Every month, the same blunt question gets typed into search boxes hundreds of thousands of times, phrased a dozen different ways: pension vs 401(k), which one is better, and why did the better one vanish? The person typing it is often a mid-career worker holding two different promises from two different employers, one a guaranteed monthly check and the other an account balance that rises and falls with the market. Sometimes the searcher is a student writing a paper on economic inequality, or a journalist looking for a clean explanation of a transformation that played out over four decades. All of them want a verdict. Before any verdict is defensible, though, the two models have to be described as machines, because pensions and 401(k) plans do not merely pay different amounts. They assign investment risk, longevity risk, and legal obligation to different parties, and Congress spent half a century rewriting those assignments one statute at a time.

The rest of this account follows the sequence: first the two machines, then the laws that changed their relative cost, then the precise transfer of risk, and finally the competing explanations that the second half of this article weighs against each other.

Pension vs 401(k): How the Two Models Actually Work

A defined benefit pension begins with a promise stated as a formula. A typical plan might pay, at retirement, a percentage of the worker’s final average salary multiplied by years of credited service: one and a half percent per year of service, for example, so that thirty years yields forty-five percent of final pay, delivered as a monthly check for life. The worker does not own a pot of money. The worker owns a contractual right to a future stream of payments, and the employer owns the obligation to make that stream real. To meet it, the employer contributes money into a trust fund, hires investment managers, and relies on actuaries to estimate how much must be set aside each year so that the fund can pay every promised check. If the investments perform poorly, the employer must contribute more. If retirees live longer than the actuaries projected, the employer must keep paying anyway. The two great financial risks of retirement, the risk that markets disappoint and the risk that people outlive their savings, sit squarely on the employer’s side of the ledger. That allocation of risk is the defining feature of the pension, and everything Congress later did to the model touched it.

The pension’s promise came with conditions that shaped workers’ lives in ways the account system never did. Benefits had to vest, meaning a worker who left too early kept little or nothing, and for most of the pension era vesting required long service. A worker who changed jobs every few years, which described a growing share of the labor force, accumulated only fragments. Lump-sum payouts were permitted in some designs but were not the norm, and the standard payout was the life annuity: a check that arrived every month until death, which was generous to the long-lived and invisible to anyone trying to compare offers across employers. Most private-sector pensions included no automatic adjustment for inflation, so the purchasing power of that check eroded year by year unless the employer voluntarily granted increases. And because the promise stretched decades into the future, its value depended entirely on the employer’s survival. A bankrupt sponsor with an underfunded plan could leave retirees with broken promises, a failure the public witnessed when the Studebaker automobile company closed its South Bend plant in 1963 and workers discovered how little stood behind their expected benefits. That episode, more than any other single event, supplied the political momentum for the federal intervention that followed.

The 401(k) inverts the machine. It begins not with a promise but with an election. The worker decides what percentage of each paycheck to divert into an individual account, choosing between pre-tax contributions that reduce current taxable income and, after later legislation, after-tax Roth contributions that promise tax-free withdrawals. The employer may add matching contributions, and most large employers do, but the match is voluntary, can be changed or suspended, and is itself subject to vesting schedules. The money in the account is then invested according to the worker’s own choices among the funds the plan offers. At retirement, the worker owns whatever the account has become: contributions plus investment returns minus fees. There is no formula, no promised check, and no party obligated to make the balance larger than the market made it. If the market falls in the year before retirement, the balance falls with it, and no one is required to top it up.

That inversion moves both great risks across the table. Investment risk belongs entirely to the worker, who must choose an asset mix without professional obligation on anyone’s part to get it right. Longevity risk belongs to the worker as well: the account must be managed through an unknown remaining lifespan, withdrawn carefully enough to last and aggressively enough to live on, a calculation that professional pension actuaries once performed for thousands of people at once and that each retiree must perform alone. Annuities can be purchased with the balance, converting it back into a lifetime stream, but nothing in the standard design requires it. What the worker gains in exchange is portability, the feature the pension never offered. A 401(k) balance can be rolled into an individual retirement account or into a new employer’s plan when jobs change, so the job-hopper who accumulated fragments under the old system accumulates a continuous balance under the new one. And because the account is legally the worker’s property, the employer’s bankruptcy does not threaten it the way it threatened an underfunded pension promise, though nothing in federal law insures the balance against investment loss. There is no pension-style insurance fund standing behind a 401(k); the guarantee that once backed the monthly check has no counterpart in the account world.

The contrast, stated as mechanics rather than as moral judgment, is this: the pension pooled risk across the workforce and the employer’s balance sheet, while the 401(k) individualizes risk into a single account and a single lifespan. The pension rewarded long tenure with a single employer and punished mobility. The account rewards mobility and punishes bad timing, poor fund selection, and long life without a plan for it. Neither design is self-evidently superior in the abstract, because each serves a different kind of working life. But the question of which system a worker gets was not left to the worker to decide. It was shaped, step by step, by statutes that changed what each design cost to operate.

The 1974 Statute That Priced the Promise

When Congress passed the Employee Retirement Income Security Act in 1974, it was responding to genuine failures. The Studebaker collapse and a series of smaller plan failures had shown that pension promises could be worthless, and the legislative answer was to make the promise enforceable through a lattice of new duties. President Gerald Ford signed the measure on September 2, 1974, as Public Law 93-406, and its requirements fell into five interlocking parts, each of which added a layer of fixed cost and legal exposure to every defined benefit plan in the country. For readers who want the full architecture of that landmark, the complete guide to the 1974 law walks through each title in detail.

First came vesting. Before the statute, an employer could require twenty or more years of service before a worker owned any benefit at all, and forfeiture on job change was routine. The new law capped the wait: plans had to vest benefits under schedules no slower than ten-year cliff vesting, under which the entire benefit vested at once after a decade, or graduated schedules that vested portions of the benefit over a longer span. Workers who left after vesting kept what they had earned, and the employer’s ability to use forfeitures as a quiet subsidy for the plan disappeared.

Second came funding. The statute imposed minimum funding standards that required sponsors to contribute enough each year to keep the plan on track toward its promises, with shortfalls amortized over periods fixed by law. Actuaries became mandatory participants in the life of every pension, because only an actuary could certify that the contributions satisfied the standards. A plan could no longer be run on optimistic assumptions and catch-up contributions; the funding obligation became a legal duty with a timetable, and missing it triggered penalties and excise taxes.

Third came fiduciary duty. Anyone exercising discretion over plan assets became a fiduciary bound to act solely in the interest of participants, to invest with the care of a prudent expert, to diversify holdings, and to avoid transactions with parties connected to the plan. The prohibited-transaction rules fenced off an entire category of dealings that had once been routine, such as lending plan money to the sponsoring company on friendly terms. Breach carried personal liability. A corporate officer who served as a plan trustee was no longer merely managing a benefit; the officer was accepting a legal standard of care enforceable in federal court.

Fourth came reporting and disclosure. Plans had to file detailed annual reports with the federal government, the filings that became known as Form 5500, and had to furnish workers with written summaries of their rights. The administrative apparatus of the pension, recordkeeping, legal review, actuarial certification, government filings, grew from a back-office function into a compliance operation.

Fifth came insurance. The statute created the Pension Benefit Guaranty Corporation, a federal entity charged with taking over the obligations of failed plans and paying benefits up to a statutory maximum. Sponsors of single-employer plans paid premiums for this coverage, beginning at one dollar per participant per year, and the premium was a new line item on the cost of every pension that had no counterpart in any alternative design. The guarantee made the promise credible to workers, which was its purpose, and it simultaneously converted a contingent risk into a certain annual cost for employers, which was its consequence.

None of these requirements was irrational on its own terms. Each addressed a real abuse or a real failure, and the statute’s defenders, then and since, argued that a promise unenforceable in law is not a promise at all. But the cumulative effect was to transform the defined benefit plan from a relatively informal employer practice into one of the most heavily regulated financial arrangements in American business. Large employers adapted by building benefits departments, hiring actuaries and ERISA counsel, and absorbing the expense. Smaller employers, for whom a pension had already been a stretch, increasingly concluded that the regulatory price exceeded the recruitment value. The statute did not order anyone to close a pension. It made keeping one a profession in itself, and the number of new defined benefit plans began a long decline from which it never recovered.

Two further features of the 1974 design compounded the cost over time, and both deserve attention because they illustrate how legal exposure, not just compliance expense, priced the pension out of reach. The first was the trajectory of the insurance premiums. The one-dollar-per-participant charge of 1974 did not stay at one dollar. Congress raised the flat-rate premium repeatedly in subsequent legislation, and the 1987 law added the variable-rate premium tied to underfunding, so that the sponsors in the weakest financial condition paid the most. A sponsor nursing an underfunded plan through a recession faced the perverse arithmetic of rising insurance bills at the exact moment cash was scarcest. The guarantee that made the promise credible thus became, for struggling sponsors, an additional weight on the plan it was meant to protect. The second feature was the asymmetry of the anti-cutback rule. Once benefits were accrued, the statute generally forbade reducing them by plan amendment, which meant the employer’s obligation could move in one direction only. A sponsor could make the formula less generous for future service, but what workers had already earned was locked. That protection was the point, and few would argue workers should lose earned benefits to a sponsor’s changed mind. But the one-way mechanism made sponsors cautious about granting generous formulas in the first place, because every increase was permanent while every decrease applied only going forward. Fiduciary litigation added a third layer of exposure. The prudent-expert standard invited lawsuits whenever plan investments lost money, and the following decades produced a steady stream of cases testing whether fiduciaries had monitored fees, diversified adequately, and avoided conflicts. Each case, win or lose, added to the legal budget of maintaining a pension and to the perceived risk of serving as a fiduciary. Corporate officers who might once have viewed pension oversight as a routine duty came to view it as a personal liability exposure, and that perception filtered into boardroom discussions about whether the plan was worth keeping.

Why did employers stop offering traditional pensions?

Congress raised the fixed cost of pensions in stages: ERISA in 1974 imposed vesting and funding duties plus insurance premiums, and 1980s tax laws capped deductions and penalized reversions, while a 1978 tax provision plus 1981 guidance made 401(k) accounts cheap to run. Employers faced a cost gap between the two designs, and new plans followed the cheaper one.

The forfeiture economy that had quietly subsidized pensions deserves a closer look, because its removal explains part of the cost shock. Before vesting rules, a meaningful share of accrued pension value never reached workers at all; employees who left before the vesting threshold forfeited their benefits, and those forfeitures stayed in the plan and reduced the employer’s funding burden. The practice was harsh on mobile workers, which was one reason Congress outlawed its harshest forms, but it had functioned as a discount on the price of the promise. When the law required that vested workers keep their benefits and take them along in deferred form, the discount vanished and the true cost of the promise became visible for the first time. Employers who had believed they were running an affordable benefit discovered they had been running an affordable benefit only because so many workers left empty-handed. That discovery, spread across thousands of plan sponsors in the second half of the 1970s, set the stage for everything that followed, because it meant the pension’s economics were already under strain before the alternative arrived.

The 1978 Provision and the Account It Made Possible

The alternative arrived almost by accident, tucked into a tax bill whose main business lay elsewhere. The Revenue Act of 1978, signed by President Jimmy Carter on November 6, 1978, as Public Law 95-600, contained a provision that added subsection (k) to section 401 of the Internal Revenue Code. The new subsection addressed cash-or-deferred arrangements: setups in which an employee could elect to receive a sum as current cash compensation or to defer it into a profit-sharing or stock bonus plan. Amounts deferred under a qualifying arrangement would not be taxed as current income, even though the employee had, in a formal sense, chosen to defer them. The provision was brief, technical, and aimed at a narrow question about the taxation of profit-sharing contributions. Nothing in its text suggested that it would reorganize American retirement.

For two years the provision sat largely unused, because its meaning in practice was unclear. Employers and their advisers could read the statute, but they could not be certain the Internal Revenue Service would accept salary reduction contributions, money the employee plainly could have taken as cash, as genuinely deferred and therefore not currently taxable. Tax lawyers are paid to resolve exactly this kind of uncertainty, and the resolution they awaited came in November 1981, when the Service issued proposed regulations confirming the treatment: elective deferrals under a qualifying cash-or-deferred arrangement would be excluded from the employee’s gross income in the year contributed and taxed only upon distribution. The proposal issued on November 10, 1981, in the ASPPA account of the regulations’ history. With that guidance in hand, the legal risk of building a plan around the provision collapsed.

The first employers to act moved quickly. A benefits consultant named Ted Benna, working with his own firm, designed what is widely credited as the first 401(k) plan in 1981, seizing on the proposed regulations to build a working model around salary deferrals matched in part by the employer. Other companies followed within months, and the design spread first through industries with sophisticated benefits staffs and then outward. The appeal was not mysterious. A 401(k) required no actuary, no funding certification, no insurance premiums, and no promise about the distant future. The employer’s obligation ended, in the typical design, with each payroll cycle’s contribution and any promised match. Administration meant recordkeeping and investment menus, not mortality tables and funding schedules. Disclosure duties existed but were lighter than the pension’s. The prohibited-transaction and fiduciary framework of ERISA still applied, since a 401(k) is an ERISA plan, but the duties attached to a far simpler pool of assets: individual accounts rather than a single fund bearing a collective promise.

Congress watched the growth and, rather than restraining it, repeatedly adjusted the tax code in ways that favored the account model. Contribution limits for elective deferrals were set at levels generous enough to make the 401(k) a serious retirement vehicle for middle and upper-middle earners, and the limits were later raised in stages. The tax treatment, deferral of income tax until withdrawal, mirrored the pension’s own tax logic closely enough that workers perceived the two as comparable benefits while employers experienced them as radically different obligations. By the middle of the 1980s, the structural contest was effectively decided for new plans: any employer establishing a retirement benefit from scratch faced a choice between a heavily regulated promise carrying funding risk, insurance premiums, and fiduciary exposure, and a lightly regulated account carrying none of those. The flow of new plan formation went overwhelmingly to the account.

It is worth pausing on why the account’s cheapness was a legal artifact rather than a natural fact. Nothing inherent in individual accounts makes them inexpensive to administer; recordkeeping for thousands of individual balances, investment education, and compliance testing all cost money. They were cheap relative to pensions because the law had made pensions expensive and had left accounts lightly touched. Had Congress instead imposed funding standards, insurance premiums, and fiduciary liability on 401(k) sponsors comparable to what it imposed on pension sponsors, the cost gap would have narrowed or reversed. The gap was a policy choice expressed through differential regulation, and recognizing that is essential to understanding the four decades that followed.

How did the Revenue Act of 1978 create the 401(k)?

The Act added subsection (k) to section 401 of the tax code, permitting cash-or-deferred arrangements in which workers could divert salary into a plan without current taxation. Proposed IRS regulations in November 1981 confirmed that elective deferrals were not currently taxable income, removing the legal uncertainty, and employers began adopting 401(k) designs within months.

The early-1980s guidance did more than bless a tax treatment; it standardized a template. Once the Service had spoken, benefits consultants could sell the same basic design to client after client with only modest customization: a deferral election form, a menu of mutual funds, a matching formula, a vesting schedule for the match, and a summary plan description. Standardization drove costs down further and made the 401(k) accessible to mid-sized employers that could never have afforded a pension actuary. The pension, by contrast, resisted standardization, because each plan’s funding needs depended on its own workforce demographics, its own benefit formula, and its own investment history. Two employers with identical payrolls could face very different pension costs, while two employers with identical payrolls faced nearly identical 401(k) costs. Predictability itself became a feature of the account model, and finance officers, asked to forecast benefit expenses, preferred the design whose costs they could state with confidence.

The choice of host vehicle mattered as much as the tax treatment. Congress did not create a new plan type in 1978; it grafted the cash-or-deferred election onto the existing profit-sharing plan, a design employers already used to distribute a share of annual profits to workers. The graft was elegant because it solved the tax problem that had blocked salary reduction plans for years. Under the constructive-receipt doctrine, income a worker could have taken as cash was generally treated as received even if the worker directed it elsewhere, which meant salary deferral elections risked immediate taxation. The new subsection carved out a statutory exception: amounts an employee elected to defer under a qualifying arrangement would not be treated as constructively received, provided the plan met qualification requirements. The November 1981 proposed regulations then supplied the operational detail, specifying how elections had to be structured, how deferrals interacted with nondiscrimination testing, and how distributions would be taxed. With the doctrine neutralized and the mechanics specified, the profit-sharing plan, once a vehicle for discretionary employer contributions paid out of profits, became the chassis for worker-directed salary deferrals matched by the employer. The transformation required no new institutions, no new insurance program, and no new federal agency. It required only that an existing, lightly regulated vehicle be given a new tax feature, which is one reason the account system scaled so fast: the infrastructure was already in place.

The 1980s Tax Laws That Tightened the Pension Further

The 1974 statute had raised the regulatory price of the pension. The tax legislation of the 1980s raised its fiscal price, and did so through provisions that are less remembered than the 1978 change but arguably just as consequential for plan sponsors weighing their options. Each of these laws was primarily a revenue or deficit measure; pensions were collateral. But collateral damage, repeated across four major enactments in five years, compounded into a pattern that plan sponsors could read clearly: the defined benefit plan was becoming a less hospitable place to shelter compensation, while the account plan faced no comparable squeeze.

The Tax Equity and Fiscal Responsibility Act of 1982, signed September 3, 1982, as Public Law 97-248, opened the sequence. Its pension provisions reduced the statutory ceilings on benefits and contributions under section 415 of the tax code, shrinking the maximum pension a highly paid executive could accrue and therefore shrinking the tax advantage that had made generous executive pensions attractive to sponsors. It tightened the rules for integrating plan benefits with Social Security, limiting the extent to which employers could count government benefits against their own promises. And it introduced the top-heavy rules, a new compliance regime aimed at plans that disproportionately favored key employees, which required minimum benefits or contributions for rank-and-file workers in plans that disproportionately favored key employees. The top-heavy regime fell hardest on small employers, precisely the sponsors for whom pension administration was already a marginal proposition, and it added a testing and correction apparatus that made small defined benefit plans still more burdensome to maintain.

The Deficit Reduction Act of 1984, Public Law 98-369, whose tax title was the Tax Reform Act of 1984, continued the pattern of constraining the pension’s tax advantages as part of a broader deficit-reduction effort. Its significance in the sequence lies less in any single provision than in the signal it sent: two major tax bills in two years had treated the pension as a source of revenue to be harvested rather than a social institution to be nurtured. Plan sponsors making decade-long commitments read signals of that kind carefully, because a pension promise cannot be repriced once made. An employer considering a new defined benefit plan in 1984 had to reckon not only with the rules on the books but with the evident direction of travel.

The Tax Reform Act of 1986, signed October 22, 1986, as Public Law 99-514, struck harder. It accelerated vesting schedules, replacing the ten-year cliff with a five-year cliff and the longer graded schedule with a three-to-seven-year graded alternative, which further reduced the forfeiture subsidy and brought the full cost of benefits forward. It tightened the section 415 limits again and imposed new nondiscrimination testing that made it more difficult to design plans favoring higher-paid workers. Most dramatically for sponsors sitting on well-funded plans, it imposed a ten percent excise tax on reversions, situations in which an employer terminated an overfunded pension and reclaimed the surplus assets. The reversion tax addressed a genuine controversy: in the early 1980s, a wave of employers had terminated overfunded plans, captured the surplus, and left workers’ benefits to the insurance program, a practice critics described as raiding. The tax was meant to deter the raid. Its side effect was to trap surplus inside plans: an employer that funded conservatively, building a cushion against market downturns, could no longer recover the cushion without paying a penalty. Prudent overfunding, the very behavior the funding rules were meant to encourage, became financially punishable, and sponsors learned to fund to the minimum rather than to the comfortable.

The Omnibus Budget Reconciliation Act of 1987, signed December 22, 1987, as Public Law 100-203, completed the decade’s work. Its centerpiece for pensions was the full funding limitation, which capped deductible contributions at one hundred fifty percent of the plan’s current liability. Contributions beyond that line lost their deduction, which meant the tax code punished both overfunding (through the reversion excise) and advance funding beyond a statutory ceiling (through the deduction limit). The corridor of rational funding behavior narrowed to a band: fund too little and face penalties and insurance premiums, fund too much and lose deductions or face excise taxes on the way out. The same law introduced variable-rate insurance premiums, so that the Pension Benefit Guaranty Corporation’s charges rose with the size of a plan’s funding shortfall, converting underfunding from a quiet actuarial condition into an immediate cash cost. It also tightened minimum funding requirements for underfunded plans, accelerating the schedule on which sponsors had to close gaps. Each provision was defensible as policy: deductions should have limits, insurance should be risk-priced, shortfalls should be repaired promptly. Together, they made the defined benefit plan the most tightly supervised financial commitment a company could undertake, supervised on the way in, on the way out, and every year in between.

Step back and the decade reads as a single movement. In 1974, Congress had made the pension a regulated promise. Between 1982 and 1987, Congress made it a fiscally constrained one, limiting the benefits it could pay, the deductions it could generate, the surplus it could hold, and the surplus it could recover. Through all of this, the 401(k) faced nothing comparable. No funding standards applied to it, because there was no promise to fund. No insurance premiums applied, because there was nothing to insure. No reversion tax could touch it, because surplus belonged to individual accounts. The account plan grew in a regulatory climate that was, by comparison, nearly silent. Sponsors did not need an economic theory to see the difference; they needed only to compare the compliance calendar for a pension with the compliance calendar for a 401(k), and the headcount each required.

Defenders of the 1980s tightening argued, with reason, that much of it corrected genuine excess. Executive pensions had grown lavish, integration with Social Security had been used to hollow out rank-and-file benefits, and reversion raids had been real. Labor advocates and pension-rights groups supported the vesting acceleration and the nondiscrimination testing as protections workers had long been owed. Employer associations countered, also with reason, that the cumulative burden was driving sponsors out of the system the protections were meant to preserve, and that a protected benefit inside a terminated plan helps no one. Both arguments were sincere, and both described real effects. The laws passed because the abuses were visible and the exits were gradual; by the time the exits were visible, the laws were entrenched.

The 2006 Law That Made the Account the Default

By the early 2000s, the account system had won the contest for new plans, but it had a problem of its own: workers were not using it well. Participation was voluntary, and voluntary systems inherit human inertia. Large numbers of eligible workers never enrolled, and those who did often contributed too little or left their money in the plan’s most conservative option for decades. Behavioral economists had documented the pattern extensively, and a growing consensus held that the design of the enrollment choice mattered as much as the tax treatment. The Pension Protection Act of 2006, signed August 17, 2006, as Public Law 109-280, answered that diagnosis with the most consequential retirement legislation since 1974, and its answer entrenched the account system for a generation. For the mechanics of what the 2006 law built, the guide to the statute that entrenched the account system covers each provision in full.

The Act’s central innovation was legal safety for automatic enrollment. Before 2006, an employer that enrolled workers in a 401(k) without their affirmative consent risked violating state wage-payment laws, which in many states prohibited deductions from pay without explicit written authorization. A handful of employers had tried automatic enrollment anyway, and the Department of Labor had issued guidance suggesting it was permissible, but the legal footing was uncertain enough to deter most sponsors. The 2006 statute settled the question by preempting state laws that would have blocked automatic enrollment under qualifying arrangements, giving employers explicit federal permission to default workers into the plan at a stated contribution rate unless the worker opted out. Inertia, which had kept workers out of the plan, then worked to keep them in it. Participation rates at adopting employers rose sharply, and automatic enrollment spread from a rarity to the standard design for large plans within a few years.

The companion innovation concerned what happened to the money of workers who were enrolled automatically but never made an investment choice. Left to themselves, such workers’ contributions had typically flowed into money market or stable value funds, the plan’s designated safe parking place, where they earned little and fell far behind inflation over a career. The statute directed the Department of Labor to define categories of diversified default investments, known as qualified default investment alternatives, into which fiduciaries could place the contributions of non-choosing participants with protection from liability for the investment outcome. The Department’s final regulations, issued in October 2007, blessed target-date funds, balanced funds, and managed accounts as qualifying defaults. The effect was to automate not only saving but asset allocation: the defaulted worker was invested in a diversified portfolio calibrated to a retirement horizon, without ever signing a form. The fiduciary safe harbor was the key that unlocked adoption, because plan sponsors had feared exactly the lawsuit the safe harbor foreclosed, a claim by a worker whose defaulted investment lost money in a downturn.

The 2006 law also made permanent the higher contribution limits that Congress had enacted on a temporary basis in the Economic Growth and Tax Relief Reconciliation Act of 2001, removing the scheduled sunset that had left the account system’s generosity in doubt. And it refined the nondiscrimination testing framework to accommodate automatic enrollment designs, smoothing the compliance path for the very arrangements it was promoting. Every one of these provisions lowered the legal friction of running a 401(k) or raised its attractiveness as a benefit, continuing the pattern established in the 1980s of legislation that eased the account’s path.

There was, however, a less noticed side to the statute, and intellectual honesty requires stating it. The Pension Protection Act also tightened the funding regime for defined benefit plans, and tightened it substantially. It replaced the older funding framework with a requirement that plans target full funding of their liabilities, measured against market-based interest rates, with shortfalls amortized over seven years rather than the longer schedules previously allowed. Plans deemed at risk faced still stricter assumptions. Benefit restrictions kicked in automatically when funding fell below statutory thresholds, limiting lump sums and benefit improvements at precisely the moments sponsors were least able to afford them. For underfunded sponsors, the law converted a manageable long-term obligation into a near-term cash demand, and a number of employers responded by freezing their pensions, closing them to new entrants while preserving accrued benefits, or terminating them outright. The statute that made the 401(k) effortless simultaneously made the pension more demanding, and the two movements reinforced each other in the same legislative package. The pattern of the previous three decades, easing for accounts and tightening for pensions, was not broken in 2006. It was enacted in a single bill.

The statute’s treatment of contribution limits illustrated the same directional pattern in quieter form. The Economic Growth and Tax Relief Reconciliation Act of 2001 had raised the ceilings on elective deferrals and added catch-up contributions for workers aged fifty and older, but it had done so on a temporary basis, with the increases scheduled to expire. The 2006 law made those increases permanent, removing the sunset that had left the account system’s generosity hostage to future legislative bargaining. The Act also created a detailed safe harbor for qualified automatic contribution arrangements, specifying the default deferral rates, the required escalation schedule, and the minimum employer contributions that would deem a plan to satisfy nondiscrimination testing. An employer that adopted the prescribed design bought not only legal safety for automatic enrollment but also freedom from the annual testing regime that had been one of the 401(k)’s remaining administrative burdens. Each of these provisions, taken alone, was a technical refinement. Taken together, they showed Congress acting as the account system’s attentive engineer, tuning its rules for performance, while the pension received, in the same statute, a stricter funding regime and automatic benefit restrictions. The two designs were not merely diverging by accident of history. They were being steered.

What did the 2006 law change about how workers join 401(k) plans?

The Pension Protection Act of 2006 gave employers legal cover to enroll workers automatically, preempted state wage laws that blocked the practice, and created a safe harbor for investing defaulted contributions in diversified funds. Participation rose because inertia then worked for saving instead of against it, and automatic enrollment became the standard design.

The behavioral logic behind the 2006 reforms deserves emphasis because it reveals how thoroughly the account system had been reconceived. The original 401(k) was a libertarian instrument: it offered a tax-favored choice and left the choosing to the worker. The post-2006 401(k) is a paternalistic instrument: it makes the favored choice the default and leaves the opting out to the worker. Nothing in the statute compels any worker to save; the opt-out preserves formal freedom. But the design acknowledges, as the pension’s designers had acknowledged a century earlier, that most people will accept whatever arrangement requires no action. The difference is that the pension’s default was a professionally managed promise, while the 401(k)’s default is a professionally selected investment menu inside a worker-owned account. The law did not return risk to the employer. It kept risk with the worker and hired better defaults to manage it.

The 2019 and 2022 Acts: Automatic Enrollment Becomes Mandatory

The next two enactments completed the transformation that the 2006 law had begun, moving automatic enrollment from a protected option to a legal requirement and extending the account system’s reach into corners of the workforce it had never served. The Setting Every Community Up for Retirement Enhancement Act of 2019, known universally as the SECURE Act and enacted as part of the appropriations package signed December 20, 2019, as Public Law 116-94, made several structural changes. It raised the age at which required minimum distributions must begin from seventy and one half to seventy-two, acknowledging longer working lives. It created pooled employer plans, allowing unrelated small employers to band together in a single plan and share administrative costs, which lowered the barrier for the smallest businesses. It required plans to provide participants with lifetime income illustrations, translating account balances into estimated monthly payments so that workers could see what their savings meant as retirement income. And it gave fiduciaries a safe harbor for selecting annuity providers, encouraging plans to offer lifetime income options inside the account. Each provision assumed the account as the setting and improved it on its own terms; none revisited the pension.

The follow-on legislation, enacted December 29, 2022, as Division T of the omnibus appropriations measure Public Law 117-328 and known as SECURE 2.0, went further. Its headline provision made automatic enrollment mandatory for 401(k) and 403(b) plans established after December 29, 2022. For plan years beginning after December 31, 2024, covered plans must automatically enroll eligible workers at a contribution rate between three and ten percent of pay, escalating by one percentage point per year until reaching at least ten percent, capped at fifteen percent. Workers retain the right to opt out, but the default is set by law rather than by employer choice. The mandate carries exceptions: businesses with ten or fewer employees, new businesses in their first three years, church plans, and governmental plans are excluded, among others. Alongside the mandate, the law raised the required minimum distribution age to seventy-three and scheduled a further rise to seventy-five, permitted employers to match workers’ student loan payments as if they were retirement contributions, and authorized emergency savings accounts linked to plans. The trajectory from 2006 to 2022 is unmistakable: Congress spent sixteen years perfecting the account, automating its enrollment, diversifying its defaults, extending its availability, and finally commanding its use, while the defined benefit plan received no comparable program of improvement in any of these statutes.

The Risk Transfer, Stated Without Euphemism

With the sequence complete, the transfer of risk can be stated precisely, because precision is what this subject most needs and what it least often receives. In a defined benefit plan, the employer bore three burdens. It bore investment risk: when markets fell, the sponsor contributed more. It bore longevity risk: when retirees lived longer than expected, the sponsor kept paying. And it bore the funding obligation itself, a legal duty to keep the plan’s assets on pace with its promises, enforceable through excise taxes and, in extremis, through the claims of the federal insurance program. Behind all three stood the Pension Benefit Guaranty Corporation, which guaranteed benefits up to a statutory maximum when sponsors failed. The worker’s corresponding position was passive but protected: contribute nothing directly in most designs, make no investment decisions, and receive a defined payment for life, subject to the plan’s survival and the insurance cap.

In the 401(k), every one of those burdens changes address. Investment risk belongs to the worker, who selects the funds and absorbs the losses. Longevity risk belongs to the worker, who must stretch a finite balance across an unknown lifespan or purchase an annuity at retail prices. The funding obligation disappears as a concept, because there is no promise to fund; the employer’s duty ends with the remittance of elective deferrals and any promised match. The insurance guarantee disappears with it: no federal program insures a 401(k) balance against market loss, and the Pension Benefit Guaranty Corporation has no jurisdiction over individual accounts. What the worker gains, set against these transferred risks, is ownership and portability. The account is the worker’s property from the start, reachable through loans and hardship withdrawals in ways pension benefits never were, and it travels with the worker across employers through rollovers, preserving the savings of the mobile worker whom the pension systematically shortchanged.

One more transferred cost deserves explicit treatment, because it is the one workers feel most directly and understand least: fees. In the pension, investment management and administrative expenses were paid from plan assets or by the sponsor, and their effect on any individual worker’s benefit was invisible, since the benefit was defined by formula rather than by account performance. In the 401(k), every layer of cost, the expense ratios of the mutual funds, the recordkeeper’s per-participant charges, the adviser’s fees where one is engaged, is borne by the account and therefore by the worker, deducted silently from returns year after year. A difference of one percentage point in annual fees, sustained over a forty-year career, compounds into a substantial reduction in the final balance, yet the fee structure of most plans was for years disclosed in documents few participants read. Later regulation improved fee disclosure, but the structural fact remains: the account system moved the cost of professional money management from the sponsor’s budget to the worker’s balance. Related to this is the question economists call the annuity puzzle: given that pensions once provided lifetime income and that longevity risk is real, why do so few 401(k) holders convert their balances into annuities at retirement? Part of the answer is price, since retail annuities carry higher loads than the wholesale mortality pooling a pension achieved. Part is psychology, since handing a lifetime of savings to an insurer feels like a loss of control. And part is the tax code’s own doing, since required minimum distribution rules force withdrawals on a schedule that discourages annuitization. Whatever the mix of causes, the result is that the worker who bore longevity risk in theory often manages it in practice through cautious underspending, which is to say through a lower standard of living than the balance might have supported. The pension’s actuary once solved this problem for the retiree. The account leaves it unsolved.

The public-sector contrast sharpens the picture and supplies the brief’s key piece of evidence about causation. While private-sector coverage moved decisively toward accounts over four decades, state and local government employment remained predominantly pension-based. Public employers operate under different legal and budgetary conditions: they cannot easily shed benefit promises, they face political rather than market pressure on compensation, and their accounting for pension liabilities follows governmental standards rather than the corporate funding rules that the 1980s and 2006 legislation tightened. The persistence of pensions where the legal and budgetary environment preserved them, alongside their disappearance where the environment penalized them, suggests that the environment did much of the work. Employee preference alone cannot easily explain why workers doing similar jobs in the public and private sectors ended up in such different systems. The law’s differential treatment of the two designs can.

What Else Was Moving While the Law Moved

A careful account must foreshadow the complication that the second half of this article takes up in full, because the legislative story, though powerful, is not the whole story. Three other forces pushed in the same direction, and any honest weighing of causes must reckon with them.

First, the structure of employment changed. The pension’s golden age coincided with the peak of large-firm manufacturing employment, where long tenures with a single employer were the norm and the pension’s reward for loyalty matched the shape of working life. As employment shifted toward services, toward smaller firms, and toward more frequent job changes, the pension’s design fit fewer and fewer workers. A benefit that punished mobility became less valuable as mobility rose, independent of anything Congress did.

Second, workers themselves placed a rising value on portability and control. The portability argument holds that younger and more mobile workers often preferred the visible, portable account to the opaque, tenure-bound promise, even when the promise was actuarially more generous. The 401(k) gave workers something the pension never had: a number on a statement, growing visibly, owned outright. That psychological difference mattered for adoption, and employers designing benefits to attract workers had reason to offer what workers said they wanted.

Third, longevity rose. Life expectancy at age sixty-five rose across the pension era, which lengthened the payout period of every life annuity and raised the true cost of the pension promise year after year. The employer that had promised a worker retiring at sixty-five a check for life found that life growing longer with each passing decade, while the employer’s ability to fund the extension depended on market returns it could not control. Rising longevity made the pension more expensive to keep even if Congress had never touched the tax code.

These competing explanations do not refute the legislative account; they qualify it. Employers did not simply wake up one morning and abandon their workers, and Congress did not single-handedly engineer the outcome. The shift emerged from the interaction of a legal framework that raised the pension’s cost, an economy that reduced the pension’s fit, workers who valued the account’s portability, and a demography that made lifetime promises dearer. How much weight each cause deserves is genuinely contested, and partisans of each story can cite evidence. The second half of this article takes up that contest directly: it lays the competing explanations side by side, examines what the public-sector evidence and the timing of the shift imply about each, and reaches a defended verdict on which system serves which kind of worker in the world the statutes made.

Where Each Risk Actually Sits

The statutory sequence that produced the modern system can be stated in a single paragraph. The 1974 law attached vesting, minimum funding, fiduciary, reporting, and insurance obligations to the defined benefit promise, raising the fixed cost and legal exposure of maintaining a pension. The Revenue Act of 1978 created the account vehicle by permitting salary reduction contributions inside a profit-sharing plan, and guidance at the start of the 1980s confirmed the tax treatment. Later tax legislation tightened the pension rules further while leaving the account system light, the 2006 statute made automatic enrollment and diversified defaults legally safe, and the 2022 retirement act converted that safe harbor into a mandate for most newly established plans, a step traced in full in the companion account of the retirement acts and the mandatory enrollment step. What remains is to state, element by element, exactly what that sequence moved and to whom.

Investment risk moved first and most completely. In a defined benefit plan the sponsor promises a formula, and the promise does not flex with markets. When the trust’s assets earn less than the actuary assumed, the shortfall does not reduce any worker’s accrued benefit; it becomes the employer’s required contribution under the funding standards that the 1974 law imposed and later legislation tightened. The employer’s obligation runs to the formula, not to the portfolio. The account plan inverts the mechanism. Each contribution buys investments the participant selects from the plan’s menu, and the balance at retirement is whatever those investments produced, nothing more and nothing less. The employer owes nothing beyond the contributions it contracted to make, and a market decline reduces the worker’s balance with no legal consequence for the sponsor. This inversion is the single most consequential reallocation in the story, and it is also the one most often misdescribed as a corporate decision rather than as the predictable result of a tax provision that severed the contribution from the promise.

The statutes also shielded the account sponsor from the investment consequences in a way the pension sponsor could never claim. Under the Labor Department’s regulations implementing section 404(c) of the 1974 law, a sponsor that offers a broad menu of diversified investments and lets participants direct their own accounts is generally not liable as a fiduciary for the participants’ investment losses. The pension sponsor enjoys no such shield: the fiduciary standards apply to every investment decision the plan’s managers make, and imprudent choices are actionable. The contrast is the point. The law holds the pension sponsor to a fiduciary standard for managing the money and then holds the same sponsor to a funding standard for the shortfall; it holds the account sponsor to neither, provided the menu was adequate and the choice was the worker’s. Risk followed responsibility, and responsibility followed the statute.

Who bears investment risk in a 401(k)?

The worker bears it, entirely. Every contribution sits in an account titled to the participant, and the balance moves with the performance of the investments the participant selects. The employer owes no make-whole payment if markets fall. That allocation was the point of the 1978 design: the employer funds no promise, so the employer carries no investment consequence.

Longevity risk, the risk of outliving one’s savings, traveled the same route. A defined benefit plan pays for life. The sponsor’s actuary estimates how long participants will live, the sponsor funds accordingly, and when participants live longer than the estimate, the sponsor pays longer. Improvements in life expectancy, welcome everywhere else in economic life, are a cost increase inside a pension plan, and that cost sits entirely with the employer. The account plan contains no such mechanism. The balance is finite, the lifespan is not, and reconciling the two is the worker’s task. A participant who reaches ninety-five with an account sized for eighty-five faces the arithmetic directly, with no sponsor obligated to bridge the gap. The worker can purchase a commercial annuity to recreate the pooling, but the purchase is voluntary, priced by an insurer, and chosen from a market the worker must navigate without the employer’s actuary.

The commercial annuity market, which the account holder must use to recreate the pension’s pooling, suffers from a selection problem the pension never faced. Only people who expect to live long buy lifetime annuities voluntarily, so insurers price for long-lived buyers, which makes the product expensive for everyone else, which drives more buyers away. The pension plan avoided this spiral because participation was automatic: every worker was in the pool, the short-lived subsidized the long-lived unknowingly, and the pricing reflected the whole population. When the statutes moved workers from the automatic pool to the voluntary market, they moved them from the one setting where longevity insurance prices fairly to the one setting where it prices punitively. This is a genuine efficiency loss of the transfer, and it belongs in any honest ledger of what the shift cost.

The funding obligation is the legal spine of the pension promise, and it has no counterpart in the account world. The 1974 law required sponsors to fund their promises on a schedule, disclose the funding position, and answer to fiduciaries and auditors; the 2006 law accelerated the timetable and tied required contributions more tightly to the plan’s funded status, so that a falling market or a low interest rate environment could trigger large mandatory cash calls in the same year the sponsor’s business was under stress. This procyclical bite deserves attention, because it explains behavior that looks like abandonment and was in fact compliance. A sponsor that froze a plan during a downturn was often responding to a funding requirement that arrived at the worst possible moment, a requirement the law imposed and the business cycle made painful. The account plan carries no such obligation. Employer contributions, including the familiar match, are voluntary and contractual; a sponsor can reduce or suspend them, and many did in recessions, with no statute violated and no penalty assessed. The asymmetry is structural: the law demands funding only where a promise has been made, and the account system was designed to make no promise.

The premium structure reinforced the same logic. The 1974 law funded its new insurer with premiums levied on sponsors, and later Congress repeatedly raised both the flat per-participant premium and the variable premium tied to underfunding. Each increase was enacted to shore up the insurer’s finances, and each increase functioned as a tax on maintaining a pension that had no counterpart in the account system, where no insurer exists and no premium is owed. A sponsor weighing whether to keep a defined benefit plan thus faced a recurring charge for the privilege of carrying the very risks the law had assigned to it, a charge that grew precisely when plans were most stressed. The procyclical trap deserves a final emphasis because it is the mechanism most often mistaken for a change of heart. Funding rules measure liabilities with interest rates, so when rates fall, the present value of promised benefits rises even though nothing about the workforce has changed; the sponsor must then contribute more at the moment credit is tightest. The 2006 law’s tighter link between funded status and required contributions made this bite sharper and more immediate. Sponsors that froze plans in the years after its enactment were not discovering a new dislike of pensions; they were discovering that the funding formula could demand enormous contributions in a recession year, and the account alternative carried no such demand at all.

Portability is the one element of the transfer that moved in the worker’s favor, and its importance is easy to understate. A pension benefit is attached to the employer. Vesting schedules mean that a worker who leaves early forfeits the unvested portion outright, and even vested benefits typically convert to a deferred annuity frozen at the salary and service levels of the departure date. The final-average-pay formulas that dominated pension design then compound the penalty: each job change restarts the clock on the salary measure that determines the benefit, so the mobile worker’s eventual pension is a fraction of what continuous service would have produced. The account has no employer attachment. The balance is titled to the participant, contributions are the worker’s property from the start in most designs, and the tax code permits rollovers into a new employer’s plan or an individual retirement account without triggering tax. For a workforce that changes jobs repeatedly across a career, this is not a minor convenience; it is the difference between an asset that accumulates and a promise that evaporates with each departure.

What happens to a pension when you change jobs?

Leaving usually strands the benefit. Unvested accruals are forfeited, and vested accruals convert to a deferred annuity payable at retirement age, frozen at the salary and service of the departure date. Some plans offered small lump sums, but the final-pay formula rewarded stayers, so each move typically shrank the eventual payment.

Portability’s virtue has a documented vice, and an honest account names it. The same tax code that lets the account travel also lets the worker cash it out at each job change, paying income tax and, for most workers under fifty-nine and a half, an additional ten percent penalty. A substantial share of job-changers do exactly that, particularly younger and lower-paid workers for whom the balance looks like spending money rather than retirement money. Each cash-out resets the compounding clock to zero, and the cumulative effect across a mobile career can erase most of the portability advantage the account was supposed to confer. The pension never offered this particular failure mode, because the deferred annuity could not be raided at twenty-eight. Later legislation tried to dam the leak: automatic rollover rules for small balances, and eventually provisions encouraging emergency savings features so that workers would stop treating the retirement account as the emergency fund. The leakage evidence complicates the portability story without overturning it. The account remains the better instrument for the mobile worker, but only for the mobile worker who leaves the money alone, and the statutes that created the account spent decades learning that portability without preservation is an incomplete design.

The vesting rules, meanwhile, tightened in the worker’s favor across the same period, which is one place the ratchet ran the other direction. The 1974 law permitted schedules as slow as ten-year cliff vesting; the Tax Reform Act of 1986 cut the permissible maximum to five-year cliff or three-to-seven-year graded vesting. Each shortening reduced the forfeiture penalty for the mobile worker without changing the deeper arithmetic: even fully vested, the departing worker’s benefit froze at the old salary, so faster vesting helped at the margins while the final-pay formula did the real damage to movers. The vesting history is a useful corrective to any story that paints the statutes as uniformly hostile to workers. Congress repeatedly intervened to protect the worker inside the pension; it simply intervened more consequentially, and more often, to raise the price of the pension itself.

Insurance backing completes the risk picture and corrects the most common misunderstanding in the public argument. Defined benefit plans carry federal insurance through the Pension Benefit Guaranty Corporation, the agency the 1974 law created under Title IV and funded through premiums paid by plan sponsors. If a sponsor fails with an underfunded plan, the agency steps in and pays benefits up to statutory caps; the guarantee is real but limited, and it is paid for by the community of sponsors rather than by general taxpayers. Account plans have no equivalent. No federal agency insures a 401(k) balance against market loss, and no statute extends the pension insurer’s guarantee to defined contribution accounts. The Securities Investor Protection Corporation, which some readers confuse with pension insurance, protects against brokerage failure rather than investment decline, and its coverage has no bearing on a falling market. The worker who assumes the account carries the same backstop as the pension is mistaken about the architecture.

The guarantee’s limits matter as much as its existence. The single-employer program caps the monthly benefit it will pay, with the cap set by statute and adjusted over time, so highly compensated workers with large accrued benefits can lose a portion of the promise even when the insurer performs exactly as designed. The multiemployer program, which insures the joint union-employer plans, historically carried far lower guarantee levels and operated on thinner finances, a disparity that became the central fact of the multiemployer crisis discussed below. Both programs are funded by sponsor premiums rather than general revenue, which means the insurance is ultimately a mutual arrangement among employers rather than a taxpayer backstop, and which also means that each premium increase Congress enacted to rescue the insurer raised the cost of the insured activity. The account system, by contrast, pays no premiums because it insures nothing; the absence of a guarantee is also the absence of a bill. Readers comparing the two systems should therefore compare complete bundles: the pension offers a capped, premium-funded insurance against sponsor failure, and the account offers no insurance at a price of zero, and the 1974 law is the reason the first bundle exists and the second does not.

The death benefit rules reveal the same philosophy of assigned responsibility in miniature. The pension promise carries spousal protections written into the Retirement Equity Act of 1984: a married participant’s benefit must be paid as a joint and survivor annuity unless the spouse consents in writing to another form, and a preretirement survivor annuity protects the spouse if the participant dies before benefits begin. The law treats the pension as household property and restrains the participant from disposing of it unilaterally. The account is individual property with a beneficiary designation. A married participant’s spouse is the presumed beneficiary under the same 1984 framework, but the designation is otherwise the participant’s to make, change, and neglect, and an outdated form can send the entire account to a former spouse or into a deceased parent’s estate with no plan administrator empowered to override the paperwork. The pension’s paternalism and the account’s autonomy are two answers to the same question of who should be trusted with the final disposition, and the statutes answered it differently for the two systems.

Divorce exposes another layer of the pension’s legal weight. A court dividing marital property can assign part of a pension benefit to a former spouse through a qualified domestic relations order, a creature of the 1984 amendments, and the plan administrator must then administer two payees, two survivor elections, and a thicket of timing rules for a single benefit. The account divides with a form and a transfer. Family-law practitioners consistently report that pension orders generate more contention per dollar than account divisions, which is another quiet tax the pension’s legal complexity levied on everyone who touched it. None of this made the pension worse for the worker whose marriage survived; it made the promise more expensive for the system that administered it, and the expense, like the funding obligation and the premiums, had no counterpart in the account world.

The Risk Transfer Table

The six elements above reduce to a single comparison. Each row names the risk or feature, identifies who holds it under each system, and names the statute that assigned it. Read across each row to see the transfer; read down the final column to see how few statutes did the work. Five legislative moments, the 1974 framework, the 1978 tax provision, the 1984 spousal-protection amendments, the 2006 funding and default rules, and the rollover provisions of the tax code, account for nearly every cell.

Risk or feature	Defined benefit pension (who holds it)	401(k) account plan (who holds it)	Statute that assigned it
Investment risk	Employer: sponsor must cover asset shortfalls through mandatory contributions	Worker: balance rises and falls with the performance of chosen investments	ERISA, Public Law 93-406, funding and fiduciary rules; IRC section 401(k), added by the Revenue Act of 1978, Public Law 95-600
Longevity risk	Employer: sponsor funds lifetime payments for as long as participants live	Worker: must make a finite balance last; exhaustion is the worker’s risk	ERISA benefit accrual and annuity rules; IRC section 401(k) account structure
Funding obligation	Employer: legally required minimum contributions on a statutory schedule	No employer obligation: contributions are voluntary or contractual only	ERISA funding standards, tightened by the Pension Protection Act of 2006, Public Law 109-280; no parallel funding duty exists for account plans
Portability	Limited: unvested accruals forfeited on departure; vested benefits typically frozen as deferred annuities	Full: balance travels with the worker and rolls over tax free between plans	ERISA vesting schedules; IRC section 402(c) rollover provisions
Insurance backing	Federal insurance through the PBGC, subject to statutory benefit caps	None: no federal insurance of account balances against loss	ERISA Title IV created the PBGC; no statute extends its guarantee to defined contribution accounts
Death benefit	Spousal protections: joint and survivor annuity defaults with written spousal consent required to waive	Beneficiary designation: account passes to the named beneficiary as individual property	Retirement Equity Act of 1984; IRC beneficiary designation rules

The Coverage Divergence

The four decades after the 1978 provision produced one of the cleanest natural experiments in American economic law. In the private sector, the account plan displaced the pension with a speed that surprised even its architects. In public employment, the pension endured. The two workforces did not differ in their humanity, their thrift, or their desire for security in old age. They differed in the legal environment their employers inhabited, and the divergence is the strongest single piece of evidence about what caused the shift.

The private-sector movement is visible in federal benefits data. Bureau of Labor Statistics figures for 1979 showed 87 percent of full-time workers at medium and large private firms participating in a retirement plan, typically a defined benefit plan. By March 2023 the picture had inverted: 49 percent of private industry workers participated in defined contribution plans versus 11 percent in defined benefit plans, and only 15 percent of private industry workers had access to a defined benefit plan at all. The transition did not proceed as a wave of terminations. It proceeded as a wave of freezes and closures to new hires: sponsors kept the promises they had made to existing workers, because the law gave them little choice about accrued benefits, and directed all new accruals into accounts. A freeze is the precise footprint of a cost-and-liability calculation. It preserves the past, which the vesting and anti-cutback rules protect, and it stops the future, which the funding and premium rules had made expensive. Employers that are often described as having abandoned their pensions in fact did something narrower and more legible: they stopped writing new pension promises under a statute that had raised the price of each new promise while the tax code offered a cheaper substitute.

Public employment tells the other half of the story. State and local government workers remained predominantly pension-based across the same four decades. In the National Compensation Survey’s state and local vintages, defined benefit access remained the norm while defined contribution access stayed the supplement rather than the center. Federal civilian workers sit in a hybrid system built by the Federal Employees’ Retirement System Act of 1986, which pairs a modest defined benefit annuity with Social Security and a defined contribution Thrift Savings Plan, but the state and local picture is unambiguous: where the pension was the established form, it stayed the established form.

The inference from the divergence is difficult to avoid and worth stating plainly. Governmental plans are exempt from the Employee Retirement Income Security Act’s funding, fiduciary, and insurance regime; the 1974 law’s cost-and-liability machinery never applied to them in the first place. Public sponsors also face a different political economy of compensation: a pension promise defers costs into future budgets, which elected officials often prefer to current salary increases, and many state constitutions protect accrued public pension benefits against impairment, which makes the promise credible in a way a private sponsor’s cannot always be. None of this means public workers prefer pensions more than private workers do. It means their employers operated under statutes and incentives that preserved the defined benefit form, while private employers operated under statutes that penalized it. If the shift had been driven mainly by a change in what workers wanted, or by a spontaneous corporate decision to abandon a valued institution, the public sector would have moved too. It did not, and the reason it did not is written in the jurisdictional limits of the statutes themselves. The pattern belongs to the wider arc of American workplace law since midcentury, in which the same workforce experiences very different rules depending on which side of the public-private line the employer stands.

The divergence also clarifies a failure both systems share: the small employer. Account plans never reached the smallest firms at anything like the rate they reached large ones, because sponsorship costs, fiduciary exposure, and administrative burden deterred adoption even without a pension’s funding obligation. The result was a persistent coverage gap in which workers at small businesses had neither a pension nor an account, a gap the statutes addressed only belatedly. The 2019 retirement act authorized pooled employer plans that let unrelated small firms join a single arrangement, and beginning in the late 2010s several states built automatic individual retirement account programs that swept uncovered workers into state-facilitated savings. Both were admissions that the voluntary account system, left to itself, did not reach everyone the pension system had once covered through union and industry-wide plans. The gap is worth noting because it rebuts the comforting version of the account story, in which every worker simply got a portable balance instead of a pension. Many workers got neither, and the law is still, decades later, building the scaffolding to reach them.

The multiemployer corner of the pension world supplies a confirming data point, and it deserves careful handling because it is the one place where Congress intervened to rescue pensions rather than to burden them. Multiemployer plans, the joint union-employer plans common in construction, trucking, and food retail, were insured by a separate Pension Benefit Guaranty Corporation program with far lower guarantee levels and far weaker funding than the single-employer program. By the late 2010s, a cohort of these plans faced insolvency within years, and the insolvency of the plans threatened the insolvency of the insurance program itself. Congress responded in the American Rescue Plan Act of 2021, Public Law 117-2, signed March 11, 2021, by creating a Special Financial Assistance program under section 9704 that authorized the Pension Benefit Guaranty Corporation to pay grants, not loans, to eligible financially troubled multiemployer defined benefit plans, with funding sufficient for plans to pay benefits through 2051. The episode cuts two ways, and both cuts support the same reading. On one side, it demonstrates that pension costs can grow beyond what contribution bases can bear, which is the employer-cost argument in its starkest form. On the other side, it demonstrates that when the pension system faced collapse, the legislature treated the defined benefit promise as worth preserving with public money, while no parallel rescue has ever been extended to defined contribution accounts, because there is no promise in an account to rescue. The law’s solicitude and the law’s burden both attach to the promise, and the account sits outside both.

There is a further wrinkle that sophisticated readers will raise, and it should be addressed rather than waved away. Some sponsors did not move from pensions to pure accounts; they moved to cash balance plans, a defined benefit design that expresses the benefit as a hypothetical account with a guaranteed crediting rate. The cash balance form kept the employer’s investment risk and the federal insurance while mimicking the account’s portability and transparency, and its spread in the 1990s suggested that sponsors valued the pension’s pooling enough to preserve it in disguise. The conversions then collided with age-discrimination litigation, most prominently the dispute over International Business Machines’ 1999 conversion, which produced years of litigation before Congress clarified the treatment of cash balance designs in the 2006 statute. The episode is instructive precisely because it failed to become the dominant form: even the hybrid that split the difference between the two systems could not escape the cost-and-liability logic, and sponsors that wanted out of the pension’s legal exposure found the clean account simpler than the defended hybrid.

A second channel amplified the statutes’ effect, and it ran through accounting rather than legislation. In 2006 the Financial Accounting Standards Board required sponsors to recognize the funded status of their pension plans directly on the corporate balance sheet, where previously much of the obligation had lived in footnotes. A plan that had been an abstract long-term commitment became a visible liability that moved with interest rates and equity markets, and chief financial officers who had tolerated the pension as a footnote began managing it as a balance-sheet risk. The timing compounded the 2006 funding statute: the same year that tightened the cash demands also forced the liability into daylight, and the combination converted the pension from a human-resources institution into a treasury problem. Freeze announcements clustered in the years that followed, and the clustering is more consistent with a financial-reporting shock layered onto a funding shock than with any gradual change in philosophy. None of this was legislated in the retirement statutes themselves, which is why it belongs in the complication rather than the sequence; but the channel only existed because the 1974 law had created the funded promise in the first place, and the account system, carrying no promise, generated no liability to recognize.

The Competing Explanations

Two simple stories circulate about the pension’s disappearance, and the brief for this article rejects both of them. The first says employers abandoned workers out of greed. The second says legislation did it all. The evidence supports a more textured account in which the statutes did the heaviest work but did not work alone, and the competing explanations deserve to be stated at full strength before the weighing begins.

The manufacturing-to-services explanation starts from the composition of the workforce. Defined benefit pensions were concentrated in large, unionized manufacturing firms, the sector where long tenures, stable employers, and collective bargaining made the pension promise natural. Over the four decades in question, American employment shifted decisively toward services, toward smaller firms, and toward non-union workplaces, none of which had ever adopted pensions at manufacturing’s rates. On this account, the pension did not so much die as lose its habitat: the firms that sponsored pensions shrank as a share of employment, and the firms that grew in their place never had pensions to lose. The correlation is real. Union density and defined benefit coverage declined on similar schedules, and the service-sector firms that came to dominate employment overwhelmingly chose accounts from their founding. The explanation’s weakness is that it cannot account for what happened inside continuing firms. The freezes that drove the coverage statistics were decisions by existing sponsors, many of them large and long-lived, to stop new pension accruals. A sectoral shift explains why new firms chose accounts; it does not explain why old firms stopped pensions. Something changed the calculus for sponsors that had been comfortable with the promise for decades, and the habitat story has no candidate for that something except the changing legal price of the promise itself.

The portability explanation starts from the worker’s side of the ledger and carries genuine force. Median job tenure in the American economy is measured in years, not decades, and each change of employer under a traditional pension meant forfeited accruals and frozen benefits. For the mobile worker, the account is not a consolation prize; it is the superior instrument, accumulating across employers while the pension would have evaporated at each departure. Worker advocates and survey researchers have long reported that younger workers value the account’s transparency and control, and the revealed preference of the labor market is suggestive: quit rates rose across the period, and the pension’s back-loaded accrual pattern, which concentrates value in the final years of a long tenure, looks increasingly like a design for a workforce that no longer exists. The explanation’s limit is that worker preference cannot easily explain the public-private divergence. Public-sector workers change jobs too, and value portability too, yet their employers kept the pension. If portability demand were the engine of the shift, it should have operated across sectors. It operated where the law permitted the alternative and stalled where the law preserved the original.

The longevity explanation is the most technically respectable of the three and deserves the most careful statement. Life expectancy at older ages rose substantially across the period, and every additional year of expected life raises the present value of a lifetime annuity promise. A pension priced for the mortality of 1974 became materially more expensive to fund for the mortality of 2004, through no fault of the sponsor and no act of the legislature. Actuaries who study plan costs attribute a meaningful share of the funding burden’s growth to this demographic fact, and it is the one cost driver that would have pressed on pensions even in a world where Congress never touched the statutes. Its limit is timing. Longevity rose gradually, on a smooth curve, while pension freezes clustered in waves that followed statutory and regulatory shocks: the funding tightening of the 2006 law, the accounting changes that forced sponsors to recognize funding positions on the balance sheet, and the market collapses that turned funding requirements procyclical. A smooth cause does not readily explain a lumpy effect. Longevity raised the water level; the statutes built the waves.

The empirical literature has tried to decompose these forces, and its results are worth reporting with their uncertainties intact. Studies that compare sponsors across industries find that funding-rule tightenings predict freezes even after controlling for sector and union status, which favors the statutory account; studies that compare coverage across cohorts find that younger workers sort into firms that never offered pensions, which favors the compositional account. Neither literature has produced a clean decomposition, partly because the forces interact: the statutes raised the cost of the promise, the cost fell hardest on the unionized manufacturers already under competitive pressure, and the resulting freezes accelerated the sectoral shift they are sometimes treated as an alternative to. Interaction is not the same as exoneration. A cause that operates through other causes is still a cause, and the question is which factor a policymaker could have changed to produce a different outcome. Changing the funding and premium rules would have changed sponsor behavior directly; changing the sectoral composition of the economy was never on the table. That asymmetry in policy leverage is one reason the legislated explanation carries the preponderance even before the timing and jurisdiction evidence is weighed.

Weighing the four accounts together, the evidence supports the legislated explanation most strongly, with the other three accounting for real but secondary shares of the outcome. The decisive considerations are timing, jurisdiction, and the footprint of the freeze. Timing favors the statutes: coverage fell fastest where and when the legal price of the promise rose, not on the smooth schedule of demographic change. Jurisdiction favors the statutes: the public sector, exempt from the 1974 law’s machinery, kept its pensions while the covered private sector shed them. The footprint favors the statutes: sponsors froze future accruals while honoring past ones, exactly the behavior the vesting, anti-cutback, funding, and premium rules would predict, and not the behavior a pure change in worker taste would produce. The competing explanations are not wrong; they are incomplete. The sectoral shift removed the pension’s natural habitat, longevity raised the annuity’s true cost, and portability made the account genuinely attractive to mobile workers. But none of the three explains why sponsors that had maintained pensions for decades stopped writing new ones, and all three are consistent with a world in which the statutes set the terms and the economy filled in the details. The causal weighting is contested among economists and policy analysts, and this article does not pretend the contest is settled; it reports that the legislated account carries the preponderance of the evidence, with the demographic and sectoral accounts carrying the remainder.

A final note on tone is owed, because the subject invites moral language and the evidence does not support it. Employer groups argued, with documentation, that the funding and premium rules made the pension promise unmanageable in competitive industries, and the freeze waves that followed market downturns corroborate the claim. Worker advocates argued, with equal documentation, that the account system left millions of workers with inadequate balances, no longevity pooling, and no insurance, and the retirement-security literature corroborates that claim too. Both arguments can be true, because they describe different sides of the same transfer. The statutes moved risk from institutions that could pool it to individuals who cannot, and whether that movement counts as efficiency or abandonment depends on which side of the transfer one stands. The law is neutral on the moral question. It is not neutral on the causal one.

The Verdict

A comparison that ends in a shrug wastes the reader’s time, so this one ends in a verdict with a named deciding factor. Call it the tenure test: the worker’s expected years with a single employer determine which system serves that worker, because the two systems are optimized for opposite career shapes. The pension rewards duration; the account rewards motion. Everything else in the comparison, the insurance, the investment control, the death benefit rules, is commentary on that central opposition.

For the long-tenure, single-employer worker, the pension wins outright. The back-loaded accrual pattern that punishes the job-changer is a subsidy to the stayer: the final-average-pay formula concentrates the plan’s value in the last years of a long career, delivering a replacement rate the account system struggles to match without heroic contribution levels. The worker bears no investment risk, makes no asset-allocation decisions, and cannot outlive the benefit. The employer’s funding obligation, described earlier as a burden on sponsors, is from this worker’s perspective the entire point: a legally enforceable promise, backed by a federal insurer, that converts decades of service into a lifetime income. The worker who spends thirty years with one employer and retires under a traditional pension has received the best retirement deal the private sector ever offered, and the account system has never replicated it for that career shape.

For the mobile worker, the account wins by an equally clear margin. The worker who changes employers every four or five years would have accumulated a pension record of forfeited accruals and frozen deferred annuities, each too small to matter and each payable at some distant retirement age. The same career in the account system produces a single growing balance that compounds across employers, visible at every login and available for rollover at every departure. The mobile worker gives up longevity pooling and the insurance backstop, and those are real losses, but the alternative was never a generous pension; the alternative was a scattering of pension fragments. Portability is not a consolation feature for this worker. It is the feature that makes retirement saving possible at all.

Which is better for a worker who changes jobs often: pension or 401(k)?

The 401(k), by a wide margin. Frequent moves forfeit unvested pension accruals and freeze vested ones at old salaries, while the account travels with the worker and compounds across employers. The mobile worker trades the pension’s longevity pooling for portability, and the arithmetic of five or six job changes makes that trade favorable.

For the worker who needs a guaranteed floor, the pension wins, and this is the verdict’s hardest edge. The account system has no answer to the problem of outliving one’s savings except the commercial annuity market, which most participants never enter, or continued work, which health may not permit. The pension’s lifetime annuity is the only private mechanism that pools longevity risk across a population, and its disappearance from the private sector removed the one instrument that guaranteed a worker could not outlive the benefit. Beneath both systems sits the public floor that Social Security provides, and for lower-earning workers that floor does more retirement-security work than either employer plan; but the floor was designed as a base, not a replacement, and the workers perched just above it, with earnings too high for the floor to suffice and savings too low to self-insure longevity, are the ones the shift harmed most. An honest verdict names them: the moderately paid, intermittently employed worker who needed a guaranteed annuity and received instead an account balance and a brochure about investment choice.

A stylized comparison makes the verdict concrete, with the caveat that the numbers are illustrative rather than drawn from any single plan. Imagine two workers who each earn the same wage path over a thirty-five-year career. The first spends all thirty-five years with one employer under a traditional final-average-pay pension accruing one and a half percent of final pay per year of service; at retirement the formula yields an annuity of roughly half of final pay, payable for life, with spousal protections and federal insurance behind it. The second changes employers seven times and participates in account plans throughout, contributing a combined ten percent of pay each year between personal deferrals and employer matches, earning a steady market return; at retirement the account holds a lump sum that must be converted, by the worker’s own decisions, into an income stream of uncertain duration. If the first worker’s formula delivers about half of final pay for life and the second worker’s account must fund thirty years of withdrawals at a prudent rate, the comparison turns on two questions the statutes answered differently: who bears the market risk along the way, and who bears the longevity risk at the end. For the stayer, the employer’s answers were better. For the mover, the account’s portability was the precondition of having anything at all. The illustration is deliberately simple, and real plans and real markets complicate every term, but the shape of the tradeoff survives the complications: duration favors the promise, motion favors the account, and the need for a guaranteed floor favors the promise for everyone.

The tenure test also resolves the internet’s most argued formulation of the question, whether the pension was better than the 401(k). The question has no general answer because the systems do not compete on the same terms. Asked of the thirty-year stayer, the pension was better. Asked of the six-employer careerist, the account is better. Asked of the worker who needs longevity insurance, the pension was better and nothing has replaced it. A verdict that names the deciding factor is more useful than a ranking, because it tells each reader which system was built for a career like theirs, and the statutes built the two systems for very different careers indeed.

The verdict’s final complication is that the legislature has spent the last two decades trying to put the pension’s virtues back into the account. The 2006 statute’s safe harbor for automatic enrollment and diversified defaults was the first retrofit: if workers would not choose to save, the law would choose for them, while preserving the formal voluntarism of the account. The 2019 act required plans to show participants a lifetime-income illustration, translating the lump sum into a monthly annuity equivalent on the benefit statement, an explicit attempt to make the account feel like the pension it replaced. The 2022 act pushed further, facilitating annuity options inside plans and expanding the provisions that let workers convert balances into guaranteed income. Each retrofit concedes, without quite saying so, that the account system is incomplete on the terms the pension once satisfied: it needs defaults to replace the compulsion, illustrations to replace the formula’s clarity, and annuity options to replace the pooling. Whether retrofits can fully substitute for the original is the live policy question, and the tenure test suggests a partial answer. Defaults and illustrations serve every worker; only genuine pooling serves the worker who needs the guaranteed floor, and pooling is the one feature the account resists by design.

How Different Readers Should Use This

The student meeting this material for the first time should read the risk transfer as a worked example of a general principle: statutes allocate risk by defining obligations, and every obligation has a price that shapes behavior. The pension story is the cleanest illustration available of how a legislature can transform an economy without ever ordering anyone to do anything, simply by making one form of promise expensive and offering a cheaper form beside it. Learn the six rows of the table as vocabulary, because the same vocabulary, investment risk, longevity risk, funding obligation, portability, insurance, and survivor protection, recurs in every retirement debate that follows.

The staffer drafting or amending retirement legislation should read the sequence as a caution about second-order effects. Every provision in the chain was defensible on its own terms: protecting vested benefits, clarifying tax treatment, encouraging saving through defaults. The cumulative effect was a system transformation that no single Congress voted for and no single Congress can easily reverse, because each step created constituencies, the account industry, the rollover market, the default-investment complex, that defend the result. Legislation that moves risk should be scored for where the risk lands, not only for the behavior it encourages in the year of enactment.

The researcher should treat the public-private divergence as the identification strategy the literature has been looking for. The two sectors shared a workforce, a macroeconomy, and a demographic curve; they differed in the statutes that governed their retirement promises. That is as close to a controlled comparison as economic history provides, and it deserves more formal exploitation than it has received. The open questions are quantitative: how much of the coverage decline each statutory change explains, how the freeze waves map onto funding-rule tightenings, and what the counterfactual coverage path looks like under alternative premium and funding designs.

The journalist covering the next round of retirement legislation should carry two facts into every interview. First, no federal insurance backs any 401(k) balance, a fact that surprises most readers and corrects most commentary. Second, the mandatory automatic enrollment provisions that arrived in the 2022 act complete a forty-year arc from voluntary alternative to default expectation, which means the account system is no longer the market’s spontaneous choice but the law’s designated successor. Sources on both sides will offer moral narratives; the statutes offer a causal one, and the causal one is the story.

All four readers should keep one structural fact in view that this article has repeated in several forms: the account system was never designed to do what the pension did. It was designed as a tax-favored savings vehicle inside profit-sharing plans, and it grew into the center of American retirement provision by accretion rather than by design. Every debate about its adequacy, every proposal to add automatic features, mandates, or guarantees, is an attempt to retrofit onto the account the protections the pension once provided by statute. Knowing which protections were legislated away, and which were never legislated at all, is the precondition for judging whether the retrofits can succeed.

Closing Assessment

The disappearance of the traditional pension reads, in retrospect, like a single event, a corporate decision repeated across thousands of boardrooms. The legal history shows it was not one decision but a sequence, and not a corporate sequence but a legislative one, in which each new statute altered the relative price of the two promises. The cost-and-liability ratchet: every major retirement statute since 1974 raised the cost or legal exposure of maintaining a defined benefit plan while lowering it for account plans, and the shift readers describe as a corporate choice is better read as a legislated one. That sentence is the article’s answer to its own One Test: a reader who can name the changes in order, state which risks moved and to whom, and defend a verdict by worker type has understood the subject as a legal history rather than as a grievance. The pension did not vanish because employers stopped caring about retirement security or because workers stopped wanting it. It vanished because the law, step by deliberate step, made the promise expensive to keep and the account cheap to offer, and rational actors on both sides of the employment relationship responded exactly as the incentives predicted.

The lesson generalizes beyond retirement. Any statutory scheme that prices one organizational form and subsidizes another will, given time, produce a migration it never explicitly ordered, and the migration will then be misread as the spontaneous choice of the migrating parties. The pension’s disappearance is the clearest American example of the pattern, which is why it rewards the close reading this article has attempted: follow the obligations, price the promises, and the economy’s slow structural changes resolve into sequences of provisions with dates, votes, and public law numbers.

That method is the article’s final claim about how to read economic change. A comparison resolved with a verdict and a named deciding factor does more work than a lament or a celebration, because it converts a grievance into a diagnosis: the tenure test tells a young worker, a mid-career switcher, and a near-retiree each something different and each something actionable. And a slow structural change traced to a specific sequence of statutory provisions does more work than either of the simple stories, because it identifies the levers. If the shift was legislated, it can be re-legislated, and the retrofits written into the 2006, 2019, and 2022 statutes, defaults, illustrations, annuity options, are the legislature’s tacit admission that the original transfer left something essential behind. Whether the retrofits finish the job is the question the next decade of retirement law will answer.

Study the Sequence

Readers comparing retirement systems, whether for a course, an exam, or a benefits decision, do best with a repeatable study path: work through the statutory sequence in order, keep the six rows of the risk transfer table at hand while reading, and test yourself against the verdict’s tenure test by arguing both sides. keep your statute notes, citations, and case chronologies together free on VaultBook for the note-taking, and practice and revise US government and civics material on ReportMedic to check whether the sequence holds together under questioning.

Frequently Asked Questions

Q: Why did pensions disappear and 401(k) plans take over?

The sequence runs through four statutes rather than one corporate decision. ERISA of 1974, Public Law 93-406, imposed vesting, funding, reporting, fiduciary, and insurance obligations on defined benefit plans, which raised the fixed cost and legal exposure of keeping one. The Revenue Act of 1978 added section 401(k) to the Internal Revenue Code, permitting salary reduction contributions, and guidance from tax authorities early in the 1980s confirmed the treatment, giving employers a cheap and legally simpler alternative. Later tax legislation tightened defined benefit rules further while the account system grew. The Pension Protection Act of 2006 made automatic enrollment and diversified default investments legally safe, and the 2022 retirement act made automatic enrollment mandatory for most newly established plans. Structural forces also mattered: employment moved from manufacturing toward services, workers valued portability across jobs, and rising longevity made lifetime promises more expensive. The causal weighting is contested, but the strongest supported reading is a cost-and-liability ratchet in which each major retirement statute raised the price of the pension while lowering it for the account plan.

Q: What law created the 401(k)?

The Revenue Act of 1978, Public Law 95-600, added section 401(k) to the Internal Revenue Code, the provision that permits employees to defer salary into a plan on a pre-tax basis. The provision was not aimed at retirement policy in the grand sense; it was a narrow salary reduction rule tucked into a tax bill. Guidance from tax authorities early in the 1980s confirmed that salary deferrals into profit-sharing plans received the intended treatment, and employers quickly adopted the design. A common error holds that ERISA of 1974 created the account plan. ERISA predated it by four years and regulated defined benefit plans rather than inventing the individual account. The Pension Protection Act of 2006 later entrenched the account system by making automatic enrollment and qualified default investments legally safe, and the 2022 retirement act extended the design further with an automatic enrollment mandate for most new plans. The 1978 revenue act remains the origin of the account plan itself.

Q: Who bears the risk in a pension versus a 401(k)?

The risk allocation is the decisive difference between the two systems. In a defined benefit pension, the employer bears the investment risk, the longevity risk, and the funding obligation: the benefit formula promises a fixed payment, and the sponsor must contribute enough to pay it even when markets fall or retirees live longer than projected. A federal insurance program, the Pension Benefit Guaranty Corporation, backstops the promise within statutory limits. In a 401(k), those risks move to the worker. The account holds whatever contributions plus market returns produce, and no one guarantees the balance or converts it into lifetime income automatically. The worker gains something the pension never offered, full portability from job to job, and pays for it with exposure to market downturns and to outliving the balance. The shift from one system to the other is therefore a transfer of investment risk and longevity risk from employer to employee.

Q: Is a pension better than a 401(k)?

Neither system wins in the abstract; the verdict turns on the worker’s career pattern. A pension suits the worker who stays with one employer for decades, because the benefit formula rewards long tenure and converts the result into a lifetime annuity that cannot be outlived. A 401(k) suits the worker who changes employers, because the account follows the worker, contributions keep accumulating across jobs, and the balance is fully portable. The pension’s weakness is its dependence on one sponsor’s survival and one worker’s longevity at a single firm; the 401(k)’s weakness is that the worker absorbs market risk and longevity risk with no guaranteed conversion into lifetime income. The deciding factor is mobility. A stable, single-employer career favors the pension; a mobile career favors the account plan.

Q: Did the shift from pensions to 401(k) plans happen by law or by choice?

By both, which is why both simple stories fail. Employers made real choices: they preferred the predictable cost of a match contribution to the open-ended liability of a benefit promise, and many adopted account plans voluntarily. But the choices were made inside a legal environment that legislation had reshaped. The 1974 statute raised the fixed cost of defined benefit plans; the 1978 tax provision created the account alternative; later tax law tightened pension rules further; and the 2006 and 2022 statutes made the account system the legal default through safe harbors and mandates. Structural change also contributed: employment moved from manufacturing toward services, workers valued portability, and longer life spans raised the cost of lifetime promises. The brief’s name for the pattern is the cost-and-liability ratchet, and the causal weighting remains contested. What the record supports is that the shift was legislated in its direction even when it was voluntary in its details.

Q: Why do government workers still have pensions when 401(k) plans replaced them elsewhere?

Public employment sits in a different legal and budgetary environment, and the divergence is itself evidence about what drove the private sector shift. Federal law exempts governmental plans from the core ERISA regime that raised private pension costs, so the ratchet that pushed private employers toward accounts never applied with the same force. Public employers also face different incentives: they can fund promises from tax bases rather than profits, they value workforce stability in teaching, policing, and administration, and collective bargaining has historically defended the benefit structure. The result is a natural experiment. Where the legal cost structure favored accounts, pensions retreated; where it did not, they persisted. Public plans face their own funding strains, and the point is not that they are immune to pressure. The point is that the private sector shift tracked the law’s incentives rather than some universal employee preference.

Q: Is a 401(k) insured the way a pension is?

No. A defined benefit pension carries federal insurance backing through the Pension Benefit Guaranty Corporation, which takes over failed single-employer plans and pays benefits within statutory limits, and administers a separate program for multiemployer plans. A 401(k) has no equivalent. The account holds securities in the worker’s name, and no federal program replaces balances lost to market declines or guarantees the account’s value at retirement. Protections that do exist operate on a different axis: plan assets must be held in trust apart from the employer’s own property, fiduciary duties govern the people who run the plan, and brokerage-level protections guard against institutional failure rather than market loss. The distinction is worth stating plainly because it is widely misread. Insurance covers the pension promise against sponsor failure; nothing insures the 401(k) balance against the market.

Q: Are multiemployer pension plans safer than 401(k) plans?

Not in any simple sense. A multiemployer pension pools investment and longevity risk across many employers and workers in one industry, and it carries federal insurance backing through the multiemployer program, but that program’s guarantees have historically been far thinner than the single-employer program’s, and the insurance addresses plan failure rather than benefit adequacy. Congress created a special financial assistance program in 2021 for deeply distressed multiemployer plans, which is itself a measure of the strain. A 401(k) has no insurance at all but also no plan-level failure mode: the account is the worker’s own, portable and intact regardless of any employer’s fate. The comparison therefore trades one kind of safety for another. The multiemployer pension offers a promised benefit with pooled risk and limited insurance; the 401(k) offers individual control with no backstop. Which is safer depends on the plan’s funding, the worker’s tenure, and how each handles a market downturn.

Q: How do vesting schedules differ between a pension and a 401(k)?

Vesting answers the question of when employer money becomes the worker’s money for good, and the two systems answer it differently. In a pension, the entire benefit comes from the employer, so the worker owns nothing until the vesting schedule is satisfied; federal law sets minimum schedules measured in years of service, and a worker who leaves before vesting walks away with no benefit at all. In a 401(k), the worker’s own contributions are always fully vested from the first paycheck, because the money was the worker’s salary to begin with. Only the employer’s matching contributions are subject to a vesting schedule, and those schedules are typically short, often measured in two or three years rather than five. The practical difference is large. A mobile worker keeps every dollar she contributed to a 401(k) no matter when she leaves, while the same worker can forfeit an entire pension by departing one year too early.

Q: What happens to a pension and a 401(k) when the employer goes bankrupt?

The two systems fail in opposite ways. When a defined benefit sponsor goes bankrupt with an underfunded plan, the federal insurance program takes over the plan and pays benefits within statutory guarantee limits; the worker keeps a pension, though possibly a reduced one. The 401(k) has no insurance program because it needs none on this axis: the assets are held in a trust that is legally separate from the employer’s property, so the employer’s creditors cannot reach them and the bankruptcy does not touch the account. The 401(k) worker’s danger is the market, not the sponsor. A pension promises a benefit and insures it against employer failure; a 401(k) promises an account and protects it by keeping it out of the employer’s hands. Each system guards against the other’s catastrophe and leaves its own exposed.

Q: Which system rewards long tenure and which rewards a mobile worker, a pension or a 401(k)?

The pension rewards long tenure and the 401(k) rewards mobility, and the mechanics explain why. A defined benefit formula typically ties the payout to years of service and late-career earnings, so the benefit accrues slowly in the early years and heavily in the final ones; the worker who stays twenty-five years earns far more than five workers who each stay five. The formula punishes departure twice over: the leaver forfeits future accrual and, under many designs, the unvested portion as well. A 401(k) accrues evenly with each contribution and belongs to the worker from the start; changing employers interrupts nothing, because the balance rolls over and keeps compounding. The career pattern therefore decides the winner. The single-employer lifer does best under the pension formula; the worker who changes jobs every few years does best under the portable account.

Q: How does the tax treatment differ between a pension and a 401(k) at contribution and withdrawal?

At the broadest level the two systems share the same tax logic: contributions go in before tax, the money grows tax deferred, and withdrawals are taxed as ordinary income in retirement. The differences sit in the details. Pension funding is entirely the employer’s affair, with deductible contributions and no employee-side choice to make. In a 401(k) the worker chooses the contribution level and, in many plans, chooses between a traditional pre-tax contribution and a Roth contribution made with after-tax dollars, which grows tax free and comes out tax free in retirement. The pension offers no Roth-style option; its payout is always ordinary income. Contribution limits also differ in structure: the 401(k) caps what the worker and employer together may add each year, while the pension’s limits are set by the benefit formula and the funding rules. The tax system treats both as deferred compensation, but only the account plan gives the worker a choice about when to pay the tax.

Q: Can a worker be covered by both a pension and a 401(k) at the same time?

Yes, and the combination is common. Nothing in federal law prevents an employer from sponsoring both a defined benefit plan and a 401(k) at once, and many employers did exactly that during the transition decades, adding an account plan while keeping the pension for existing workers. A worker can also hold both across a career: a frozen pension from an earlier employer pays a fixed benefit at retirement while the current employer’s 401(k) holds the growing account balance. The two systems stack rather than cancel. Where the combination creates complexity is in the benefit design: some employers reduced pension accruals as they introduced matching contributions, effectively trading one for the other. But as a legal matter the systems are fully compatible, and a worker receiving a pension check while drawing down a 401(k) is one of the more secure retirement pictures the comparison produces.

Q: How do death benefits differ between a pension and a 401(k)?

The pension treats death as a contingency inside the benefit formula, while the 401(k) treats the account as property that passes to a beneficiary. A married pension participant’s default is a joint and survivor annuity that continues payments to the surviving spouse, and federal law requires the spouse’s written consent before the worker can waive that protection; an unmarried worker’s pre-retirement death typically triggers a survivor annuity for a designated beneficiary. A 401(k) balance simply transfers to whoever the worker named, with the spouse holding protected rights that require consent to name someone else. The difference matters most for the surviving spouse. The pension’s survivor annuity is automatic unless waived, which protects the inattentive; the 401(k) requires the worker to name a beneficiary and keep the designation up to date, which rewards the organized and punishes the forgetful.

Q: How did the 1978 provision that created the 401(k) affect workers who already had pensions?

It changed nothing about their existing pensions and everything about their employers’ future choices. The provision added section 401(k) to the tax code as a salary reduction rule; it did not amend a single accrued pension benefit, and the 1974 statute’s protections for benefits already earned remained fully in force. What changed was the menu. Employers that had offered only a pension could add an account plan cheaply, and many did, layering 401(k)s on top of existing pensions before later freezing the pension plans. A worker who already had a pension therefore often gained a second vehicle rather than losing the first, at least in the early years of the shift. The provision’s real effect on pension holders was indirect and slow: it gave employers the alternative that made pension freezes thinkable, and the freezes followed over the next two decades.

Q: How should a job seeker evaluate an employer’s choice between offering a pension and a 401(k)?

By pricing the offer rather than admiring the label. A pension’s value depends on staying: the job seeker should ask how the benefit formula accrues, how long vesting takes, and what the projected payment looks like at the tenure she realistically expects, not the tenure the brochure assumes. A 401(k)’s value is easier to read: the match formula, the vesting schedule on the match, and the fee burden on the investments. A generous match with immediate vesting can outweigh a modest pension accrual for anyone who will not stay decades, while a strong pension formula can dominate for the worker who will. The comparison also turns on the employer’s financial health, since the pension is only as reliable as its sponsor and its funding. The disciplined question is always the same: how much retirement income does this job actually buy me if I leave in five years, and how much if I stay for twenty-five.

Q: Why do small employers favor 401(k) plans over pensions?

Cost predictability and administrative simplicity. A defined benefit plan obligates the sponsor to fund a promised benefit whatever the markets do, which means actuarial valuations, funding rules, annual premiums to the insurance program, and legal exposure when assumptions miss. For a small firm those fixed costs and open-ended liabilities are disproportionate to the payroll. A 401(k) converts the obligation into a known line item: the employer contributes a match if and when it chooses, with no funding target, no insurance premium, and no promise to make up market shortfalls. The account design also fits how small-firm workers actually behave, since shorter tenures make portability more valuable than a backloaded pension formula. Federal law reinforced the preference at every step: each statute that raised the regulatory price of the pension lowered the relative price of the account plan, and small employers, who feel fixed costs most sharply, responded first and most completely.

Q: What does a pension freeze mean for benefits already earned, and how does that differ from a 401(k) account?

A freeze stops the future, not the past. When an employer freezes a defined benefit plan, workers keep every dollar of benefit accrued through the freeze date; federal law’s anti-cutback rule forbids reducing benefits already earned, so the accrued promise is locked in and typically paid at retirement age under the original formula. What ends is further accrual: years of service after the freeze add nothing, and the employer usually directs new contributions into a 401(k) or similar account instead. The contrast with a 401(k) is structural. An account balance cannot be frozen in the same way because it is already the worker’s property; the employer can stop matching, but cannot stop the balance from belonging to the worker or erase its past growth. The freeze is therefore the pension’s version of a plan change: the employer caps its future liability while the law protects what workers already earned.

Q: How do lump-sum buyout offers fit the comparison between a pension and a 401(k)?

A lump-sum offer converts the pension into the 401(k)’s logic. The employer offers the worker a one-time payment, calculated as the present value of the promised annuity, in exchange for surrendering the lifetime benefit. Accepting moves investment risk and longevity risk from the sponsor to the worker, which is the same direction as the broader historical shift: the guaranteed stream becomes a balance the worker must manage. The offer suits the worker who values control, portability, or estate planning over a guaranteed stream, and it suits the employer by extinguishing a long-dated liability. The comparison turns on what is lost in the exchange. The annuity insures against outliving one’s money; the lump sum does not, and a worker who accepts it takes on the same two risks that every 401(k) holder carries. The buyout is the pension system borrowing the account system’s risk allocation for a single transaction.

Q: What role does Social Security play alongside a pension versus a 401(k)?

Social Security is the public floor beneath both systems, and its role grows as the private leg shrinks. Its benefit formula replaces a larger share of earnings for lower-paid workers, which makes it the most progressive element in most retirees’ income. Paired with a pension, Social Security is the second of two annuity-like streams: the worker retires with a government benefit and an employer benefit that both pay for life. Paired with a 401(k), Social Security is the only guaranteed lifetime income the worker has, while the account supplies the variable leg that rises and falls with the market. That difference explains why the account system’s rise increased the importance of the public program. The pension worker could treat Social Security as a supplement; the 401(k) worker often must treat it as the foundation, because nothing else in the package converts savings into income that cannot be outlived.