:

Synthetic private credit versus public leveraged loans:

Portfolio allocation examples:

Technical Stack

• Python
• pandas
• NumPy
• matplotlib
• statsmodels
• scikit-learn
• scipy
• yfinance
• openpyxl
• Jupyter

Running The Notebooks

Install dependencies:

python3 -m pip install -r requirements.txt

The notebooks are stored with outputs so the analysis can be reviewed without rerunning every data collection step. Some internal and sponsor provided files are intentionally excluded from the public repository. See DATA.md for details.

Notes

This repository is for research and portfolio demonstration. It is not investment advice and should not be interpreted as a recommendation to buy or sell any security, fund, or strategy.