Guide to Public Pension Disclosure Data | LP Data

Why Pension Data Matters

Public pension funds disclose their PE investments due to public records laws (FOIA and state equivalents). Unlike self-reported databases, pension data is audited and includes underperformers. No survivorship bias.

LP Data Coverage:

Metric	Count
Pension funds tracked	161
Fund holdings	140,000+
Unique funds	15,500+
Fund managers	3,500+

Here's what pension disclosures contain, where they fall short, and how to use them.

What Pension Funds Disclose

Standard disclosure fields:

Fund name and manager — Which PE funds the pension committed to

Commitment amount — Total capital pledged to each fund

Vintage year — When the fund started investing

Strategy — Private equity, private credit, real estate, infrastructure, venture capital

Performance metrics — IRR, TVPI, DPI (varies by pension)

NAV — Current valuation of the LP's stake

Contributions — Capital called to date

Distributions — Cash returned

Major disclosing pensions:

Pension	Est. Holdings	Frequency
CalPERS	900+	Quarterly
CalSTRS	700+	Quarterly
Washington State	600+	Quarterly
Texas Teachers	500+	Quarterly
New York Common	450+	Annual

Quarterly disclosers have the freshest data. Annual disclosers are still useful for tracking long-dated fund performance.

Data Limitations and How We Handle Them

1. Time lag (3-6 months)

Pension disclosures reflect prior quarter-end valuations. A September release shows June 30 data. For rapidly changing portfolios, this matters.

2. Inconsistent metric definitions

Some pensions report net IRR, others gross. Some include co-investments in fund totals, others separate them. LP Data normalizes where possible and flags inconsistencies.

3. Naming variations

"Blackstone Capital Partners VII" appears as "BCP VII," "Blackstone VII," and variations. LP Data uses fuzzy matching to consolidate 15,500+ funds from raw disclosures.

4. Incomplete LP coverage

A fund with 50 LPs may have only 5 that publicly disclose. Cross-referencing multiple pensions improves confidence.

5. LP-specific economics

Fee arrangements and side letters vary by LP. Two pensions in the same fund may report different net IRRs.

What we do about it:

Fuzzy matching consolidates fund names (15,500+ funds from messy raw data)

Cross-reference multiple pensions on the same fund when available

Track time-series to catch reporting anomalies

Classify strategies using rules + LLM validation

Practical Applications

1. Performance triangulation

When multiple pensions report on the same fund, compare figures. Significant discrepancies suggest reporting differences or LP-specific economics.

2. Manager due diligence

Track a GP's performance across fund families. A manager with 5 funds in our database shows a pattern—consistently strong, improving, or declining.

3. LP commitment patterns

See which pensions repeatedly back the same managers. Long-term LP-GP relationships often indicate satisfaction with prior funds.

4. Market timing signals

Track aggregate pension PE allocations by vintage year. Large commitments in 2007 and 2021 preceded challenging environments.

5. Emerging manager identification

When CalPERS or CalSTRS commits to a first-time fund, it signals institutional validation.

Example query from LP Data:

Find all 2018 vintage buyout funds with IRR data from 3+ pensions:

47 funds meet this criteria

Median IRR: 13.8%

Q1 IRR: 19.1%

Q3 IRR: 8.3%

This multi-pension consensus is more reliable than single-source data.