The DOL's safe harbor hands participants the private-credit mark

Private credit's next marginal buyer cannot negotiate a spread, demand a covenant, or sell on a Friday afternoon, because that buyer is the 401(k) participant and the Labor Department's proposed safe harbor for alternatives in defined-contribution plans is the door being held open for them.

The proposal is being argued as an access question, and on access the instinct is sound: a participant with three decades of contributions ahead of them has a better claim on a liquidity premium than a retiree drawing income off a bond ladder. The process safe harbor, though, answers a different question — when the mark moves, whose problem is it — with documentation: follow the steps, keep the minutes, file the diligence, and the committee's duty is discharged by the record rather than by the result.

There is nothing underhanded in that construction and no bad faith required of anyone who wrote it: prudence is a standard applied in hindsight, a miserable thing to be judged by, while procedure is judged on the file, a far better thing to be judged by. The consequence is that a rule about investment quality becomes a rule about record-keeping quality, and the committee's incentive shifts from asking whether the asset is priced right to asking whether the paperwork is complete.

The demand side explains why a more forgiving standard is worth so much to the sellers, since an illiquid loan book wants liabilities that behave like equity—capital that arrives on schedule, leaves rarely, and cannot be pulled in a bad quarter. Retirement plans are funded by payroll deduction, their participants are told to think in decades, and their flows are steadier than anything a wealth manager generates from client referrals; private credit has been looking for that liability for years, and this proposal is the mechanism that would supply it. That is why the same rule can be defended as access and read, from the borrower's side, as a new funding line.

Inside committee memos, the number carrying the argument is private credit's headline default rate, which PWD's tracking puts at 0.8%, a real number asked to hold more weight than it can. A default rate tells you what has been declared; a mark tells you what has been priced, and in a privately held book the distance between the two is set by negotiation rather than by a tape. The gap in its plainest form is that marks on the smallest borrowers have repriced twelvefold since 2023 while the headline rate sits still, and the lenders holding the most of that tail are already repositioning around the workout.

The 0.8% measures the loans that have reached a formal conclusion, and formal conclusions come last in a private credit cycle. What precedes them is a stretch in which a troubled loan still has alternatives to a default, each of which keeps the loan out of the default column while changing what it is worth. Some of that work is good credit management and protects borrower and lender both, but the narrower point for a plan committee is that a default rate cannot do the work of a valuation, and a memo that treats it as one has swapped a flow statistic for a price.

A multiple flatters a small base, so the same ratio off a thin discount is a different event from the ratio off a wide one, and the move in the tail, by itself, falls short of a distress call. What it establishes is that the tail is moving ahead of the headline, and in a credit cycle the tail moves first: the smallest borrowers have the fewest ways to avoid being priced and the least standing to argue about the number.

That tail would be an internal valuation argument among professionals if the paper lived only in closed-end pools with long memories and no redemption rights, but it does not. Private-market wrappers are being assembled for individual balance sheets at a pace the filings themselves show: three Form Ds from Brown Advisory inside twenty days reported $57.8 million sold for client-facing venture vehicles whose offering amounts the forms never state, and Baceline's two filings reported $394 million sold in fourteen days on the same blank line. Neither figure is frightening on its own; the blank offering amount is the tell, because an open shelf is built to keep selling.

The lenders most exposed to the tail have the clearest reason to want a wider buyer base, and the work they are doing suggests they know it, since firms repositioning around the workout are organizing ahead of a default cycle rather than waiting for one to announce itself. A retirement plan entering the asset class now would be buying the part of the market that has already repriced and the part that has not, at a price the seller administers.

Redemption caps are the least discussed feature of private credit's distribution into wealth management and the most consequential, because a vehicle that limits how much capital can leave in a quarter cannot easily be forced to print a clearing price: the manager's mark stands because nobody is required to trade against it. That is arithmetic, with no scandal required, and the result is that a great deal of private credit has been sold to individuals without the market establishing the price at which they would sell it.

Plans run on a different clock from the funds that would sit inside them, and a participant can move money between menu options while the underlying loans cannot be sold, so two regimes would describe the same book and only one of them involves anyone transacting. The proposal arrives before the first clearing prices of this cycle have been established, which is the sequence worth objecting to, more than access and more than fees.

One day's flow makes the point from the other side: $914 million moved into the iShares 0-3 Month Treasury Bond ETF, and whatever source that money came from, duration was the last thing it wanted. It is a fair picture of what the marginal dollar asks for when nobody is selling it a liquidity premium.

A defined-benefit plan can carry illiquidity because an actuary's schedule sits on the other side of it, but a defined-contribution plan cannot make that trade on the plan's behalf: the money moves when the participant moves, and the vehicle has to absorb contributions and withdrawals at the same time. That is a harder liability structure for an illiquid book to sit inside, and it is why the valuation question matters more in a 401(k) than it does in a pension.

The mismatch would surface at the plan level, where a participant who retires, changes jobs, or rolls over needs cash on the plan's timeline rather than the fund's and a redemption cap does not distinguish among those events. The plan wears the timing risk first, and if the gates bind through a bad stretch, the residual cost lands on the participants who stayed.

There is a template for the rest of it, and the industry has run it before: an institutional asset gets a wrapper sized for individual accounts, the wrapper gets a distribution force, and the distribution force gets a rationale for why the end client should own what the institutions were already holding. Interval funds were the second step for private credit; the safe harbor is the third, with one difference—a retirement plan arrives with payroll deduction attached and a participant who may never make an active decision about the allocation.

None of this means participants should be barred from private markets: the case for access is real, and a plan menu restricted to daily-liquid options is not obviously the right menu for a 35-year horizon. Horizon matters, though, only after valuation has been settled, because time protects an investor from a bad quarter only if the price they bought at was real, and the price inside these vehicles is administered rather than discovered. Open the retirement channel before that gap closes and the first honest clearing price in private credit may be set by a recordkeeper explaining a statement to a plan sponsor.

A mark that has to be tested against real bids would not stay inside the vehicle that printed it; it would become the reference for every other mark in the asset class, and for the allocation decisions of every advisor who recommended private credit on the strength of its reported numbers. The safe harbor does not address that exposure, which is why an access argument cannot settle a valuation question.

The objection worth making is narrower and harder than the one the debate has produced, and it starts from a simple gap: a process safe harbor standardizes the committee's homework; it does not standardize the number the committee is buying at. If the department means to open retirement capital to private credit—and the access argument for doing so is real—the safe harbor worth writing is the one that pairs the paperwork with a valuation requirement: an independent mark on a stated schedule, the methodology disclosed to the plan, and the committee obliged to read it. That version would still expand access; it would also stop the expansion from doubling as a transfer of mark-to-market risk from the firms that made the loans to the households that end up holding them.

The proposal as described converts fiduciary judgment into a paperwork standard, and the effect is a liability transfer whether or not that is the intent. None of it requires a villain: the manager earns a fee on committed capital, the committee earns a defense, and the participant earns the mark.

Watch two things from here: whether a final safe harbor pairs its paperwork with a valuation requirement, and the first private-credit mark inside a retirement plan that has to be restated. The restatement will arrive with the minutes in order and the diligence filed.