The 401(k) advice fight has moved to the login

The legal fight over participant-chosen 401(k) advice has been replaced by a technical one—who controls the login—as Pontera put the permissibility case on a webinar with Lisa Gomez, the former head of the Employee Benefits Security Administration. Fidelity's December cutoff of account access answered with the fact underneath it: a recordkeeper holds the account, and permission implied by a statute does not open a door the data owner keeps shut. Read together, the contest has moved off the statute and onto the plumbing.

What a webinar can settle is limited: the argument for participant-chosen advisers, made by a former enforcement chief to an audience deciding whether to build a business on it, travels with no ruling. What the format does suggest is that ERISA permissibility is no longer where the uncertainty sits—the people who spent careers enforcing the statute are comfortable making the affirmative case from a vendor's podium. The audience for that case is a plan sponsor holding a recordkeeping contract and an adviser deciding whether participant work is a product line or a favor.

The cutoff is the harder document. Participant advice runs on position-level data—what is owned, what it cost, where it sits across the current plan and a previous employer's account—and in most plans that information lives in one place while the recordkeeper decides who else may read it. Remove the access and what remains is generic guidance delivered to a specific person, a hard thing to charge for and harder to defend when the participant asks why the recommendation did not match the statement. Access is the inventory in this business; without it the adviser is selling commentary.

Permissibility was the easy half, argued in public, on the record, by people with standing to argue it; the access question is the half that decides revenue. If the December cutoff holds and is copied, participant advice gets priced by data agreements rather than by opinion letters—what the recordkeeper will feed, under whose consent, in what format, and for how long. An adviser reading this as a legal story is waiting on a ruling that would not change the economics either way.

The practical sorting is already visible: firms with a working data arrangement—sponsor blessing, executed consent, a feed the recordkeeper maintains—can take an assignment on Monday, while firms whose main asset is a view about the law can describe what they would do if allowed. December published a split that was forming anyway, and it concentrates risk in a way worth naming: a practice built on reading participant accounts through a single connection learns quickly how much of its client experience depends on a decision made in an office it has never visited.

Edelman Financial Engines' first confidence report lands on the demand side and complicates the standard account of participant apathy: most pre-retirees, according to the survey, want to run their own retirement decisions, yet 60% of them cannot say what they own. Both findings have to hold at once for the report to mean anything—the appetite for control is not the constraint, and the missing inventory is.

Plan-level education was built for a different problem: a webinar, a projection module, a nudge toward a higher deferral rate each addresses a population and each can explain how a savings rate behaves in the aggregate without ever telling a participant what the statement in front of them holds. What the Edelman finding describes is narrower and more consequential—a diagnosis, delivered once, by someone who can see the account, which is a service a participant will pay for and a plan generally will not. That is exactly why it belongs to the outside adviser, and exactly why the outside adviser needs the feed before it needs anything else.

The document itself is worth a moment's context: the firm's first confidence report is a vendor-financed reading of the psychology of the population the firm advises—self-interested, and still more specific about what participants do not know than a participation count is. Engagement gets tallied as logins, deferrals and meeting attendance, while the Edelman pairing points at comprehension as the variable that decides whether an advice offer lands, and comprehension is the one nobody has been measuring.

The procurement differs in a way that matters: education is bought by the plan, while diagnosis is bought by the participant who wants to know what they hold. A sponsor that has budgeted for the first and is now hearing demand for the second has a hole in its service menu that no education vendor closes, because the vendor's engagement metrics measure the wrong outcome. Outsourcing that work to an adviser the participant picks is the path of least resistance for the sponsor, and the path that runs straight through the recordkeeper's data policy.

The exit years sharpen the same gap: Schroders' 2026 survey puts the income target at $5,094 a month and finds 51% of retirees reporting no strategy for converting savings into income, awkward timing for the guaranteed-income shelves being built for a cohort that has not yet decided anything. Vanguard's recent work on decumulation traces the drift into required minimum distributions to plan design rather than the product shelf, and together the two directions leave participants where the defaults and the nearest available answer put them, which is rarely where the menu pointed.

The money attached is not small: defined contribution assets reached $15 trillion on the latest ICI count, an 8.7% gain for the quarter, and plan design has not moved at anything like that speed. The system's accumulation defaults keep working while its decumulation defaults barely exist, which leaves the largest single decision most participants will make—what to do with the balance—to whoever is close enough, and permitted, to answer. Rollovers have been the visible answer; the adviser inside a participant's account is the less visible one, and the one that requires access before it requires anything else.