Co-Investment Looks Free. Here's What It Actually Costs.

GPs pitch co-investment as a fee-free enhancement to LP returns. The math is seductive: no management fee, no carried interest, direct exposure to a handpicked deal. What's not to like? Plenty, as it turns out — and the full cost only becomes visible after years of tracking realized vs paper returns.

Co-investment opportunities have become a standard feature of the institutional private equity offering. Large LPs expect them. Mid-sized LPs demand them. And the market has responded: co-investment capital deployed annually has roughly doubled over the past five years, with dedicated co-investment funds and separate-account vehicles proliferating.

The headline economics look extraordinary. A direct co-investment alongside a PE sponsor, on the same deal, at the same entry price — but without the 2% management fee and 20% carry that applies to capital deployed through the main fund. On a deal returning 2.5x gross, the fee differential translates to roughly 500-700 bps of incremental net IRR to the LP.

So why doesn't everyone deploy all their PE capital as co-invest?

Cost #1: Adverse selection

The deals that GPs offer to LPs as co-investment are not a random sample of their portfolio. They're the deals where the GP cannot fully deploy its fund commitment — because the check size required exceeds the fund's concentration limits, or because the GP wants to reduce its own risk, or because the deal characteristics don't quite fit the fund's thesis.

This sounds benign. In aggregate, it isn't. Academic studies of co-investment performance consistently show that co-invest deals underperform the GP's main-fund deals on a gross basis by 100-300 bps of IRR. The fee savings, properly measured, offset only part of this gross underperformance. Net-of-fee, co-investment portfolios often match or slightly underperform the main fund.

This is adverse selection in its textbook form. The GP offers you deals it has a reason to share. Most of the time, that reason is neutral. Some of the time, it isn't.

Cost #2: Information asymmetry

When you invest in a GP's main fund, you're invested alongside an institutional due diligence team, a portfolio construction discipline, and a reserve for follow-on investment. When you co-invest in a single deal, you're sharing the same underlying asset — but you're on your own for everything else.

Three specific asymmetries hurt LPs:

Diligence timelines. The GP has been studying the target for months. You get the data room, a call with the deal team, and typically 2-4 weeks to commit. The information processing deficit is real and compounds decision quality over time.

Reserve capacity. Main-fund capital reserves for follow-on investment in the same asset: bridge financing, recapitalization, growth capital. Co-invest capital typically doesn't. When the portfolio company needs 15% of its original equity as additional capital, the main-fund LPs participate pro-rata; co-invest LPs are diluted or have to find fresh capital on short notice.

Governance rights. Main-fund LPs access the fund's LPAC. Co-invest LPs often access nothing. If the deal goes sideways, your ability to influence outcomes — valuation methodology, restructuring decisions, timing of exits — is limited to what's in the co-invest documentation, which is usually thin.

Cost #3: Portfolio construction drag

Professional co-investment requires a dedicated team, processes, and systems. A $500m pension fund building a thoughtful co-invest program needs 2-3 senior investment professionals dedicated to deal evaluation, monitoring, and reporting. This overhead — easily $1-2m annually — is a real cost that offsets the headline fee savings.

Smaller investors who co-invest without this infrastructure face a worse outcome: they say yes to deals they don't have time to evaluate, because the alternative is saying no and losing access to future co-invest allocations from the GP. This is a recipe for a portfolio of deals selected by pressure, not conviction.

Cost #4: Concentration and vintage risk

A main-fund allocation diversifies across ~15-25 portfolio companies and deploys over 3-5 years. A co-invest portfolio is typically more concentrated (fewer, larger tickets) and more vintage-exposed (deals clustered in whatever 12-18 month window the LP was active). The result is higher expected return variance and lower-quality diversification.

Studies consistently show that co-invest portfolios exhibit 30-50% higher return dispersion than main-fund portfolios of comparable size. Some co-invest programs generate exceptional returns; others disappoint. The median outcome is typically close to main-fund net returns with materially higher variance.

Cost #5: Tax and operational complexity

Each co-invest deal is typically structured as a separate SPV with its own legal documentation, tax structuring, reporting, and K-1 equivalent (or local equivalents). For an LP executing 5-10 co-invests per year, this is real operational overhead: tax filings in multiple jurisdictions, separate audit obligations, bespoke reporting.

Institutional LPs absorb this cost. Smaller LPs often discover it only at tax season.

The total

Stack the costs properly:

• Gross return drag from adverse selection: ~100-300 bps of IRR
• Execution drag from information asymmetry: ~50-100 bps, variable
• Program overhead: ~20-50 bps on small programs, 5-15 bps on large ones
• Vintage/concentration risk premium: variable, manifests as higher dispersion, not direct fee cost
• Operational and tax overhead: $50-200k per year for a small program

Against headline fee savings of ~500-700 bps vs main-fund deployment.

The net math, for a well-run institutional co-invest program, is still positive. But the margin is narrower than the brochures suggest — perhaps 100-250 bps of net IRR uplift, not the 500-700 bps the fee savings alone imply.

Who should be co-investing

Should: Institutional investors with dedicated PE teams, meaningful ticket sizes ($100m+ per deal), long-established GP relationships, and the infrastructure to execute professionally.

Should probably not: LPs without dedicated staff, inconsistent access to deal flow, or pressure to say yes to maintain relationships. Small-ticket co-invest programs routinely underperform — the fixed costs eat the fee savings, and deal selection suffers.

In between: Fund-of-funds and multi-LP co-invest vehicles. These structures mutualize the infrastructure cost and can deliver the theoretical benefits of co-investment at scale. But they reintroduce a fee layer — typically ~50-75 bps — that offsets some of the headline savings.

The takeaway

Co-investment is a powerful tool when executed properly and a drag on returns when executed casually. The "free" label on fee-free allocations is one of the more misleading pieces of marketing in private markets. The right question for any LP considering a co-invest program isn't "what's the fee savings?". It's "do we have the infrastructure, deal flow, and discipline to capture the savings without eating them through worse deal selection?"

For most allocators under $2-3bn of PE exposure, the honest answer is no.