Optimal fund structures for private debt: why fund structure for private debt is a governance decision

The closed-ended structure is the traditional private markets vehicle. Investors commit capital, the manager draws it down over a defined investment period, deploys it, and returns capital plus gains over a fixed life, typically eight to ten years for direct lending. Its great virtue is alignment. The capital base is predictable, the investment period is finite, and the carried interest mechanics line up with a defined realisation timeline. It’s the cleanest structure to govern, because there’s no ongoing redemption pressure to manage and the liquidity profile of the fund matches the illiquidity of the loans almost by definition. The cost is the J-curve and having to raise a new fund every few years, which is a commercial burden rather than a compliance one.

The open-ended structure is conceptually attractive for a strategy that deploys continuously. Capital comes in and goes out, the manager isn’t forced back to market every few years, and a strong track record compounds. The problem is the one that runs through everything in private credit: liquidity mismatch. An open-ended fund offers redemptions, but the underlying loans can’t be sold quickly without loss. If redemptions arrive faster than the loan book naturally amortises, or faster than new subscriptions replace them, the manager is pushed into distressed sales or has to gate. Plenty of funds have been caught by exactly this. It’s the central reason the FCA and other regulators are wary of open-ended vehicles holding illiquid assets, and why an open-ended private credit fund carries the heaviest governance burden of the three.

The hybrid or evergreen structure is the notable recent development. It tries to capture the deployment efficiency of open-ended capital while engineering the liquidity to fit the assets. The toolkit is now well established: staggered liquidity windows, semi-annual or annual, with long notice periods of 90 to 180 days; multi-tiered gates, capping individual redemptions at perhaps 10 to 25 percent and aggregate fund redemptions at 10 to 20 percent in any window; slow-pay mechanics that return capital in line with actual asset realisations rather than on demand; and vintage sleeves that let the manager track performance and align capital with the deals it funded. Whether any of that works comes down to the discipline behind the gates. A poorly built evergreen fails under stress like any open-ended fund.

For a long time, structure was a matter of commercial preference. AIFMD II changes that for funds that originate loans. From 16 April 2026, the directive introduces a structural presumption that loan-originating AIFs should be closed-ended, unless the manager can demonstrate to its regulator that the fund’s liquidity risk management system is compatible with an open-ended structure. That’s a meaningful shift. The default for an originating fund becomes the closed-ended structure, and openness becomes something you justify with evidence about your liquidity management, not something you simply choose.

The UK is taking its own path rather than copying AIFMD II wholesale, and the FCA has signalled a more outcomes-focused approach. But the underlying supervisory concern is shared on both sides of the Channel: an open-ended vehicle holding illiquid loans can be forced into a fire sale to meet redemptions it never had the cash to fund. Any UK manager marketing into the EU, or running parallel EU vehicles, has to plan around the AIFMD II presumption now. And any UK manager running an open-ended or evergreen credit fund should expect the FCA to ask, under FUND 3.6, whether the redemption terms are supportable by the liquidity of the book.

For retail-facing or DC-pension capital, the relevant UK structure is the Long-Term Asset Fund under COLL 15.8, designed specifically to hold illiquid assets like private credit with a redemption frequency and notice period matched to those assets. The LTAF is, in effect, the regulator’s blessed template for offering broader access to illiquid strategies without recreating the liquidity mismatch. Even where you’re not using the LTAF itself, its design constraints are a fair guide to what the regulator will accept if you’re structuring for a wider investor base.

The right structure depends on the strategy, and the fit is more than a matter of taste. It determines whether the fund can function under stress.

Direct lending, with loan tenors typically of three to five years, fits a closed-ended structure of eight to ten years comfortably, or a well-designed hybrid with conservative gates. The cash flows from amortisation and repayment give a closed-ended fund a natural liquidity rhythm. Mezzanine, with longer tenors and complex equity upside that takes time to crystallise, leans closed-ended, because forcing a redemption mechanism onto an instrument whose value depends on a future realisation invites valuation and fairness problems. Specialty finance, with rapid asset turnover, can support a hybrid structure, because the underlying assets recycle quickly enough to generate real liquidity. Distressed debt, where realisation timelines are unpredictable, needs either a closed-ended structure or a hybrid with extended lock-ups of three to five years, because offering liquidity against assets whose timing you can’t control is the one thing you never want to do.

The structure sets a liquidity promise; the strategy determines whether you can keep it. A mismatch here forces distressed sales or gating.

Every structure embeds its own conflicts, and the structure decision is partly a decision about which conflicts your governance function will spend the fund’s life managing. A closed-ended series, where you raise Fund I, then Fund II, then Fund III, creates the classic successor-fund conflict. When does the new fund start investing while the old one is still deploying? How do you allocate a deal that both funds could legitimately fund? If the older fund is winding down and holds an asset the newer fund would like, on what terms, and valued by whom, does it move? These are SYSC 10 conflicts, and they need an allocation policy that produces fair, evidenced outcomes rather than after-the-fact justification. The closed-ended structure is the cleanest on liquidity, but it generates the most allocation conflict over time, and a manager running a series has to hold a documented, pre-trade allocation methodology and an exception log that shows the rules were followed.

The evergreen structure swaps allocation conflict for valuation conflict. Because investors enter and exit at a net asset value struck between liquidity windows, the price at which they transact is set by the manager, and the manager’s fees often depend on that same NAV. An incoming investor allocated units at a stale or generous valuation is subsidised by, or subsidises, the standing investors. That’s a direct conflict, and it sharpens the requirement under FUND 3.9 for independent valuation, because in an evergreen fund the valuation isn’t only a reporting number, it’s the transaction price between investors. Closed-ended structures concentrate allocation conflict; evergreen structures concentrate valuation conflict. Build the function for whichever one the structure hands it.

Few private credit funds are a single clean vehicle. Most sit alongside co-investment arrangements, separately managed accounts for large investors, and side-by-side vehicles for the team’s own capital or for investors with specific tax or regulatory needs. Each of these is part of the real structure, and each multiplies the governance. A large investor with a separately managed account that invests alongside the main fund creates an allocation question on every deal: how much goes to the fund, how much to the account, and is the split fair across the cycle rather than just on the deals everyone wanted? Co-investment, where the manager offers favoured investors the chance to put extra capital into specific deals, creates the same conflict in a sharper form, because the deals offered for co-investment are chosen by the manager, and the choosing is where the conflict lives.

The answer isn’t to avoid these arrangements, which are commercially essential, but to build the structure with the conflicts visible and the allocation rules written down before the deals arrive. The managers I see get into trouble are the ones who treat co-investment and managed accounts as commercial side deals sitting outside the fund’s governance perimeter. Under SYSC 10 they’re squarely inside it, and the structure documentation has to reflect that the whole constellation of vehicles, not just the flagship fund, is governed as one conflicted system.

The structure doesn’t only decide liquidity and conflicts. It decides how the manager gets paid, and so what the manager is incentivised to do. A closed-ended fund typically charges a management fee on committed capital during the investment period, then on invested capital afterwards, with carried interest paid through a distribution waterfall once investors have received their capital back and a preferred return. That aligns the manager with realisations: the carry is earned when capital is returned, which is the outcome investors care about. The governance function’s job is to make sure the waterfall is calculated correctly, that the preferred return and catch-up are applied as the documents say, and that carry isn’t crystallised early on unrealised gains.

An evergreen fund usually charges its management fee on net asset value, and may take performance fees periodically against NAV growth rather than against realised distributions. That’s where the incentive risk concentrates. If the manager earns a performance fee on an unrealised NAV that the manager itself sets, the structure has built a direct link between the manager’s valuation judgement and the manager’s pay. Under FUND 3.9 the valuation has to be independent, and in an evergreen structure that independence is what separates a fair performance fee from a fee paid on a number the manager was motivated to inflate. So the structure decision is also a decision about how hard your valuation governance has to work: the more the fee depends on a manager-struck NAV, the more the function has to prove the NAV was set independently and the fee calculated correctly.

Take an evergreen direct lending fund offering semi-annual redemptions with a 90-day notice period, a 20 percent individual gate and a 15 percent aggregate gate. On paper the terms look conservative. What matters is what happens when a redemption window opens in a stressed market.

In a fund that has done the work, the manager has modelled a window where redemption requests hit the 15 percent aggregate cap while new subscriptions have stopped. The model shows where the cash comes from: scheduled amortisation, any revolving facility headroom, and, if it’s needed, the slow-pay mechanism that returns redeeming investors their capital as assets realise rather than on demand. The board has seen this analysis, has signed off that the gates and slow-pay mechanics are sufficient, and the offering documents disclose plainly that redemption is subject to these constraints. When the window opens, the fund absorbs the stress without forcing a fire sale.

In a fund that hasn’t done the work, the gates exist in the documents but the cash flow analysis behind them doesn’t. When the window opens, the manager discovers that the aggregate gate still requires more cash than the book can generate without selling loans at a discount. The gate slows the damage but doesn’t prevent it, and the investors who don’t redeem in this window are left holding a fund whose best assets have been sold to meet the ones who did.

This insight is provided for general informational purposes only and doesn’t constitute legal, investment, or regulatory advice.