Diligence the Fund Before You Sign Your Next PE Gig

Many PE CEOs and C-suite leaders misread how their investors might manage and support their business, so diligence your future fund as much as they've diligenced you.

Two CEOs. Both good operators. Both failed.

The first ran a premium consumer business. The plan was international expansion. His backer was a small mid-cap fund. It could not capitalise the expansion properly. When the market turned, it could not reinvest either. The company stalled. He left.

The second ran a larger business. His fund had one priority: exit. They harvested. He executed. The sale cleared. The fund did fine. He walked away with little upside and a gap on his CV.

Neither failure was about the business. Both were about the fund.

CEOs diligence the company. They study the market, the team, the numbers. Then they sign. Most skip the harder question: what does the fund actually need from this company, right now?

The fund is the real employer. The company is the instrument.

The diagnostic

Before signing, a CEO should answer nine questions.

• Fund vintage. What year is the fund in its life? A year-two fund can back ambition. A year-six fund in a ten-year life needs exits. The same “growth plan” means different things in each.

• Dry powder and reserves. Is follow-on capital actually reserved for this company? How much, on what terms? No reserves mean no massive expansion if opportunities arise but capital is needed. It also means no rescue in a downturn.

• Fundraising cycle. When did the GP last close a fund? When is the next one due? A GP raising the next vehicle needs distributions to LPs. That changes every conversation you will have with them.

• What’s their IRR, MOIC and DPI. LP’s have lots of GP partners to choose from, how does your potential fund compare to it’s peers ?

• Fund size fit. A small company in a big fund gets orphaned. A big bet in a small fund gets starved the moment it wobbles. Check the cheque size against the rest of the portfolio.

• Strategy type. Growth, buyout, value, turnaround, harvest. Each wants a different CEO outcome. A harvest mandate is legitimate. It also caps your upside. Know which one you are in.

• Equity package. Vesting. Leaver provisions. Ratchets. Anti-dilution in down rounds. Institutional strip versus sweet equity. A good PE lawyer pays for himself ten times over.

• Deal partner. How many boards do they sit on? Distracted partners kill companies. Talk to three to five former portfolio CEOs. Include the ones who left unhappy. The ones who will not take your call tell you the most.

• Capital structure. Leverage constrains every strategic choice. Over-levered deals cannot pivot. Read the covenants.

• Exit scenario. What does a good outcome look like to them? IPO, trade sale, secondary, continuation vehicle. Each has different economics for you.

The question most CEOs miss

The sharpest question is about DPI.

IRR rewards speed. DPI rewards cash actually returned to investors. Early in a fund, GPs chase IRR. They want mark-ups, growth, a narrative. Late in a fund, or when raising the next one, LPs stop caring about paper marks. They want distributions. DPI becomes the only number that matters.

That shift is where CEOs get crushed.

The plan you were hired to execute was an IRR plan. Invest. Grow. Build value. The plan you are now being asked to execute is a DPI plan. Sell what you can, when you can, at a price that clears.

Same company. Same CEO. Opposite incentives.

The fund did not lie. The fund’s priorities changed under you.

The real question is not what kind of fund it is. It is what the fund needs from this specific company right now. A growth story or a liquidity story. Those are different jobs. The business card says the same thing either way.

Signs the liquidity story is coming:

• The rest of the portfolio has underperformed and needs a winner to exit.

• The GP is talking about a continuation vehicle.

• The next fundraise is imminent and DPI is thin.

• Marks have been held flat for two reporting periods.

The test

Before you sign, write down the fund’s likely exit path and timing. Write down what that means for your equity. Compare it to what you have been told in the interview.

If the two do not match, you have a problem. Ask again. If they still do not match, walk away.

The hardest part of this job is walking away from the wrong deal. CEOs fall in love with companies. They should fall in love with funds.

How to actually do this research

Most CEOs assume this data is locked behind a PitchBook or Preqin subscription. Some of it is. But more is free than you would think. Public pension funds that invest as LPs are legally required to disclose fund-level performance. That data is free, quarterly, and detailed.

What works well

For US-domiciled mega and large buyout funds with mature vintages, four sources do most of the work:

• WSIB (Washington State Investment Board) — sib.wa.gov/reports.html. The single richest free source. Net IRR, TVPI and DPI for every fund WSIB has invested in, calculated by Hamilton Lane. Start here.

• CalPERS — calpers.ca.gov. Updated quarterly. Hundreds of funds. Cash in, cash out, remaining value, net IRR.

• SFERS — mysfers.org. San Francisco pension. Good for mid-cap and Asia-focused funds.

• Maryland SRPS — sra.maryland.gov. Broad coverage, quarterly.

For market context, read Bain’s annual PE report for the relevant region. It tells you whether the fund is outperforming or underperforming its peers in the same vintage.

Where it breaks down — honestly

I tested the prompts against ten funds spanning size, vintage, geography and strategy. Performance data came back for four. Six returned nothing on IRR or DPI. Here is what the test exposed.

• Geography bias. The four sources above are US public pensions. They hold mostly US funds. European and Asian funds frequently miss them. Inflexion (£1.25B UK buyout, 2018 vintage) returned no performance data from any of the four. Its LPs are mostly European and Asian institutions that disclose nothing publicly.

• Vintage bias. CalPERS itself flags any fund with a 2021 or later vintage as “Not Meaningful” — too early in the J-curve. EQT IX (2021 vintage) returned a 9.6% IRR but with that caveat attached. Below year-three, treat any number as noise.

• Date staleness. WSIB published TA XIII at a 56.85% IRR in December 2021. That number was the freshest free figure I could find. PAG Asia II went from 19.95% IRR in WSIB’s December 2021 report to 10.25% in the December 2024 report — same fund, same source, three years later, halved. Always carry the date next to the number. If the data is more than eighteen months old, treat it as historical, not current.

• Size cliff. Hit rate falls off a cliff somewhere between $1B and $3B. Sub-$1B funds rarely appear in the four sources. The cliff is even higher for non-US funds — use the smaller funds prompt below for any non-US fund under roughly $3B.

• Gross vs net. For some Asian and African funds, the only public numbers leak through trade press (AVCJ, DealStreetAsia, Mergermarket) and are usually gross IRR, not net. Net IRR is typically five to seven points lower. Do not compare the two.

Sources the standard list misses

• UK LGPS pools — Border to Coast, LGPS Central, Brunel Pension Partnership and the others. Hold significant European PE. FOI-able.

• Nordic AP funds, APG, PGGM — Continental European pensions with much better disclosure than most US peers for European funds.

• AVCJ, DealStreetAsia, ION Analytics — Trade press that frequently leaks Asian fund performance, usually as gross multiples.

• DFI project pages — For African and emerging-market funds, FMO, EIB, IFC, Proparco, DEG, BPI, Swedfund, Finnfund, SIFEM and EBRD all publish project descriptions confirming commitments and basic terms. They will not give you IRR. They will confirm the fund exists, who else is in it, and at what scale.

• SEC Form D and Form ADV — Even small non-US funds with US LPs file Form D (one-time) and Form ADV (annual, if registered adviser). For August Equity V (£300M UK fund), Form D confirmed the US LP raise of $170.5M. Companies House confirmed the UK structure. No IRR, but enough to validate the fund is real and well-supported.

Two prompts

I have written two structured research prompts. Paste either into Claude or ChatGPT, fill in the fund name, and get a first draft of your diligence pack in twenty minutes.

• Generic prompt — for US-domiciled buyout, growth, credit, real estate or infrastructure funds at $1B+ with mature vintages. Will return useful data most of the time.

• Smaller funds prompt — for sub-$1B US funds, sub-$3B non-US funds, first- or second-time funds, regional managers, or anything you suspect won’t appear in major US public pension databases. Leads with regulatory filings and DFI sources. Ends with an explicit data confidence rating, including “None” as a legitimate outcome.

Both prompts are below. A note on what to expect: in my testing, the generic prompt returned performance data for mots mega and large US funds. The smaller funds prompt returned regulatory and structural data but almost no IRR for any of the four small or first-time funds tested. That is not a flaw in the prompt. It is the disclosure regime. The absence of public performance data for a small fund is itself useful information — it tells you the fund’s LPs are private institutions, sovereigns or DFIs that do not publish, and that you will need to ask the GP directly and triangulate from portfolio company filings.

Use the prompts before you sign. Treat what they return as a starting point, not a complete picture. Then go ask the GP the questions the data raised.

———

Generic Private Equity Fund Research Prompt

For US-domiciled funds at $1B+, mature vintages. Replace the bracketed fields with your target fund.

I want to research [FUND NAME] managed by [GP NAME], a [STRATEGY] fund with a [GEOGRAPHIC FOCUS] focus, vintage year approximately [YEAR]. Build a comprehensive picture using only publicly available sources. For every data point, cite source name, date and URL. Always carry the data date next to every IRR, TVPI and DPI figure. If data is more than eighteen months old, mark it STALE. Do not present stale data as current.

• Fund basics. Size, vintage, fund number, domicile, strategy, status.

• GP overview. AUM, strategies, ownership, senior personnel changes in last three years.

• Performance — US sources. Net IRR, TVPI, DPI from WSIB (sib.wa.gov), CalPERS (calpers.ca.gov), SFERS (mysfers.org), Maryland SRPS (sra.maryland.gov). Note: any vintage 2022 or later may be flagged “Not Meaningful” — carry that flag forward.

• Performance — non-US sources. For European funds also check Border to Coast Pensions Partnership, LGPS Central, Brunel Pension Partnership (UK LGPS pools — FOI-able), AP funds (Sweden), APG and PGGM (Netherlands). For Asian funds check AVCJ, DealStreetAsia, ION Analytics — these often carry gross figures only, do not present as net.

• Series trajectory. Pull the same metrics for Fund I, II, III, IV. Build a trend table. Note whether improving, stable or declining. Always include the report date for each data point.

• Confirmed LPs. Commitment size and source for each.

• Peer comparison. Two to three funds of comparable vintage, geography and strategy. Same metrics, same sources.

• Fundraising environment. Bain regional PE report for the relevant year.

• Secondary market activity. Any LP tender or continuation vehicle. Pricing and structure.

• Key personnel changes. Departures, timing relative to fundraising.

• Portfolio companies. Notable exits, write-downs, underperformers.

For every data point: cite source, distinguish confirmed from derived, note paywalled sources and free alternatives. If the fund is non-US and under $3B, switch to the smaller funds prompt — the four US sources will likely return nothing.

———

Smaller Funds Research Prompt

For sub-$1B US funds, sub-$3B non-US funds, first- or second-time funds, regional managers, or any GP unlikely to appear in major US public pension databases.

I want to research [FUND NAME] managed by [GP NAME], a [STRATEGY] fund with a [GEOGRAPHIC FOCUS] focus, vintage approximately [YEAR], estimated size [FUND SIZE]. Public performance data on smaller funds is thin and often absent entirely. Treat this as triangulation, not lookup. For every data point cite source, date, URL, and flag whether confirmed, derived or inferred. Where data is absent, say so explicitly. Do not improvise IRR or DPI from press leaks unless the source names a specific gross or net figure with a date.

• Regulatory filings first. Start here, not with performance. SEC Form D (sec.gov/edgar) for any fund with US LPs — confirms fund existence, size, US LP raise. SEC Form ADV (iapd.sec.gov) for any US-registered adviser with $150M+ AUM. UK Companies House (find-and-update.company-information.service.gov.uk) for the management company and fund vehicles. Luxembourg RCS, Irish CRO, Guernsey and Jersey registries for offshore vehicles. ESMA and FCA registers for regulatory status.

• Fund basics. Size, vintage, fund number, domicile, vehicle structure. Sources: GP website, press releases at final close, PitchBook and Preqin free previews, PEI Media, Buyouts, regional trade press (AVCJ for Asia, LAVCA for Latin America, AVCA for Africa, Invest Europe for Europe).

• Performance — try in this order. Tier 1 (US public pensions, broad): WSIB, CalPERS, SFERS, Maryland SRPS, Oregon Treasury, Texas TRS and ERS, Virginia Retirement System. Tier 2 (US public pensions, smaller and regional): Pennsylvania PSERS, New Mexico ERB, Montana BOI, South Carolina RSIC, New Jersey, Connecticut, Kentucky, Iowa PERS. Tier 3 (US city and county): NYCERS, LACERS, Fairfax County, Orange County, Ventura, Chicago Teachers, Boston Retirement. Tier 4 (US endowments, low yield): UTIMCO has good disclosure; UC Regents occasionally. Yale, Harvard, MIT, Michigan rarely disclose individual funds — do not waste time. Tier 5 (international LPs, often the best route for non-US funds): UK LGPS pools (Border to Coast, LGPS Central, Brunel — FOI-able), Nordic AP funds, APG, PGGM, Australian super funds, Ontario Teachers, CPP Investments, BCI, OMERS, PSP. Tier 6 (DFIs, the only route for many emerging market funds): FMO, EIB, IFC, Proparco, DEG, BPI, Swedfund, Finnfund, SIFEM, EBRD, BIO, OeEB, ADB. Project pages confirm commitment size and basic terms. They will not give IRR.

• Portfolio company triangulation. If LP sources return nothing, identify 3–5 portfolio companies. Pull financials from Companies House (UK), EDGAR (US), state registries. Find acquisition price, debt at entry, dividend recaps, follow-on capital, exit price. Build rough gross multiple per asset. This is time-intensive — budget an hour per company.

• Series trajectory. Apply the above to Fund I, II, III. For first-time funds, look at predecessor track record at partners’ previous firms.

• Confirmed LPs. Public pension minutes, FOI requests (US public pensions and UK LGPS — typically 20 working days), DFI project pages, press releases at fund close. Distinguish confirmed from inferred.

• Peer comparison. 2–3 funds of comparable size, vintage, geography, strategy. Peers may also be data-poor — that is itself the finding.

• Fundraising and personnel signals. Time on the road (over 24 months signals trouble). Fund size vs. target. Step-up vs. predecessor (a small step-up signals weak performance even with no IRR public). Senior departures in the 3 years before or during fundraising. Placement agent changes mid-raise.

• Secondary market and continuation vehicles. GP-led CVs signal exit pressure. LP tenders signal liquidity problems. Sources: Secondaries Investor, PEI, Lazard and Jefferies annual secondary reports.

• Time-box and stop rule. If after 30 minutes of work no IRR, TVPI or DPI has been found from any source, stop searching for performance. Report fund basics, LP base structure, regulatory filings, and date the absence of performance data explicitly. Do not pad. Do not improvise.

Data confidence rating — mandatory closing line.End the output with one of: HIGH (multiple LP sources within 12 months agree on IRR/TVPI/DPI), MEDIUM (one LP source, or portfolio triangulation only, or data older than 18 months), LOW (regulatory filings and press only, no performance numbers), NONE (fund exists but no public data beyond size and vintage). A LOW or NONE rating is itself a finding — small GPs with sovereign, DFI or family-office LPs systematically disclose nothing publicly. Absence is a signal about the LP base, not a verdict on the