As France debates the future of its pension system once again, and in particular how much room to give to funded retirement savings, I wanted to take a close look at what is happening on the other side of the Atlantic. The United States is the country where funded retirement saving has grown most spectacularly: US retirement assets reached $51.2 trillion as of June 30, 2026, compared with $2.561 trillion in reserves at the end of 2025 for Social Security, the public pay-as-you-go pension system. It is also where the question of opening that savings pool to private equity is most hotly contested today.
I expected a debate among financiers. I found a debate among lawyers. The starting point is a paradox: an American employee has practically no access to private equity in their workplace retirement plan, the 401(k), while public pension funds, which pay public employees a pension defined in advance, allocated an average of 13.9% of their assets to it in fiscal 2025, up from 3.6% in 2001. To understand that gap, you have to wade into a legal debate of great complexity, which nonetheless comes down to a simple question: who answers for the choice if private equity disappoints? In the United States, it is the employer, or the provider to whom it has delegated the plan, that selects the investments offered to employees, and it is the employer that employees can sue, with its own assets on the line, if they consider that choice imprudent. Yet the law does not say what is prudent: that is left to the Department of Labor, and then to the courts. In the absence of a clear rule, and under the threat of class actions, most employers stay away. It is this balance that a president is now trying to shift by executive order: he intends to open to the more than 90 million Americans who save through these plans the assets already enjoyed by public pension funds, and to put an end to what he presents as lawsuits brought by “opportunistic trial lawyers.” All of this under the gaze of a public opinion that some, it seems, have tried to manufacture.
This is the journey I trace here, taking the time to explain to a European reader what a 401(k) is and why a 1974 law still weighs on the savings of tens of millions of Americans. France has chosen the regulatory route: since a ministerial order of July 1, 2024, the managed option of French retirement savings plans (PER), the one that applies when savers do not choose their investments themselves, must include a minimum share of unlisted assets, mandatory since June 30, 2026 for company plans; the obligation applies only to that managed option. At European level, the ELTIF 2.0 regulation has relaxed the framework for unlisted funds open to retail investors. The French debate is not limited, however, to comparing the returns of private markets with those of other asset classes. The report of the French National Assembly’s committee of inquiry “into the predation of French productive capacity by speculative funds,” tabled in June 2026, blames the funds for the debt they load onto companies, the jobs cut, the dividends and fees extracted, the tax advantages they enjoy and the loss of sovereignty they allegedly cause. I contested its conclusions at length, point by point, in my analysis No. 27, “Predation of productive capacity by speculative funds”: the report put to the test of facts (in French), published on this site on July 4, 2026. The United States, for its part, is trying to remove a litigation barrier that Europe, lacking a comparable class-action culture, does not face: there, the obstacles lie in distribution and liquidity, not in the courtroom. Two paths for the same question, one that will be at the heart of any pension reform that makes room for funded savings: should employees’ savings finance unlisted companies, and on what terms?
On March 30, 2026, the US Department of Labor (DOL) published a proposed rule designed to make it easier for private market assets to enter 401(k) plans. Six months later, the file is making headlines for an entirely different reason: nearly 12,000 comments supporting the proposal appear to have been fabricated.
1. The 401(k) and ERISA: what are they, for readers who are not US lawyers?
A paragraph of the tax code that became the backbone of American retirement. The 401(k) takes its name from a simple reference: subsection (k) of section 401 of the Internal Revenue Code, added by Congress in 1978 and effective from 1980. The arrangement took off only with the regulations proposed by the IRS on November 10, 1981, which allowed employees to direct part of their salary into it: that is the birth date of the modern 401(k).
The principle is that of an individual retirement savings account opened through the employer. Each employee decides, individually, to put part of their salary into it; the employer may add its own contribution; depending on the option chosen, tax is paid either when the money goes in or when it comes out. Participation remains voluntary, but it is increasingly automatic: the SECURE 2.0 Act, signed into law on December 29, 2022, has required since January 1, 2025 that plans created after its enactment automatically enroll eligible employees at a contribution rate of at least 3% of salary, rising by one percentage point a year, with each employee free to opt out or change the rate; businesses with ten or fewer employees and those less than three years old are exempt. In 2026, an employee can contribute up to $24,500 a year, plus $8,000 more from age 50. In the United States, the 401(k) has become the main vehicle for workplace retirement saving.
Defined contribution versus defined benefit. Two families of plans coexist. In a defined benefit plan, the employer promises a pension, usually calculated on salary and length of service, and bears the financial risk. In a defined contribution plan such as the 401(k), only the savings effort is set: the retirement income depends on what has been paid in and what the investments have earned, and it is the employee who bears the risk. The shift from one to the other has been massive. In March 2025, 70% of private-sector workers had access to a defined contribution plan, against 14% to a defined benefit plan, and 72% of workers had access to at least one plan, some to both; 28% had access to none. The reason is simple: no federal law requires an employer to offer a retirement plan. Access therefore depends on the company and the job: 55% of workers at establishments with fewer than 50 employees had access, against 90% at those with more than 500; 47% of part-time workers, against 81% of full-time workers; 49% in the lowest-paid quartile, against 91% in the highest-paid. As of June 30, 2026, 401(k) plans held $10.8 trillion, against $3.2 trillion for private-sector defined benefit plans. Public defined benefit plans, covering federal, state and local government employees, hold $10.4 trillion. They differ from 401(k)s in three respects. Savings are managed collectively, by professional investment teams, rather than split among individual accounts. Risk is borne by the public employer, and ultimately by the taxpayer, rather than by the employee. And they fall outside ERISA, which does not apply to government plans, and are governed by the law of each state. They therefore have neither to offer daily liquidity to each employee nor to fear the class actions specific to that law.
The employee chooses, but usually by default. The employer selects a list of investments, the “menu,” among which the employee allocates their savings. Since the Pension Protection Act of 2006, an employer can enroll its employees automatically and, absent instructions from them, place their savings in a default option protected by a safe harbor, most often a target-date fund. A target-date fund is not an individual fund: it is a pooled fund, identified by a retirement year and shared by all employees aiming for that date. The plan offers a series, one per vintage, and each employee is placed in the one matching their expected retirement year: everyone planning to retire around 2055 holds units of the same “2055” fund. The employee chooses neither the series nor its manager. It is the employer, or the committee that administers the plan, that, acting as fiduciary, compares the series offered by asset managers, selects one and designates it as the default option, and must then monitor its performance and fees regularly. The chosen manager allocates the savings among stocks, bonds and other investments, then automatically makes the portfolio more conservative as the target date approaches, reducing equities in favor of bonds and cash along what Americans call a “glide path.” The employee has no decisions to make: the fund makes them, from the first contribution to retirement. It is through these funds, more than through employees’ individual choices, that private markets could enter 401(k)s on a large scale. As of June 30, 2026, hybrid mutual funds, a category that includes target-date funds, accounted for $1.7 trillion in 401(k) plans on their own.
ERISA, a law that protects and exposes. The 401(k) is governed by a 1974 federal law, the Employee Retirement Income Security Act (ERISA). It was born of a trauma: in 1963, the closing of the Studebaker plant in South Bend, Indiana, deprived more than 8,500 workers of all or part of their promised pensions. President Ford signed the law on September 2, 1974. ERISA does not require any company to offer a retirement plan; it regulates those that exist: disclosure to participants, rules on participation and funding, duties of those who manage the assets, and the right to sue. It preempts all state laws relating to employee benefit plans.
At the heart of the text lies the concept of the fiduciary: the employer, or the committee administering the plan, must act “with the care, skill, prudence, and diligence” of a prudent man acting in a like capacity. ERISA sets out no list of permitted or prohibited investments. Everything is judged against this standard of prudence, and it is the courts that ultimately say what it requires.
Why so many lawsuits. ERISA gives participants the right to sue a fiduciary who breaches their duties, and the fiduciary is “personally liable” to make good the plan’s losses. In 2015, the Supreme Court made clear that the obligation does not end with the initial selection: a fiduciary has “a continuing duty to monitor trust investments and remove imprudent ones.” Combined with the American class-action mechanism, this explains employers’ caution. A low-cost index fund offers little purchase for such a suit; a fund that includes private equity, more expensive and harder to value, offers more. It is this barrier that the reform seeks to loosen.
2. Why an agency, and why a public comment process
Since ERISA does not say which investments are prudent, the question is one of interpretation, and it is settled at the federal level: neither New York nor California can legislate on the point. Congress gave the Secretary of Labor the power to issue the regulations “necessary or appropriate” to carry out the law.
The DOL exercises that power through an agency dedicated to workplace retirement and welfare plans, the Employee Benefits Security Administration (EBSA), which signed the proposed rule. It does not act alone, however. The executive order of August 7, 2025 also enlists the SEC, tasked with considering ways to facilitate retail investors’ access to alternative assets, and the Treasury.
A procedure must still be followed. A federal agency cannot adopt a binding rule on its own say-so. The Administrative Procedure Act requires it to publish its proposal in the Federal Register and then give any interested person the opportunity to submit “written data, views, or arguments.” The agency must then consider these submissions and set out, in the final rule, the basis and purpose of its choices. Without that step, the rule could be struck down by a court.
The DOL’s proposal is no exception. It implements the executive order and explicitly aims to reduce the litigation risk that would otherwise deter fiduciaries from exercising their judgment. The Department invites the public to comment on the relevance of the six factors it proposes (performance, fees, liquidity, valuation, benchmarking, complexity), but also on questions it leaves open: should a factor relating to participants’ profile be added, should the construction of the investment menu as a whole be addressed, should good practice for monitoring after selection be spelled out? The comment process thus serves to test the technical soundness of the text, and then to build the record that judges will examine if it is challenged.
It is in this setting that the fraud surfaced.
3. A rigged comment process?
The public consultation did not ask a closed question of the “for or against private markets in 401(k)s” type. It concerned the proposed rule itself: the safe harbor, its six factors and the questions the DOL had left open. Anyone could comment on all or part of the text; it was the massive use of form letters that turned the exercise into a head count of supporters and opponents. The comment period closed on June 1, 2026. The DOL received more than 47,000 comments, about 43,000 of them form letters: roughly 31,000 against the proposal and 12,000 in favor.
In August, Bloomberg took a closer look at those 12,000 supporters. They carry neither signature nor location, repeat five text templates verbatim, and arrived at the Department in nearly identical volumes every day for a week before stopping abruptly. Reporters tried to reach three dozen of the signatories: five vehemently denied any involvement, for themselves or for family members, and at least two comments appear to have been filed in the names of people who had died. One of these “signatories” had died five months before their support was recorded. Bloomberg found no sign of fabrication in the main opposition campaign.
The fallout is political and could become criminal. Democratic Representatives Bobby Scott and Jamie Raskin, together with Senator Bernie Sanders, referred the matter to the FBI and the Attorney General; Elizabeth Warren and Bernie Sanders demanded answers from the DOL by October 8, 2026.
4. A comment process is not a referendum
It is still worth gauging what the fraud really changes. The comment process is not a vote: the DOL must respond to arguments, not count letters, and a letter reproduced thirty thousand times carries no more weight than a single one.
On that front, the record has not moved. The Department received more than 200 submissions of at least 500 words, some running to over 100 pages: 121 support the proposal, 101 oppose it. On one side, asset managers, recordkeepers, employers and academics; on the other, labor unions, advocacy groups and state attorneys general.
The fraud therefore does not weaken the future rule legally. It weakens its legitimacy, at the very moment the DOL must finalize it, and it hands opponents a ready-made argument to challenge the final rule in court or to have it repealed.
Two routes would lead to repeal. The first runs through the executive branch: in the United States, “the administration” means the president and the executive departments. A future president could have the rule repealed by their own Department of Labor, after a new notice-and-comment procedure, without a vote in Congress. The second runs through Congress, which can overturn a rule by a joint resolution of both chambers; a rule overturned this way cannot be reissued in substantially the same form without a new law. But such a resolution must, like a bill, be signed by the president, who can veto it. As long as the current president is in office, that is until January 2029, this second route remains theoretical, whatever the outcome of the November 3, 2026 midterm elections, which will renew the entire House of Representatives and 35 of the 100 Senate seats.
5. The substance: two classes of savers
The real debate lies elsewhere. An employee covered by a defined benefit plan already benefits, without knowing it, from private equity when the plan’s manager has included it in the allocation, as more and more pension funds do; the employee whose retirement depends on a 401(k) has practically no access to it. The orders of magnitude are well known. Among the 73 largest state public pension plans tracked by The Pew Charitable Trusts, the share of alternative investments rose from about 20% of assets in 2009 to 35% in 2022, and 43 of these plans allocated more than 10% of their assets to private equity alone. The California Public Employees’ Retirement System (CalPERS), which in March 2024 set its long-term private equity target allocation at 17%, had $123.6 billion invested in it as of June 30, 2026, or 19.4% of its assets, as I noted in the Week 38 press review. So if private equity remains outside 401(k)s, it is not because a law forbids it: none does. What blocks it is fear of litigation. The DOL’s proposal itself notes the sharp rise in class actions over the past decade.
Recent history illustrates the swings of the pendulum. On June 3, 2020, responding to a law firm acting for two private equity managers, Pantheon and Partners Group, the DOL stated in an information letter that a fiduciary does not breach their duties merely by offering, among the plan’s options, a professionally managed diversified fund (target-date, target-risk or balanced fund) that includes a private equity sleeve. The door opened only cautiously: private equity can never be offered to employees on its own, and the fiduciary must cap its share, ensure the fund remains liquid enough to meet withdrawals, require independent valuation and take fees into account. On December 21, 2021, the new administration half-closed it. The DOL clarified that its 2020 letter should not be read as an endorsement. The criterion it applied was not size, which it did not quantify, but experience: only an employer already evaluating private equity for its defined benefit plan, preferably with a qualified adviser, would be equipped to offer it in its 401(k). That is a minority. For all the others, whom the DOL described as fiduciaries of “small” plans, the conclusion was unequivocal: they were likely not in a position to evaluate a private equity investment. That caution was rescinded on August 12, 2025, five days after President Trump signed, on August 7, 2025, the executive order on 401(k) investors’ access to alternative assets.
The March 2026 proposal creates no privilege for private markets. It sets out a process: a fiduciary who objectively, thoroughly and analytically considers six factors (performance, fees, liquidity, valuation, benchmarking, complexity) will have their judgment presumed reasonable and receive significant deference. The text has its limits, however: it covers the selection of each option, not the construction of the menu as a whole.
6. The real lock is at the Supreme Court
The real decision will not come from the DOL. On October 6, 2026, the Supreme Court hears Anderson v. Intel. The plaintiffs accuse Intel’s fiduciaries of investing part of the plan’s target-date funds in hedge funds and private equity, resulting in higher fees and lower returns than comparable funds.
The question is simple: is it enough to allege underperformance or higher fees to sue a fiduciary? The DOL argues that it is not: plaintiffs would have to show underperformance against a “meaningful benchmark,” that is, an investment with comparable objectives and strategy.
If the Court follows that line, the safe harbor and the case law will reinforce each other, and litigation risk will recede markedly. If it rejects it, comparing the fees of a fund that includes private markets with those of an index fund will be enough to launch a lawsuit, and no DOL rule will change much. Practitioners interviewed by PitchBook (a Morningstar company) are under no illusion: a DLA Piper lawyer considers that, from a litigation standpoint, the proposal does not solve the problem, while an Alight Solutions consultant sums up large employers’ caution by noting that the bigger the plan, the bigger the target on its back.
7. What this means for private equity
A deep pool, not a tidal wave. As of June 30, 2026, US defined contribution plans hold $15 trillion, of which $10.8 trillion is in 401(k)s alone. Deloitte estimates that private markets could exceed $1 trillion there by 2030, or about 6% of assets, after about 2% in 2027; PwC arrives at the same order of magnitude with a 5% allocation. Scope matters: these figures cover all private markets, including private credit, real estate and infrastructure. In the target-date fund modeled by Georgetown University, private equity represents, early in a saver’s career, only a third of the private markets sleeve, on a par with real assets and private credit.
Relative to fundraising, the effect remains measured. Private markets raised $1.3 trillion worldwide in 2025, close to the 2024 total. Spread over four to five years, $1 trillion in assets would represent $200 billion to $250 billion a year, or 15% to 19% of global private markets fundraising in 2025. That is significant, but gradual, and concentrated toward the end of the period.
Above all, these flows will not be evenly distributed. Paid in by American employees and invested by American target-date fund managers, they will go mostly to North American managers and assets. North America already dominates: it raised $432 billion of the roughly $616 billion raised in 2025 by closed-end private equity funds worldwide, or 70%. These $616 billion are not comparable with the $1.3 trillion cited above, which covers all private markets, including private credit, real estate and infrastructure, and comes from a different source. The vehicle matters as much as the amount. A 401(k) operates continuously: employees contribute every payday, switch options, retire or leave the company, and each of these movements requires an up-to-date value and available cash. A traditional closed-end fund, which calls capital as it invests and locks it up for about ten years, is poorly suited to this. The DOL’s 2020 letter already required fiduciaries to ensure the fund remains liquid enough to meet withdrawals and transfers, and the 2026 proposal makes liquidity one of its six factors. Hence the appeal of semi-liquid private equity vehicles, open continuously and redeemable periodically within limits: they raised $204 billion in the United States in 2025, more than double the 2023 figure. By way of illustration, if private equity received a third of the flows, as in the Georgetown model, it would capture $65 billion to $85 billion a year, or between a third and two-fifths of what US semi-liquid private equity vehicles raised in 2025. The additional capital will therefore benefit American private equity first and widen its scale advantage over Europe.
Patient capital just as liquidity runs short. In 2025, distributions to investors remained stuck at 14% of net asset value, and the industry holds 32,000 unsold companies valued at $3.8 trillion. Open-ended vehicles fed every month by employee contributions would be natural buyers in the secondary market and in continuation funds. That is also the opponents’ most formidable argument: making savers carry what institutions no longer want. The answer exists, and it is in the text: the valuation factor requires regular and accurate valuation, free from conflicts of interest.
A premium for platforms. Only managers with open-ended vehicles, frequent valuation and agreements with recordkeepers and target-date fund managers will be able to serve this market. Consolidation in the industry will accelerate; the mid-market will gain access only indirectly, through funds of funds, secondaries and co-investment.
Fees, the heart of the matter. In 2024, 401(k) participants paid an average of 0.26% a year on equity funds and 0.29% on target-date funds. PwC estimates the annual revenue generated by $1 trillion in private markets assets at $7 billion to $19 billion, or 0.7% to 1.9%: between 2.4 and 6.6 times the average cost of a target-date fund. The fee factor in the safe harbor will force every basis point to be justified. The traditional model of management fees and carried interest is likely to undergo substantial change.
Liquidity, the Achilles’ heel. On December 1, 2022, Blackstone had to limit redemptions from its $69 billion unlisted real estate fund for individual investors, as requests exceeded its caps of 2% of assets per month and 5% per quarter. The mechanism worked as designed. In a 401(k), the same episode would become a political affair.
Three milestones will decide the fate of the reform: the DOL’s response to the senators, due October 8; the final rule, still in preparation; and above all the Intel ruling, expected before the end of the Court’s term in June 2027.
Appendix. American retirement in figures
The United States combines a public pay-as-you-go system, Social Security, with funded retirement savings that are twenty times larger than that system’s reserves.
| Pillar | Indicator | Value | Date |
|---|---|---|---|
| Pay-as-you-go: Social Security | Beneficiaries | 70 million people | December 2025 |
| Pay-as-you-go: Social Security | Covered workers | 185 million people | 2025 |
| Pay-as-you-go: Social Security | Workers per beneficiary | 2.9 | 2026 |
| Pay-as-you-go: Social Security | Benefits paid | $1.609 trillion | 2025 |
| Pay-as-you-go: Social Security | Income | $1.449 trillion | 2025 |
| Pay-as-you-go: Social Security | Reserves | $2.561 trillion, against $2.721 trillion a year earlier | End of 2025 |
| Pay-as-you-go: Social Security | Payroll tax | 6.2% employee and 6.2% employer, on up to $184,500 of earnings | 2026 |
| Pay-as-you-go: Social Security | Replacement rate at full retirement age | 79% (very low earners), 43% (average earners), 28% (earners at the cap) | Retiring in 2026 |
| Pay-as-you-go: Social Security | Reserve depletion | Retirement trust fund: late 2032; combined with the disability fund: 2034; thereafter, about 78% of scheduled benefits payable | 2026 projection |
| Funded | Total retirement assets | $51.2 trillion | June 30, 2026 |
| Funded | of which individual retirement accounts (IRAs) | $19.9 trillion | June 30, 2026 |
| Funded | of which employer defined contribution plans | $15 trillion, of which $10.8 trillion in 401(k)s | June 30, 2026 |
| Funded | of which public defined benefit plans | $10.4 trillion | June 30, 2026 |
| Funded | of which private defined benefit plans | $3.2 trillion | June 30, 2026 |
| Funded | Private-sector workers with access to an employer plan | 72% | March 2025 |
| Funded | Private equity share of public pension plans | 13.9% of assets, against 3.6% in 2001 | Fiscal 2025 |
Total retirement assets include other categories not detailed here. Social Security data cover retirement and disability.
References
12. Bloomberg, “Trump 401(k) Proposal Shows Signs of Phony Supportive Comments,” August 13, 2026
14. Internal Revenue Service, “401(k) plans,” information page
16. Internal Revenue Service, news release IR-2025-111, November 13, 2025
18. U.S. Department of Labor, “Employee Retirement Income Security Act (ERISA),” information page
21. U.S. Department of Labor (EBSA), “Target Date Retirement Funds — Tips for ERISA Plan Fiduciaries”
22. Pension Benefit Guaranty Corporation, “History of PBGC”
23. Supreme Court of the United States, Tibble v. Edison International, No. 13-550, May 18, 2015
25. Gibson Dunn, “U.S. DOL’s 401(k) Rule — Analysis of Public Comments,” June 22, 2026, p. 2
26. Senate Banking Committee (minority), Warren-Sanders press release, September 29, 2026
27. 401(k) Specialist, “Labor Critics Demand Investigation into Fake Submissions,” September 23, 2026
28. 5 U.S. Code § 801 (Congressional Review Act), Legal Information Institute, Cornell Law School
29. Bipartisan Policy Center, “The 2026 Midterms: Key Dates and Events”
36. 401(k) Specialist, “Anderson v. Intel Case Heads to Supreme Court Tuesday,” October 2, 2026
40. PwC, “Private markets in 401(k) plans: The $1T opportunity”
42. Bain & Company, Global Private Equity Report 2026, press release of February 23, 2026
46. Bipartisan Policy Center, “2026 Social Security Trustees Report, Explained,” June 9, 2026
47. Social Security Administration, “Understanding the Benefits,” publication No. 05-10024, January 2026
