On December 9, 1963, the Studebaker Corporation announced it was closing its main automobile plant in South Bend, Indiana. The company had been building things to ride in since before there were engines to put in them — wagons, in the 1850s — and for the people on the line the pension was not a perk. It was the arrangement. You gave the company the usable decades of your body, and at the end of them the company gave you an income you could not outlive. That was the deal, and in South Bend a great many people had organized an entire life around believing it.
When the plant closed, the plan was terminated, and the arithmetic underneath it became visible for the first time. There was not enough money in it. Roughly 3,600 workers who had already reached retirement age received their full pensions. Around 4,000 more, aged forty to fifty-nine with at least a decade of service, received a lump sum worth about fifteen cents on the dollar of what they had been promised. The rest, the youngest, the ones with the most working life still ahead of them and the most years to have counted on the promise, got nothing at all. More than 8,500 people in one city learned in roughly the same week that the thing they had built their lives around was a sentence the company was not legally required to finish.
What happened next is the part worth holding onto. South Bend became a national argument. It fed a decade of congressional investigation — a presidential committee under Kennedy had already been picking at corporate pension practice, and Senator Jacob Javits introduced reform legislation in 1967 — and it produced, in 1974, the Employee Retirement Income Security Act, signed by Gerald Ford. ERISA imposed minimum funding and vesting rules, placed fiduciary duties on the people running the plans, and created the Pension Benefit Guaranty Corporation, a federal insurer modeled on the one that backs bank deposits, whose entire purpose was to make certain a Studebaker could never happen to anyone again.
So the country built a machine. A large, specific, expensive machine, assembled out of a real catastrophe, whose single job was to keep the risk of growing old from being dumped on the individual worker the way it had been dumped on the people in South Bend.
And then, over the following forty years, the machine was quietly taken apart. Not by failing. Not by being repealed. By being routed around. The thing that replaced the pension was never built to replace the pension; it was a tax provision, and nobody who wrote it was thinking about the people in South Bend, or about you. The protection grew more elaborate, and the thing it was protecting was moved out from under it.
This is the same argument as the last sixteen posts. It just stopped being an argument and became an amount of money someone else is allowed to keep.
The Thing That Replaced The Pension Was Never Built To
Section 401(k) entered the Internal Revenue Code as a minor provision of the Revenue Act of 1978. Its purpose was narrow and slightly boring: it was meant to put limits on a kind of cash-or-deferred arrangement that companies had been using to let highly paid people shelter compensation. It took effect on the first day of 1980, and it would probably have stayed minor and boring except that a benefits consultant named Ted Benna, at a small firm outside Philadelphia, read it that year and saw that the language could be turned around. Read aggressively — his word, later — it could be the basis for an entire employer-sponsored salary-deferral savings plan, with a matching contribution as the hook to pull ordinary employees in. The IRS issued rules in 1981 that made the salary-deferral mechanics work, and the thing was loose in the world. None of the people involved were trying to replace the pension. There is essentially no legislative record of anyone debating whether a tax provision should become the primary way a hundred million people would attempt to grow old. The man most associated with it spent a good part of his later life saying, with visible regret, that he had helped pry open a door for the financial industry that he had never meant to open.
That is the origin of the instrument. The origin of the swap — the deliberate replacement of the defined-benefit pension by the defined-contribution account across the private-sector economy — is a separate thing, and it happened on purpose, and it was sold with a specific story. The story was modernization. The pension was cast as a relic of a one-employer-for-life economy that no longer existed; the 401(k) was freedom, portability, ownership, control, your money with your name on it that you could carry from job to job. Almost every word of that pitch was about the worker’s autonomy. Almost none of it was about what was actually changing hands. Private-sector defined-benefit coverage fell from somewhere near 38 percent of workers at the start of the 1980s to roughly 15 percent now, and at no point in that collapse was the disclosure honest about the part of the transaction that mattered, which was never who controlled the money. It was who was holding the risk.
A few honest qualifications before any of that, because the argument does not survive without them. This is not a case against the 401(k), or against owning the account with your name on it. Individual ownership solved a real problem: the old defined-benefit pension locked you to one employer, because leaving early forfeited benefits you had been promised, and a worker who changed jobs four times could retire having genuinely earned a pension and still collect almost nothing. Portability was not a marketing fiction. It was a real gain, and a system that pretends otherwise is lying in the other direction. Nor is this nostalgia for a pension golden age that did not exist. Defined-benefit plans were chronically underfunded by the companies that sponsored them, vesting rules were designed to shed workers before the promise matured, and when a sponsor went under the promise went with it — which is not an abstraction, it is the entire reason the opening scene happened and the entire reason the protective machinery had to be invented. The argument is narrower and harder than “pensions good, 401(k)s bad.” It is that a transfer of three risks onto the worst-positioned party was sold as a convenience upgrade, that the disclosure was dishonest about what was actually changing hands, and that a permanent toll was installed on the money in between and then defended for forty years. Saying the swap was a bad deal is not the same as saying the thing it replaced was a good one. Both were bad in different directions, and the people who profited engineered the second badness on purpose.
And this one is a consensus, not a partisan project, which the post is not going to disguise in either direction. ERISA passed with broad bipartisan support and a Republican signature. The 401(k) was a bipartisan accident — a 1978 provision and a 1980 reading — entrenched across administrations of both parties, and the fee structure that is the quiet deliverable has been protected, or left unprotected, by both as well; the rule requiring advisers to act in the saver’s interest has been written, struck down, rewritten, and struck down again across administrations of both parties, with the next attempt already back on the regulatory calendar. There is no clean villain to point at for the swap itself, and inventing one would be its own kind of dishonesty. But one piece is asymmetric, and flattening it into “both sides” would be the opposite dishonesty. The specific claim that Social Security is “going bankrupt” is not what the numbers say — the 2025 Trustees Report projects the retirement trust fund’s reserves depleting in 2033, after which incoming payroll taxes still cover roughly 77 percent of scheduled benefits; “depleted” and “bankrupt” are not the same statement, and the gap between them is the entire argument. The single change that would close most of the shortfall — lifting or removing the payroll-tax cap, which in 1983 covered about 90 percent of all covered wages and now covers about 83 percent because the income above it grew so much faster — has been kept off the table, and the campaign to keep it there and to sell the “bankrupt” framing instead has come overwhelmingly from one direction. The consensus is real. So is the asymmetry. The work is to say both without using either to launder the other.
What Moved, And Who Was Standing In The Gap
The cleanest way to see what the swap actually did is to stop arguing about it and itemize it, the way you would itemize anything transferred from one party to another. Read it as a ledger. There are four entries. None is an accident and none requires a conspiracy; each is the predictable behavior of named institutions responding to incentives that statute and regulation built and then declined to unbuild. The first entry is the one everything else sits on top of.
The first entry is the risk itself, and there are three kinds of it, and the worker held none of them before and holds all of them now. Investment risk: in a defined-benefit plan the employer and a pooled fund carry whatever the market does; in a defined-contribution account, you do, alone. Longevity risk: a pension pays until you die, however long that takes, because the pool covers the people who live a long time out of the people who do not; an individual account has no pool, and since you cannot know your own death date you are left to either underspend out of fear or outlive the money. Sequence-of-returns risk: a pension does not care what the market did the year you turned sixty-five, and an account cares about almost nothing else — two workers with identical contributions and identical average returns retire with wildly different outcomes depending only on whether the bad years fell at the start of the drawdown or the end. The individual is the worst-positioned party in the entire economy to absorb any one of the three, having no pool, no actuary, and no second chance, and the swap placed all three on the individual and called it control. The pitch said you were being handed the steering wheel. What you were handed was the liability.
The second entry is the one that never appears on any statement you will receive. The Department of Labor publishes its own neutral illustration of what fees do: a worker with thirty-five years to go and a $25,000 balance, 7 percent average returns, no further contributions. At total costs of half a percent, that becomes about $227,000. At one and a half percent, it becomes about $163,000. A single percentage point, compounded across a working life, removes 28 percent of the final number, and that is the government’s own restrained example — real plans, especially the ones offered by small employers, run worse. The recipient is a permanent industry: fund managers, recordkeepers, advisers, whose take does not depend on whether the account rises or falls, on whether the decade was good or lost, on whether you retire able to eat. And because 28 percent of everything is worth defending, the defense has been continuous. A 2016 Department of Labor rule that would have required advisers to act in the saver’s interest was vacated by the Fifth Circuit in 2018. A narrower 2024 replacement was blocked in court and, by early this year, abandoned. The next version is already back on the regulatory calendar, and the fact that each successive version is drafted to be just survivable enough to clear the same litigation is the tell. Nobody fights this hard, this long, over a fee that is incidental to the design.
The third entry is the one that proves “evolution” is the wrong word. The existing stock of pension promises was not allowed to lapse gently; in a great many cases it was actively taken. The 1980s produced a reversion industry in which overfunded plans were terminated specifically so the sponsor could pocket the surplus. The bankruptcy route produced a different mechanism: a firm enters Chapter 11 and hands its pension obligations to the PBGC, the way United Airlines did in 2005, terminating all four of its plans, underfunded by something like ten billion dollars, the largest pension default in the country’s history to that point, with the workers’ promised benefits marked down to the federal guarantee ceiling. The multiemployer wave produced a third: trucking deregulation, the 2008 crash, and declining union density pushed a set of collectively bargained plans toward collapse, and the resolution, in 2021, was the Butch Lewis provision of the American Rescue Plan — the public, through the PBGC, writing grants to keep those plans paying through 2051. Read that last one slowly. The cost of the stripped promises did not vanish. It was moved onto the same public that had been told the private system was the responsible one. The machine from the opening did not fail here. It was used as the place to put the bill.
The fourth entry is the one that makes the failure look like it was yours. The entire model rests on an assumption: that workers will voluntarily set aside ten to fifteen percent of income for forty years. On the wages this series documented in Part 3, most workers structurally cannot, and the balances prove they do not. The honest figure is not the average, because averages are dragged upward by a small number of very large accounts and are built to reassure. The median is the honest one. In the 2022 Survey of Consumer Finances, only about 54 percent of families have any retirement account at all, and the median household aged fifty-five to sixty-four — people for whom retirement is not theoretical — holds on the order of ten thousand dollars in those accounts. Roughly half the country approaching retirement will enter it dependent almost entirely on Social Security. The Congressional Budget Office has estimated that the shift from defined-benefit to defined-contribution accounts explains about a fifth of the entire increase in American wealth inequality between 1989 and 2019. And the design’s most valuable feature is what it does to the explanation. A system that asks the structurally unable to do the structurally impossible, and then records their failure to do it as a personal shortfall — a discipline problem, a character flaw, you should have saved more — has arranged for its own outcome to be heard as the saver’s fault. That recoding is the cruelty mechanism Part 16 took apart in full, the one that lets hardship register as a moral verdict instead of a designed result; the complete treatment is there and this post hands it back rather than rebuilding it. The point that belongs here is narrower and worth saying flatly: the fiction is not a flaw in the model. The fiction is what the model is for.
What’s Real
A few honest qualifications, because the strongest version of the argument is the one that doesn’t pretend the counter-arguments aren’t there.
The defined-benefit pension the swap replaced was not the thing nostalgia remembers. Its great structural cruelty was that it rewarded staying and punished leaving: benefits accrued back-loaded and vested on cliffs, so a worker who gave a company twelve good years and then left, because the plant moved or the marriage moved or the better job was elsewhere, could walk away having earned on paper a pension worth almost nothing. The economy the pension was built for, in which one firm carried one worker across one career, was already disappearing when the swap began, and for the large and growing share of people who would change employers many times, the pension’s promise was frequently a promise the structure was designed to let them never quite reach. The 401(k) genuinely solved that. An account with your name on it that travels with you is a real good, made by serious people for a real reason, and a critique that cannot say so plainly is not credible.
There is also a serious, non-cynical case that individual ownership carries value the pension never offered: a balance you can see, direct, and leave to someone, rather than an opaque promise sitting on the books of an employer whose solvency you cannot audit and whose actuarial assumptions you will never be shown. Public-sector defined-benefit plans, the usual counterexample, have their own well-documented pathology — elected officials promising future benefits they decline to fund in the present, leaving the bill for a later legislature and a later taxpayer. The argument of this post is not that every old pension was good and every account is bad. It is the narrower and harder thing: that a transfer of three quantifiable risks onto the least-equipped party was disclosed as a convenience, and that a permanent toll was installed in the path of the money and defended for four decades by people who knew precisely what they were defending. The portability was real. It was also the wrapper.
What They’re Paying For
What the swap actually produced, set against what it was sold as, is four things stacked together. Each is worth real money to someone specific, and the someone is nameable in every case, which is the difference between this and a complaint about Wall Street.
It is a balance-sheet liability converted into your problem. A defined-benefit pension is a debt the company owes the future, and it sits on the books as one; the public corporation that froze its plan and pushed its workers into a 401(k) did not make that debt smaller, it made it disappear from its own ledger by relocating it onto roughly fifty million household ledgers, one anxious retirement at a time. The beneficiary is not abstract. It is the specific employer who turned a funded promise into a quarterly line item that ends the day the paycheck does, and booked the difference as savings.
It is a permanent toll on the largest pool of middle-class money in the country. Trillions of dollars in defined-contribution accounts now run through an industry of fund managers, recordkeepers, and advisers whose take depends not on the outcome but only on the assets sitting there, and whose entire decade-long political project, the fight over the fiduciary rule, has been the defense of the right to be paid for advice that does not legally have to be in your interest. They are not skimming in spite of the system. The skim is what the system was built to deliver.
It is a structural failure issued back to the people it happened to as a personal one. The savings-rate fiction does not merely produce inadequate balances; it produces an explanation for them, and the explanation is you. That engine is the one Part 16 disassembled, and it is not reassembled here; what matters for this layer is only that the retirement system and the cruelty theology turn out to be the same machine seen from opposite ends. One asks you to do the impossible. The other explains that failing to do it was a defect in your character.
And it is a backstop kept deliberately weak so that none of the above has to be survivable. Social Security is the only universal, pooled, inflation-protected retirement income most Americans will ever have, and it is the one piece of the structure that could simply be expanded into the hole the swap dug. It has instead been described, for forty years and overwhelmingly from one direction, as a program on the verge of bankruptcy, a description the 2025 Trustees Report does not support, since the projected 2033 depletion of the retirement trust fund still leaves ongoing payroll taxes covering roughly 77 percent of scheduled benefits — a serious problem, and not the same statement as “it will not be there.” The single change that closes most of the gap, lifting the payroll-tax cap that covered about 90 percent of wages in 1983 and about 83 percent now, is the change kept furthest from a vote. A backstop is most useful to the people who built the swap when it is frightening enough to look doomed and weak enough to never actually catch anyone.
The Fixes Are Boring
The fixes below are structural and economic. None is a budgeting tip, and the post is not going to pretend “financial literacy” or “save more” belongs on the list — those are not solutions to a designed risk transfer, they are the fourth ledger entry wearing a helpful voice. The fixes sound impossible because the people who profit from the current arrangement have spent forty years making them sound impossible, not because they are radical or untested. Most of them are approximately what other wealthy democracies already do. Roughly cheapest and most immediate to genuinely hard:
- Hard statutory fee caps and a fiduciary standard with teeth. Cap what any fund offered inside a retirement plan can charge, and make everyone advising on retirement money legally obligated to act in the saver’s interest rather than merely to avoid the most flagrant conflicts. The 2016 rule that attempted the second half of this was vacated in 2018, the 2024 replacement is effectively dead, and the version being drafted this month is being written narrow enough to survive the same industry — which is the argument for writing it wide.
- Auto-enrollment and auto-escalation as a federal floor, not an employer’s option. Default participation roughly triples it, and the 2022 retirement law made it mandatory only for new plans while exempting small employers; the state programs show it works and also show its ceiling, since low-paid workers still opt out at high rates and the resulting balances stay small. This is a real partial fix, and it is honest to call it partial — it improves the swap, it does not undo it.
- Lift or eliminate the Social Security payroll-tax cap and raise the floor of benefits. Removing the cap closes most of the shortfall the 2025 Trustees Report describes, and does it without touching anyone below roughly the ninety-fourth percentile of earnings. Every “grand bargain” floated since 1983 has paired a token revenue increase with benefit cuts that were the actual payload; the revenue side can simply be done by itself.
- Outlaw the pension-reversion and obligation-dumping mechanics and fund the PBGC to do its job. The surplus-stripping terminations of the 1980s and the bankruptcy hand-offs that ran through United in 2005 and the multiemployer plans into the 2021 public bailout were specific transactions enabled by specific rules. Close the rules; stop operating the federal guarantor as the discount window where stripped promises get marked down.
- Build the universal, portable, low-fee public default that already half exists. The federal Thrift Savings Plan proves a government-run, near-zero-fee account works at scale; the state auto-IRA programs prove an account that follows the worker across employers is administrable. The pieces have been kept deliberately separate, because assembled they compete directly with the toll.
- Restore a real employer funding obligation, not an offer obligation. Australia requires employers to contribute roughly twelve percent of wages on top of pay; the American “match” is optional, skewed toward the workers who need it least, and structured to look like a contribution while obligating nothing. The optional match was the diluted version, and it was diluted on purpose.
- Put longevity risk back into a pool. Collective defined-contribution, or mandatory annuitization of part of every balance, restores the one thing the swap specifically removed: a structure in which the people who live a long time are covered by the people who do not, so that no individual has to either starve out of caution or gamble on their own death date. This is the genuinely hard one, and it is last for a reason — it is the piece the industry will fight hardest, because a pool is the part of the money it cannot charge a per-account fee to stand next to. Every voluntary version has failed for the precise reason that it was voluntary.
Six of the seven are statutory or regulatory — rules a Congress or an administration could write or enforce without inventing a single new institution. The seventh is what the country had, imperfectly, before it decided not to. None of them is exotic, and none of them is the radical position; the radical position, historically speaking, is the one we are living in, in which the entire risk of growing old was moved onto the people least able to carry it and the move was called freedom. The reason the fixes feel impossible is not that they fail elsewhere. It is that the people they would cost have spent four decades funding a conversation in which proposing them sounds naïve.
Who Is This For
The retirement system told you it was handing you control. What it handed you was three risks, a permanent toll, and an explanation, prepared in advance, for why the result was your fault. The next post stays on the money but moves up one layer, into the tax code itself, where the same logic runs without even needing a story about freedom to carry it — and the cleanest door from here to there is the 401(k) again, because the deduction that subsidizes it is itself one of the largest tax expenditures in the entire code and one of the most upside-down, worth the most to the highest earners in the highest brackets and almost nothing to the workers the program is publicly justified by. The instrument that moved the risk onto you is also a subsidy aimed over your head. Part 18 is about how much of the real damage in the tax code is done by provisions exactly like that one — the ones nobody is fighting about — while the fights you are shown, the ones about billionaires’ headline rates, mostly are not where the money is. Different layer. Same question.
Even that's Odd 
Leave a comment