An Amendment with Two Halves Pointing in Opposite Directions

Few statutes try to pull the same policy in two opposite directions at once, and fewer still succeed at both. The Pension Protection Act of 2006 did. One half of the statute looks backward, to the traditional pension promise, and makes that promise harder to break by making it harder to underfund. The other half looks forward, to the account-based system that was already replacing pensions, and rewrites the legal rules so that millions of workers save by default instead of by deliberate choice. Congress passed both halves in a single enactment because both halves answered the same fear: that American workers were arriving at retirement with less than they had been led to expect.

The fear had concrete roots. In the years before 2006, several large defined benefit plans failed and handed their obligations to the federal pension insurer, and the failures forced Congress to confront how the funding rules written a generation earlier had let sponsors promise benefits first and fund them later, sometimes never. The old rules let a plan measure its obligations against asset values smoothed across time and pay off its shortfalls gradually, which meant that a plan could look adequately funded on paper while sinking in reality. At the same time, a different problem was gathering in the account-based world. Participation in 401(k) plans was stagnant because enrollment required paperwork, paperwork required initiative, and initiative was exactly what most workers never supplied. Two failures, one of funding and one of inertia, and the 2006 act answered both.

That pairing is what makes this the amendment-stage article of the series. The original statute, ERISA, had been built to police the pension promise: disclose it, fund it, insure it. By 2006 the pension promise itself was receding, and the question before Congress was no longer only how to protect pensions but how to make the system replacing them actually work. The Pension Protection Act is therefore best read not as a new regime but as a midlife revision, a statute that took the machinery of an earlier law and repurposed it for conditions the earlier law had never imagined. The repurposing worked better in one half than in the other, and the difference between the two halves is the story of American retirement after 2006.

The One Test

Ask a reader who has finished this article to explain the Pension Protection Act in three claims, and the article has done its job if the reader can say these things. First, the single largest enactment of behavioral economics into American law is a pension statute, not a consumer protection bill or a tax credit: Congress wrote the research finding that defaults dominate participation directly into the code. Second, three safe harbors turned automatic enrollment, target-date defaults, and escalating contribution rates from legally risky practices into legally protected ones, so that a fiduciary who enrolled a silent worker and invested the money sensibly was shielded rather than exposed. Third, the same statute’s funding rules made traditional pensions more expensive and more volatile to sponsor, which is why a serious reader must weigh the argument that a law named for protecting pensions accelerated their disappearance.

The three claims belong together because the statute’s defenders and its critics each need all three to make their case. The behavioral design claim explains why the act matters even to workers who never heard of it. The safe harbor claim explains the mechanism. The freeze claim explains the cost. An honest account of the Pension Protection Act refuses to drop any of them.

Why did the 2006 act clear legal obstacles instead of mandating enrollment?

Because a mandate would have fought employer resistance while a cleared path harnessed it. Preempting state wage laws and shielding fiduciaries let sponsors adopt automatic enrollment voluntarily, which proved faster than compulsion: the research showed defaults work through inertia, and inertia works best when the employer chooses the default willingly rather than under order.

The legal problem was subtle enough that many employers misunderstood it, and the misunderstanding ran in the comfortable direction. Because ERISA broadly preempts state law, plan sponsors assumed federal law had cleared the field, and some enrolled workers automatically without incident. But wage payment is an area where states have historically regulated, and benefits advisors had asserted that state wage-withholding statutes could forbid automatic deductions outright, giving employers reason to doubt that ERISA’s general preemption reached those statutes. A payroll deduction taken without the affirmative written consent the state statute demanded was, in the cautious lawyer’s reading, an unlawful withholding of wages. Multiply that exposure by every paycheck of every automatically enrolled worker, and the prudent course was to ask permission first. The entire 401(k) system then ran on permission, and permission was the bottleneck.

The research behind the design, the automatic enrollment findings of Brigitte Madrian and Dennis Shea and the escalation findings of Richard Thaler and Shlomo Benartzi, is examined in full later in this article. What matters for the mechanism is that the drafters treated that research as a specification rather than a decoration: people save too little not because they prefer to retire poor but because saving requires a chain of decisions that inertia breaks at every link, and a well-chosen default repairs the chain by making the right outcome the outcome of doing nothing.

Congress wrote those findings into law through three safe harbors, each one removing a specific legal risk that had kept a specific behavioral practice at the margins. The first safe harbor addressed the withholding problem directly. The act provided that state wage-withholding laws are preempted with respect to automatic contribution arrangements that meet the statute’s conditions, which meant an employer that followed the federal script no longer had to fear the state-law exposure described above. Preemption is a dry word for a decisive move: it converted automatic enrollment from a practice that required a brave general counsel into a practice that required only compliance.

The second safe harbor addressed the fiduciary problem. A fiduciary who invests a participant’s money badly can be sued for breach of duty, and before 2006 the safest way to invest a silent participant’s money was to put it in the lowest-risk option available, which in practice meant a money market fund or stable value fund that preserved principal while losing ground to inflation year after year. The act created the qualified default investment alternative, a category of diversified default investments, and provided that a fiduciary who defaulted a non-electing participant into a qualifying alternative was protected from liability for the investment outcome. The logic inverted the old caution. Instead of punishing the fiduciary who put a young worker’s money in an age-appropriate mix of stocks and bonds, the law now protected that fiduciary and left exposed the one who let inflation quietly confiscate the account. The consequence, exactly as the drafters intended, was the target-date fund becoming the default holding of American retirement saving: a single fund that adjusts its mix of growth and safety as the worker ages, chosen by nobody and held by millions.

The third safe harbor addressed the contribution-rate problem through the automatic contribution arrangement with escalating default rates. The statute blessed arrangements in which the default contribution rate rises over time unless the worker opts out, which is Thaler and Benartzi’s escalating-contributions finding rendered as federal law. A worker enrolled at a modest initial rate would see that rate step upward on a schedule, each increase small enough to go unfelt, until the account was accumulating at a level the worker would never have chosen affirmatively. The worker kept the absolute right to stop the escalator at any point, which is what made the arrangement a nudge rather than a mandate, but the empirical record on which the statute rested said a substantial fraction of workers would not stop it.

The act also made permanent the low-income saver credit, the tax credit available to lower-earning workers who contribute to retirement accounts. It was the smallest of the defined contribution provisions and the most conventional, a tax incentive rather than a behavioral device, and its mechanics are examined in the defined contribution half below.

Step back and the design principle is stark enough to state plainly. The default is the policy: the act moved retirement outcomes for millions of workers swept into automatic plans without changing a single contribution requirement or tax rate, by changing only what happens when a worker does nothing. That is the clearest demonstration in American law that default rules are substantive policy rather than plumbing. Every mandate the statute declined to impose, it achieved more effectively by rewriting the path of least resistance, and the path of least resistance turned out to be the most powerful regulatory instrument Congress had.

The Statutory Identity: Public Law 109-280 at the Amendment Stage of ERISA’s Life

The Pension Protection Act of 2006 is Public Law 109-280, signed August 17, 2006, and it amends both the Employee Retirement Income Security Act (/2012/07/15/erisa-1974-complete-guide/) and the Internal Revenue Code. The dual amendment is not a formality. American retirement law lives in two titles at once: ERISA supplies the fiduciary standards, the funding rules, the disclosure duties, and the insurance system, while the Code supplies the tax qualification rules that make employer plans worth sponsoring in the first place, rules whose modern shape the series traces to the Tax Reform Act of 1986 (/2013/05/01/tax-reform-act-1986-complete-guide/). A reform that touched only one title would have left half the machine running on the old logic, so the 2006 act rewired both.

The enactment itself moved with unusual speed for a bill of its scope. Introduced as H.R. 4 in the 109th Congress, the measure passed the House on July 28, 2006, by 279 to 131, and the Senate on August 3 by 93 to 5, and was signed into law as Public Law 109-280 on August 17, 2006. The Congressional Research Service’s summary treated it as the most comprehensive pension-law reform since ERISA itself. The scale of the rewrite is visible in the provisions it added: section 624(a) created a new ERISA section 404(c)(5) for the default-investment safe harbor, section 902(f) added ERISA section 514(e) for the preemption of state wage laws, section 902 created the tax code’s automatic contribution arrangement provisions, and sections 812 and 833 made the saver’s credit permanent. An amendment-stage statute does not repeal its parent. It keeps the parent’s architecture and changes the operating assumptions inside it, the way a renovation keeps the foundation and moves the walls. The Pension Protection Act kept ERISA’s architecture: the fiduciary duties, the insurance corporation, the tax qualification framework. It moved the walls: what counts as funded, what counts as a prudent default, what an employer may do when a plan slips underwater.

The amendment stage carries the series’ recurring finding with it. Across these articles, the reforms that changed procedure or defaults have consistently outperformed the reforms that changed commands. A command tells an actor what to do and invites evasion; a default changes what happens when the actor does nothing and harnesses inertia instead of fighting it. The 2006 act is the purest illustration in the series because it ran the experiment in both directions at once. On the defined contribution side it changed defaults and transformed outcomes. On the defined benefit side it changed commands, tightening the funding mandates, and produced consequences its drafters are still arguing about. The two halves of one statute, enacted on the same day, test the series thesis against itself.

Enactment: The Making of Public Law 109-280

The Pension Protection Act arrived with a speed that belied its size. Introduced as H.R. 4 in the 109th Congress, it passed the House on July 28, 2006, by 279 to 131, cleared the Senate on August 3 by 93 to 5, and was signed on August 17, 2006, as Public Law 109-280. The Congressional Research Service’s summary treated it as the most comprehensive pension-law reform since ERISA itself, Public Law 93-406, and the comparison is instructive. ERISA had created the architecture: fiduciary duties, funding standards, disclosure, and the insurance corporation. The 2006 act kept every element of that architecture and changed the calibration inside it, which is why the statute reads as hundreds of surgical amendments rather than a new code. Section 624(a) added the new ERISA section 404(c)(5) that underpins the default-investment safe harbor. Section 902(f) added ERISA section 514(e), the preemption clause. Section 902 created the tax code’s automatic contribution arrangement provisions, sections 401(k)(13), 401(m)(12), and 414(w). Sections 812 and 833 made the saver’s credit permanent. Each of these is a small textual insertion with an outsized behavioral consequence, which is the amendment stage’s signature move.

President Bush’s signing statement captured the statute’s theory of action in a single verb: the act, he said, “encourages” employers to automatically enroll workers in 401(k) plans. Encourages, not requires. The choice of verb was the policy. Congress had the votes for a funding crackdown, where the moral case against underfunded promises was clear and the losers were diffuse, but it did not have the votes, or the desire, for a savings mandate, where the regulated party would have been every employer in America. So the statute encouraged where it could not command, and it made the encouragement expensive to refuse: testing relief for the QACA, liability protection for the QDIA, preemption for the payroll deduction. The 93-to-5 Senate vote suggests how little opposition this approach provoked. An enabling statute built on published research, with something for sponsors in every provision, is the kind of bill that passes nearly unanimously.

The act’s provisions also took effect on different clocks, and the stagger matters for the story. The preemption of state wage-withholding laws took effect on enactment, August 17, 2006, which meant the legal doubt over automatic payroll deductions lifted immediately. The funding target phased in across 2008, 2009, and 2010, which meant the defined benefit discipline arrived gradually and then collided with the financial crisis. The QDIA relief and the automatic enrollment provisions applied generally for plan years beginning after December 31, 2007, which gave the Department of Labor time to write the regulation, finalized October 24, 2007, and gave sponsors a full plan year to redesign their defaults. A statute whose halves took effect on three different schedules was really three statutes sharing a public law number, and each schedule shaped how its half was received.

The Defined Benefit Half: How Full Funding Became the Law

To understand the defined benefit half of the act, begin with the elementary mechanics of a pension promise and the precise point where the old law let the mechanics fail. A defined benefit plan is a promise to pay: a monthly check in retirement, calculated by formula, for as long as the retiree lives. The employer sponsoring the plan must hold assets today against payments that will not come due for decades. Funding measurement is the discipline of comparing those two quantities, the assets in hand and the present value of the promises outstanding, and the discipline only works if both quantities are measured honestly.

The old regime measured them loosely. Asset values could be smoothed across years, which muted the signal when markets fell. Liabilities could be discounted at rates that flattered the plan’s position. And when a shortfall did appear, the sponsor could amortize it, pay it down in installments, over periods as long as 30 years, a horizon long enough that the installments felt painless and the shortfall never quite closed. The old rules also excused deficit-reduction contributions for plans at least 90 percent funded, which meant a plan could sit just above the threshold indefinitely without ever repairing the gap beneath it. The system had a name for what it was doing, and the name was funding, but the economic reality was forbearance. A sponsor could promise benefits in a negotiation, book the liability at a flattering valuation, and leave the actual funding for a future management team. When the future arrived and the money was not there, the federal insurer absorbed the loss and the workers absorbed the benefit cuts.

The 2006 act attacked that forbearance at each joint. First, it established a full funding target: the plan’s assets were to cover the entire measured value of its benefit obligations, not a fraction of them, and the measurement was to reflect the plan’s actual position rather than a smoothed or flattered one. The target turned funding from an aspiration into a standard. A plan was either at full funding or it was short, and short had consequences.

Second, it shortened the amortization of shortfalls. Where the old rules had let sponsors stretch the repayment of a funding gap across a long horizon, the new rules compressed the schedule to seven years, which meant larger required contributions sooner. The compression was deliberate. A shortfall that must be closed quickly disciplines the sponsor in a way that a shortfall payable over decades never can, because the pain of the contribution arrives while the executives who made the promise are still in the building. But the compression also introduced the volatility that would become the act’s most disputed feature. Required contributions now moved with asset values and interest rates on a shorter fuse, so a market decline could translate into a sharply higher funding bill within a few years rather than a gently rising one across a generation.

Third, the act restricted what an underfunded plan could do with its benefits. A plan below 80 percent funded could not pay lump sums beyond limited amounts, and a plan below 60 percent funded could not accrue additional benefits at all, while amendments increasing a plan’s liability were prohibited while a funding waiver was in effect. The logic was airtight and, to sponsors, infuriating. A plan that cannot afford the benefits it has already promised has no business promising more, and a lump sum paid out of an underfunded plan is money removed from the pool that remains for everyone else. Before 2006, sponsors of troubled plans had sometimes sweetened benefits in labor negotiations, effectively spending the insurer’s money, because the downside of a deeper shortfall fell on the federal backstop rather than the company. The restrictions closed that moral hazard by tying the sponsor’s hands at exactly the moment the temptation was greatest.

Fourth, the act revised the premiums that sponsors pay to the Pension Benefit Guaranty Corporation, the federal insurer that takes over failed plans. The act made permanent the $1,250-per-participant termination premium created by the Deficit Reduction Act of 2005, redefined the variable-rate premium’s “unfunded vested benefits” base to match the new funding target, and capped variable premiums for small employers, while the flat-rate increase to $30 per participant (from $19) was enacted earlier by the Deficit Reduction Act of 2005, not this act. The changes served two purposes at once: they shored up the insurer’s own finances, which had been strained by the large plan failures of the preceding years, and they priced the insurance more accurately, so that sponsoring an underfunded plan cost more than sponsoring a well-funded one. The premium provisions were the market signal inside a command-and-control statute. Everything else in the defined benefit half told sponsors what they must do; the premiums told them what it would cost to keep doing it the old way.

The intent behind all four changes was singular and stated plainly: to stop plans from promising benefits they had not funded. Congress had watched the old funding rules permit a slow-motion default, in which each year’s forbearance became next year’s shortfall, and it chose to make the rules unforgiving. An unforgiving rule has a virtue and a vice. The virtue is that it ends the forbearance. The vice is that it ends it by making the underlying activity more expensive, more volatile, and less attractive, which is precisely what happened next.

Walk through the measurement in practice and the vice becomes visible. A plan’s liabilities are the present value of decades of future checks, and present value moves inversely with interest rates: when rates fall, the same future checks are worth more today, and the measured shortfall widens even if nothing about the workforce or the assets has changed. A plan’s assets move with the markets. Under the new, shorter amortization, both movements fed into required contributions quickly. The result was a funding bill that could swing sharply from one valuation to the next on the strength of a bond-market move the sponsor neither caused nor controlled. For the chief financial officer of a manufacturing company with a large legacy plan, the pension had become a second business, one whose earnings were hostage to interest rates and whose losses arrived as cash calls. The old rules had hidden that business inside smoothing and long amortization. The new rules exposed it.

None of this was an accident. Stricter funding is procyclical by construction: required contributions rise hardest in downturns, when sponsors can least afford them. Congress accepted the tradeoff because the alternative, the slow forbearance of the old regime, had already produced the failures the act was written to prevent. But accepting a tradeoff is not the same as controlling its consequences, and the consequences arrived in the form of the freeze.

How did stricter funding rules change the economics of keeping a pension open?

They raised the cost and the volatility of sponsorship at once. Required contributions rose when asset values fell, which typically meant the bill arrived when the company’s own business was weakest. Meanwhile, the premium changes made underfunding costlier to insure, so finance officers increasingly asked why the company bore pension risk at all.

A freeze, in the vocabulary of pension law, is the sponsor’s decision to stop the plan’s growth: no new participants admitted, or no new benefits accrued, or both, while the existing obligations are paid off over time. A frozen plan is not a terminated plan. The checks already earned still get written. But a frozen plan is a plan with no future, a promise being wound down rather than renewed, and the freeze is the last decision a sponsor makes before the plan becomes history. In the years after 2006, freezes accelerated: Pension Benefit Guaranty Corporation data show plans with any accrual or participation freeze provision rising from 8,059 of 28,876 insured plans (27.9 percent) in 2008 to 10,220 (39.9 percent) in 2011.

The freeze trend itself is documented across official sources, and the documentation is what makes the causal argument both serious and genuinely disputed. The Pension Benefit Guaranty Corporation’s own records show the insured plan population shrinking as sponsors exited the system; the corporation insured just over 35,200 defined benefit plans in 2001, down from an all-time high of 114,400 in 1985, according to an Employee Benefit Research Institute analysis of the corporation’s data. The corporation’s data tables break the freeze provisions into categories that matter for the story: some plans were hard-frozen, with all accruals stopped; some were partially frozen, with accruals continuing for some participants but not others; and some were closed to new entrants while existing participants kept earning benefits. Each category represents a different degree of retreat, and the growth across all of them after 2006 is what the competing explanations fight over. The categories also matter for what workers experienced. In a hard freeze, every participant stops earning new benefits at once, and the plan becomes a closed pool paying off past promises. In a partial freeze, the pain is distributed unevenly, with some cohorts grandfathered and others cut off, which is why partial freezes generated the sharpest disputes in bargaining. In a closure to new entrants, existing workers keep accruing while new hires are offered only the 401(k), creating the two-tier workforce that became the standard corporate compromise: the pension for those who already had it, the account for everyone after. The trend is not in doubt. What is in doubt is how much of it the 2006 act caused.

The case that the act accelerated the freeze runs through the mechanism described above. Critics of the funding rules argued that the combination of a full funding target, seven-year amortization, and the premium changes would make defined benefit sponsorship untenable for companies whose core business was not finance, and that the pattern after 2006 matched the prediction: sponsors facing volatile contribution requirements chose to cap their exposure by freezing plans, converting the open-ended pension promise into a fixed, amortizing obligation. On this reading, the statute protected pensions the way a strict building code protects old houses, by making them so expensive to maintain that the owners tear them down.

The competing explanation holds that the freeze was already underway and that the act merely governed a retreat already in motion. On this reading, the forces that had been eroding defined benefit sponsorship for decades, the shift toward a mobile workforce that valued portable accounts, the accounting rules that had already exposed pension volatility on corporate balance sheets, the simple preference of younger workers for benefits they could see and control, were the true drivers, and the 2006 funding rules arrived too late to be the cause. The accounting point deserves a sentence of its own. Pension accounting standards had, years before the act, begun forcing sponsors to recognize pension volatility on their financial statements rather than smoothing it away in footnotes, which meant the market was already punishing companies for pension risk before the statute repriced it in cash terms. A chief financial officer watching pension swings hit reported earnings needed no statute to sour on the defined benefit promise; the 2006 act merely converted a reporting headache into a funding bill. Labor economists in this camp point out that plan coverage had been declining since long before 2006, a decline visible in the insured-plan counts falling from 114,400 in 1985 to just over 35,200 in 2001, and that the act’s contribution was marginal against the larger tide.

Both explanations are politically useful, which is exactly why the causal question must stay open. Opponents of funding discipline cite the freeze as proof that the act destroyed what it claimed to protect. Defenders of the act cite the same freeze as proof that the old system was already dying and that the statute at least made the dying honest. The neutral posture, and the one this article adopts, is to present the funding rules, the freeze trend, and the competing explanations with their sources named, and to label the causal question open. One piece of timing complicates every account: the funding target phased in across 2008 through 2010, years that coincided with the 2008 financial crisis, which battered asset values just as the new measurement took hold. Disentangling the statute’s contribution from the market’s is part of why the causal question stays open. What can be said without qualification is that after 2006, sponsoring a traditional pension became a more expensive, more volatile, and more transparent undertaking than it had been before, and that fewer employers chose to undertake it. Whether the statute caused the decline or merely priced it honestly is the dispute, and the dispute is genuine.

The Funding Mechanics in Detail

The defined benefit half repays closer inspection, because its individual provisions show how an amendment-stage statute works: not by announcing a new philosophy but by recalibrating the numbers inside an old one. Start with the phase-in. The full funding target did not arrive all at once. The statute moved plans toward it on a schedule: 92 percent for 2008, 94 percent for 2009, and 96 percent for 2010, on the way to 100 percent. The phase-in gave sponsors a short runway while committing them to the destination, and it meant the new discipline bit hardest just as the financial crisis was battering the asset side of every plan’s balance sheet. That collision of a tightening rule with a falling market is the mechanical core of the procyclicality complaint, and it is also the reason the phase-in schedule matters to the freeze debate: the years the new measurement took hold were the years sponsors could least afford what it measured.

Consider what the new measurement replaced. Under the prior rules, a funding shortfall could be amortized over periods as long as 30 years, and deficit-reduction contributions were excused for plans at least 90 percent funded, which let a plan hover just above the threshold indefinitely. The 2006 act replaced the 90-percent world with a 100-percent target and the 30-year horizon with seven years. The arithmetic of that compression is the whole story in miniature. A shortfall amortized over three decades is a footnote in the annual budget; the same shortfall amortized over seven years is a line item the chief financial officer must explain every quarter. The discipline is real, and so is the volatility: every move in interest rates and asset prices now repriced the obligation on a seven-year fuse instead of a thirty-year one.

The premium provisions deserve the same careful reading, because they are the most misattributed part of the statute. The flat-rate premium increase from $19 to $30 per participant and the creation of the $1,250-per-participant termination premium belonged to the Deficit Reduction Act of 2005, which took effect for the 2006 plan year, before the Pension Protection Act was signed. What the 2006 act did was narrower and more technical, and it is worth stating precisely. Section 401(a) redefined “unfunded vested benefits,” the base for the variable-rate premium, to conform to the new funding target, effective in 2008, so the premium base and the funding measure would move together. Section 401(b) made the $1,250 termination premium permanent, payable for the year of termination and each of the next two years. Section 405 capped variable-rate premiums for small employers, those with 25 or fewer employees, at $5 multiplied by the number of participants, effective for plan years beginning January 1, 2007. And a special $2,500-per-participant termination premium applied to commercial-airline plans terminating within a five-year window, a carve-out for an industry whose pension failures had helped motivate the legislation. None of these provisions raised the headline premium rate. All of them changed who paid what for the insurance, which is how a statute prices discipline without announcing a tax.

Two quieter provisions of the defined benefit half show the amendment stage operating at its most characteristic. The act clarified, prospectively, the legal status of hybrid and cash-balance plans, the account-styled defined benefit designs whose treatment under the pension rules had been contested. The clarification applied going forward, ending years of uncertainty for sponsors that had adopted the designs without settling the litigation risk. And the act made permanent the pension provisions of the Economic Growth and Tax Relief Reconciliation Act of 2001 that had been scheduled to expire in 2010, including the higher contribution and deductible limits for 401(k) plans and IRAs. That permanence is easy to overlook beside the funding drama, but it mattered enormously for the account-based system: the higher limits the defined contribution half relied on would otherwise have sunset just as the automatic enrollment machinery was switching on. An amendment that tightened the old world while quietly locking in the tax economics of the new one is the 2006 act in a single gesture.

Walk through the measurement as a sponsor would have experienced it, and the abstract discipline becomes a concrete cash flow. Imagine a plan whose promised benefits, discounted to present value, exceed its assets. Under the old rules, the sponsor measured the assets with smoothing that muted market declines, discounted the liabilities at rates that flattered the position, and amortized the resulting shortfall over up to three decades. The required contribution was the product of three dampeners, each one shrinking the bill. Under the new rules, the assets are measured closer to market, the liabilities are measured against the full funding target, and the shortfall is amortized over seven years. The required contribution is the product of three amplifiers. The same economic reality produces a small bill under the old measurement and a large one under the new, and the difference is not a difference in the plan’s health but in the honesty of the ruler.

The smoothing dampener deserves particular attention, because it was the most technical and the most abused of the three. Asset smoothing let a plan value its holdings not at the market price on the valuation date but at an average of market values over several years, sometimes with further adjustments that pulled the reported value toward expectations. In a rising market, smoothing understated assets and flattered no one; in a falling market, it overstated them, which is exactly when the overstatement mattered. A plan that had lost a fifth of its portfolio in a crash could report an asset value that reflected mostly the pre-crash years, and the required contribution would be calculated against that fiction. The 2006 act’s move toward market-related measurement did not eliminate every averaging technique, but it narrowed the gap between the reported number and the price at which the assets could actually be sold. For the first time, the funding standard asked the question a worker would ask: if the plan had to pay everyone tomorrow, how much is there?

Now add interest rates, because they are the hidden variable that made the new ruler swing. A pension liability is a stream of payments stretching decades into the future, and its present value moves inversely with the discount rate: when rates fall, the same future checks are worth more today, and the measured shortfall widens without a single worker retiring or a single asset being sold. Under thirty-year amortization, a rate-driven widening was a distant problem, payable in installments long after rates might have recovered. Under seven-year amortization, it was an immediate cash call. The sponsor had not made a worse promise and had not managed the assets worse; the bond market had moved, and the statute translated the move into a bill within a few valuation cycles. This is the procyclicality mechanism in its entirety, and it explains why the funding rules bit hardest in downturns: falling rates and falling asset prices arrive together in a crisis, and the new measurement added them instead of averaging them away.

The benefit restrictions then operated as the emergency brake on the same vehicle. A plan below 80 percent funded that paid lump sums was, in economic substance, letting the healthiest or best-informed participants withdraw their share at full value while the remaining participants split a diminished pool. Barring those payouts protected the workers who stayed, but it also removed one of the features that had made traditional pensions attractive to mobile workers, which is why the restriction shows up on both sides of the freeze debate. A plan below 60 percent funded that kept accruing benefits was making new promises it could not fund, so the statute froze accruals. And the prohibition on benefit-increasing amendments while a funding waiver was in effect closed the specific abuse the drafters had in mind: the troubled sponsor that sweetened benefits in bargaining, spending the insurer’s money because the downside of a deeper shortfall fell on the federal backstop. Each restriction is a response to a moral hazard the old regime had tolerated, and each one made the traditional pension a less flexible instrument for the sponsor.

The waiver rule shows the moral hazard logic at its most precise. A funding waiver is the Internal Revenue Service’s permission to defer required contributions, granted when the sponsor demonstrates temporary business hardship. Under the old regime, a sponsor holding such a waiver could still amend the plan to increase benefits, effectively writing new promises on the insurer’s credit while excused from funding the old ones. The 2006 act prohibited any amendment increasing plan liability while a waiver was in effect, which closed the exact transaction the drafters feared: the sponsor that pleads hardship to skip contributions and then spends the savings on richer benefits in bargaining. The rule is narrow, technical, and revealing. It assumes the sponsor will exploit any gap between the funding obligation and the benefit promise, and it seals the gap at the moment of greatest temptation.

The premium base redefinition completed the pricing. By conforming “unfunded vested benefits” to the new funding target, the statute tied the variable-rate premium to the same honest measurement as the funding rules, so a sponsor could not look well-funded for premium purposes while looking underfunded for funding purposes. The small-employer cap at $5 multiplied by the number of participants shielded the smallest plans from the variable rate’s full force, a recognition that the compliance burden falls hardest where the administrative capacity is thinnest. And the airline carve-out, a $2,500-per-participant termination premium for commercial-airline plans terminating within five years, priced the industry’s specific history of pension failures into its own exit cost. The premium title of the act is a miniature of the whole statute: measure honestly, price the risk, and let the sponsor decide whether the activity is worth its true cost.

By 2011 the phase-in was complete and the target stood at 100 percent, up from the 90 percent world the old rules had tolerated. The transition had taken three years, spanned a financial crisis, and rewritten the funding standard for every defined benefit plan in the country. Whatever the freeze debate concludes about causation, the measurement that emerged from the transition stayed honest, and that honesty is the defined benefit half’s durable achievement even on the critics’ account.

The Honesty Dividend

Call it the honest half of the honest half. Whatever the freeze debate concludes about whether the funding rules accelerated the retreat from traditional pensions, no participant in the debate disputes what the rules did to measurement: they made it honest. The 90 percent threshold that had excused deficit-reduction contributions is gone, replaced by a 100 percent target. The 30-year amortization horizon that had turned shortfalls into footnotes is gone, replaced by seven years. The waiver that had let a sponsor skip contributions while sweetening benefits is sealed. The premium base that had floated free of the funding measure is conformed to it. Each of these is a small technical change, and together they are a large moral one: the law no longer permits the fiction that a promise is funded when it is not.

The dividend of that honesty compounds in ways the freeze debate tends to overlook. An honestly measured shortfall is a shortfall that can be closed, because the sponsor, the regulator, and the insurer are finally looking at the same number. The Pension Benefit Guaranty Corporation’s premium base, now tied to the funding target, prices the insurance on the true risk rather than the smoothed one, which means well-funded sponsors are no longer subsidizing the underfunded as heavily as the old measurement forced them to. The freeze data itself, the tables showing the rise from 27.9 percent to 39.9 percent of plans with freeze provisions between 2008 and 2011, exists in its current form because the measurement regime produces cleaner counts. Transparency is the precondition of every other reform, including the reforms the critics prefer.

There is a final irony in the dividend, and it belongs in the name-against-the-effect account. The honesty the funding half imposed made the traditional pension look worse before it could make it better, because the first honest measurement of a long-forbearing system is a large shortfall. Sponsors that saw the true number for the first time in 2008, with markets falling around them, can be forgiven for concluding that the promise was unaffordable, even if the promise had been affordable all along under honest funding from the start. The statute’s defenders make exactly this point: the funding rules did not create the shortfalls, they revealed them, and the revelation was painful because the forbearance had been long. Whether that revelation accelerated the freeze or merely priced it, the dividend remains. A pension system that measures honestly can be reformed honestly. A pension system that measures by fiction can only fail by surprise, which is what the old regime had been doing for years.

The Defined Contribution Half: Making Automatic Saving Legally Safe

The other half of the Pension Protection Act points in the opposite direction from the funding rules, and it is the half that changed the daily machinery of American retirement saving. Where the defined benefit provisions tightened commands, commanding plans to fund what they promised, the defined contribution provisions changed what happens when nobody acts. Congress took four steps: it preempted the state laws that had made automatic enrollment legally doubtful, it created a fiduciary safe harbor for defaulting silent participants into diversified investments, it built an automatic contribution arrangement with escalating default rates into the tax code’s safe harbor rules, and it made the low-income saver’s credit permanent. None of these steps ordered any employer to enroll anyone, none of them raised any contribution requirement, and none of them changed a tax rate. Their combined effect was to make the path of least resistance the path into the plan.

Understanding this half requires seeing the legal problem it solved, because the problem was not that employers were forbidden to enroll workers automatically. The problem was that doing so was legally doubtful in enough ways, across enough states, to keep most plan sponsors on the sidelines.

The Two Legal Obstacles to Automatic Enrollment

Before 2006, automatically deducting wages for 401(k) contributions without each worker’s written consent collided with state laws that treated unauthorized payroll deductions as unlawful, including garnishment statutes that did not distinguish saving from debt collection. Employers with workers in multiple states could not enroll everyone uniformly, and even favorable federal tax guidance did not settle the state-law exposure.

The doubt had two layers, and both mattered. The first layer was state wage-payment law. Many states required an employee’s express written authorization before an employer could withhold anything from a paycheck, and some states’ garnishment and wage-deduction statutes made no exception for retirement plan contributions that the employee had never affirmatively elected. The conceptual difference between taking wages to pay a judgment and taking wages to fund the worker’s own retirement account did not always survive contact with the statutory text. Plan sponsors with employees in multiple states faced a patchwork: an automatic enrollment feature lawful for the headquarters workforce might have been a wage-law violation for the same employer’s workers one state over. Because the tax code’s nondiscrimination rules generally required uniform plan terms, a sponsor could not simply auto-enroll in the friendly states and leave the others out. Some benefits advisors went further and asserted that state wage and payroll laws prohibited automatic contributions outright in the absence of a signed salary deferral agreement. The Internal Revenue Service had issued a favorable letter ruling for automatic enrollment in 1998, which addressed the federal tax treatment, but a letter ruling is not a statute and it could not silence the state-law argument. Employers were being asked to bet their payroll practices on a contested reading of dozens of state codes.

The second layer was fiduciary risk, and it was the more paralyzing of the two. A worker who is automatically enrolled is, by definition, a worker who has made no investment election. The employer must put that worker’s money somewhere. Before the act, the legally cautious choice was a money market or stable value fund: principal-preserving, lawsuit-avoiding, and terrible as a decades-long retirement investment for a young worker. Defaulting the silent participant into equities, or into a balanced fund with equity exposure, invited the question of what would happen when markets fell and a participant who had never chosen the investment sued over the loss. Under the Employee Retirement Income Security Act, the fiduciary who selects plan investments bears responsibility for the prudence of the selection, and the existing participant-direction shield of ERISA section 404(c) did not clearly cover a participant who had never directed anything. The Department of Labor’s own later account of the period observed that firms were reluctant to make recommendations about retirement saving that might cross the line from education into advice, because advice carried liability. So the default fund problem and the enrollment problem reinforced each other: even a sponsor willing to risk the wage-law question faced a fiduciary question about where the money would go.

Section 902(f) of the act answered the first layer with unusual directness. It added a new section 514(e) to ERISA, codified at 29 U.S.C. section 1144(e), providing that ERISA supersedes any state law that would directly or indirectly prohibit or restrict the inclusion of an automatic contribution arrangement in an ERISA-covered plan. The preemption was effective on the date of enactment, August 17, 2006, which the Joint Committee on Taxation’s technical explanation confirmed while adding that no inference was intended about the effect of conflicting state regulations before that date. A 2008 Department of Labor advisory opinion later clarified that the preemption was not limited to arrangements meeting the act’s qualified safe harbor designs; it covered automatic contribution arrangements in individual account pension plans generally. The legal doubt was not resolved by a court test or a clever reading. It was resolved the way legislatures resolve things: by a sentence declaring that the state laws in question do not apply.

That sentence deserves attention for what it reveals about the statute’s method. Congress did not mandate automatic enrollment, and this article repeats that point because the misconception is persistent. It removed a legal obstacle and then surrounded the newly cleared ground with incentives, a pattern the defined contribution half repeats at every step. The preemption is the negative liberty of the regime: the state may not stop the employer from enrolling workers automatically. Everything else in this half is the positive architecture: reasons for the employer to do it, and protections for the employer that does.

The Qualified Default Investment Alternative: Protection for the Silent Participant’s Money

With enrollment cleared, the fiduciary question remained, and it was the harder one. A plan sponsor that automatically enrolls workers must invest the accounts of people who have chosen nothing, and the research available by 2006 had documented exactly what happened when sponsors made the legally cautious choice. Madrian and Shea’s study of automatic enrollment found that the default fund allocation, typically a money market fund, anchored the investment behavior of automatically enrolled workers, producing substantially more conservative portfolios dominated by cash-like holdings rather than stocks. The cautious default was not neutral. It was a choice with consequences, quietly steering the least engaged workers, who were disproportionately younger and lower paid, into the lowest-returning asset available to them.

The act’s answer was the qualified default investment alternative, the QDIA, and it worked by extending the logic of participant direction to participants who had never directed. The statute amended ERISA section 404(c) to provide that an automatically enrolled participant would be treated as exercising control over the investment of the account when the default investment was consistent with Department of Labor regulations, which meant the fiduciary would not be responsible for losses associated with investing the account in accordance with that guidance. The mechanism is worth stating precisely because it is easy to misread: the law did not bless any particular fund. It created a conditional shield. Meet the conditions, and the fiduciary is protected from liability for the investment outcome of a choice the participant never made.

The Department of Labor, directed by the act to issue regulations, issued the final regulation on October 24, 2007, effective December 24, 2007, codified at 29 C.F.R. section 2550.404c-5. The regulation defined three types of long-term investment that could serve as a QDIA: a target-date or life-cycle fund keyed to the participant’s age or expected retirement date, a balanced fund mixing equity and fixed income exposure, and a professionally managed account allocating assets among the plan’s investment options. The three categories encode three theories of the silent participant. The target-date fund assumes the worker’s age is the best available proxy for risk tolerance, and automates the glide path from growth to safety without further decisions. The balanced fund assumes a single moderate allocation is defensible for almost anyone, and trades personalization for simplicity. The managed account assumes the worker’s situation deserves individual attention, and delegates the allocation to a professional working with the participant’s data. A sponsor choosing among them is choosing a theory of its workforce, and the regulation’s pluralism is deliberate: the statute did not know which theory was right, so it protected all three and let sponsors, advised by their fiduciaries, decide. What the regulation conspicuously did not bless was the old cautious default: money market and stable value funds do not qualify as QDIAs, a deliberate exclusion that pushed plan sponsors away from parking silent participants in cash. Investments made in such funds before the regulation’s effective date were grandfathered with transitional relief, and the regulation allowed a narrow 120-day capital preservation option for the period when newly enrolled workers were most likely to opt out, but the long-term message was unmistakable. The protected default was a diversified, age-appropriate investment, not a mattress.

The conditions a plan must satisfy to earn the shield are where the statute’s care shows, because each condition answers a specific abuse. The participant must have had the opportunity to direct the investment of the account and must have failed to do so, which preserves the participant’s primacy: the default only operates in the silence it is designed for. The plan must furnish a notice with specified information about the QDIA in advance of the first investment, generally thirty days before, and annually thereafter, so the silent participant is not kept silent by ignorance of the right to speak. Participants must be able to direct investments out of the QDIA as frequently as out of other plan investments, and at least quarterly, which keeps the default from becoming a trap. The plan must offer a broad range of investment alternatives no more expansive than what section 404(c) already requires, so the default exists inside a genuine menu rather than replacing one. And the QDIA itself must be managed by an investment manager, trustee, or plan sponsor acting as a named fiduciary, or be a registered investment company or equivalent regulated product, which keeps the selection inside the fiduciary system even as the liability for the outcome is lifted. The Department’s guidance stressed that choosing the QDIA remained a fiduciary act requiring an objective, thorough process with attention to competing providers and to fees and expenses. The shield covered the investment’s performance, not the prudence of selecting it. The QDIA relief and the automatic enrollment provisions applied generally for plan years beginning after December 31, 2007, while the preemption took effect on enactment.

The practical consequence was a migration of defaults that the regulation’s drafters clearly intended. Plan sponsors that had parked defaulted contributions in money market funds to avoid lawsuits now had a reason to move them into target-date funds instead, because the regulation named exactly the kind of fund that made the legally safe choice and the financially sensible choice the same choice. The act did not require any plan to offer target-date funds, and it did not require any participant to hold one. It made the target-date fund the default that a prudent fiduciary could select without fear, and then let inertia do the rest. The research on what defaults do to silent participants, which a later section examines in full, predicted the rest of the story: most of the automatically enrolled would stay where they were put.

How did the escalating default solve the adequacy problem?

By making the default contribution rate rise over time instead of sitting flat. The qualified automatic contribution arrangement starts workers at a minimum 3% default, escalating at least one point a year toward 6% and beyond, letting the inertia that kept workers enrolled keep their savings growing. A flat default anchored saving at 3%; a rising one anchors it upward.

Automatic enrollment solved the participation problem and immediately exposed a subtler one, and the subtlety had been documented before Congress acted. When Madrian and Shea studied the large employer that switched to automatic enrollment, they found not only that participation rose but that the default contribution rate, set at 3 percent, became the contribution rate for a large share of automatically enrolled workers. The distribution of contribution rates shifted dramatically away from 6 percent and higher levels, where it had sat under the old opt-in regime, toward exactly 3 percent. Workers who would have chosen higher rates under the old system, including workers chasing a full employer match, ended up saving less because the default anchored them. The default giveth participation and the default taketh away adequacy. A statute that stopped at automatic enrollment would have enrolled millions of workers into under-saving.

The escalating default was the answer, and its intellectual source was the Save More Tomorrow program designed by Shlomo Benartzi and Richard Thaler, which asked workers to commit in advance to directing a portion of future pay raises into their retirement accounts. The design exploited two findings at once: people resist cuts to their current take-home pay because of loss aversion, but they will commit future raises they have not yet received, and once committed, inertia keeps them in. In the first implementation at a midsize manufacturing company, participating workers raised their savings rates from 3.5 percent of income to 13.6 percent by the fourth pay raise after joining, while the average across the site rose from 4.4 percent to 10.6 percent. Congress did not enact Save More Tomorrow by name. It enacted the mechanism, which is the more durable form of borrowing.

The vehicle was the qualified automatic contribution arrangement, the QACA, added to the tax code as section 401(k)(13). The design is a bargain stated in numbers. The plan must apply a uniform default deferral percentage to all eligible employees: at least 3 percent of compensation in the first year of participation, rising by at least one percentage point in each of the next three years to reach 6 percent in the fourth year and beyond, with a ceiling of 10 percent. The employer, in exchange for adopting this escalating default, must make either matching contributions at a prescribed formula, 100 percent of the first 1 percent of compensation deferred plus 50 percent of the next 5 percent, or a nonelective contribution of 3 percent of compensation for eligible employees. Those employer contributions must vest fully within two years of service, faster than many plans had previously required, and the plan must give employees the required notices about the arrangement. The payoff for the employer is the one the statute dangles most visibly: a plan with a qualified automatic contribution arrangement is treated as satisfying the actual deferral percentage and actual contribution percentage nondiscrimination tests, and the top-heavy rules do not apply to a plan consisting solely of such contributions. Testing relief is the currency in which Congress bought employer cooperation, and it bought a great deal of it.

Two companion features completed the arrangement. The act recognized the eligible automatic contribution arrangement, under which employees could withdraw their automatic contributions, with earnings, within 90 days of the first automatic withholding, a pressure valve for workers who discovered the enrollment only on their pay stub and wanted out without the usual distribution machinery and without the usual 10 percent early-withdrawal penalty. The eligible arrangement also gave plans a longer window, six months rather than two and a half, to distribute excess contributions without triggering the 10 percent excise tax. And the broader automatic contribution arrangement, even without the qualified safe harbor’s testing relief, still benefited from the preemption of state wage laws, which meant employers could adopt simpler auto-enrollment designs without paying the full price of the QACA’s employer contributions. The architecture offered a ladder: any automatic enrollment got legal clearance, the qualified version got testing relief in exchange for escalation and employer money. The ladder matters because employers are heterogeneous. A large employer with sophisticated benefits counsel might adopt the full QACA for the testing relief; a smaller employer wary of the matching cost might choose the EACA for its simplicity and keep running the annual tests; an employer testing the waters might start with a plain automatic contribution arrangement under the preemption umbrella alone. The statute did not pick a winner among these designs. It priced each rung and let sponsors climb to the level their economics supported, which is why adoption spread across firm sizes instead of clustering among the largest plans. The escalating schedule is the detail that shows the statute had absorbed the research. A flat default anchors saving at the default. A rising default uses the same anchoring force in the opposite direction, and the 10 percent ceiling keeps the mechanism from becoming confiscatory by inertia.

The act’s design story is usually told as three safe harbors, preemption, the QDIA, and the automatic contribution arrangement, but the statute created the last of these in two forms, the EACA and the QACA, each with its own conditions. The safe harbor table:

Safe harbor created by the act	The legal risk it removed	Conditions a plan must meet	The behavioral finding it implements
ERISA section 514(e) preemption of state wage-withholding laws for automatic contribution arrangements	State-law liability for deducting retirement contributions from pay without an affirmative election	The arrangement must be an automatic contribution arrangement in an ERISA-covered individual account plan	Automatic enrollment dominates participation, so the legal doubt suppressing it had to be removed first
Qualified default investment alternative under ERISA section 404(c)(5) and 29 CFR 2550.404c-5	Fiduciary liability for the investment performance of a defaulted participant’s account	Default into a qualifying diversified investment, advance and annual notice, quarterly exit rights, prudent selection and monitoring of the QDIA	Silent participants stick with the default fund, so the default must be diversified rather than cash
Eligible automatic contribution arrangement under IRC section 414(w)	Testing and distribution complications for uniform automatic enrollment	Uniform default percentage for all covered employees, annual notice, 90-day penalty-free permissible withdrawals	Workers accept automatic enrollment when the exit stays easy, so participation gains do not require lock-in
Qualified automatic contribution arrangement under IRC sections 401(k)(13) and 401(m)(12)	Annual ADP and ACP nondiscrimination testing and the top-heavy rules	Minimum escalating default rates starting at no less than 3 percent and rising at least one point a year to no less than 6 percent with a 10 percent ceiling, required employer matching or nonelective contributions, annual notice	Escalating defaults beat flat defaults because workers accept gradual increases they would reject as a single deduction, and inertia sustains saving

The Testing Relief Bargain

The QACA’s testing relief is the provision sponsors cared about most, and it repays explanation, because nondiscrimination testing is the tax code’s oldest bargain with employers and the reason automatic enrollment needed a safe harbor at all. The bargain works like this. The tax code gives retirement plans generous treatment, deductible employer contributions, tax-deferred growth, on the condition that the benefits do not skew too heavily toward the highly paid. To police the condition, the code requires plans to run annual tests comparing the deferral and contribution rates of highly compensated employees against those of everyone else. A plan that fails the tests must return excess contributions to the highly paid or make corrective contributions to the rest, an expensive and embarrassing outcome that benefits counsel are paid to avoid.

Automatic enrollment threatened this bargain from both directions. A sponsor considering automatic enrollment worried that defaulting lower-paid workers at 3 percent might not lift their participation enough to keep the tests passing, while the highly paid continued to max out their deferrals. The rational response for a risk-averse sponsor was to skip automatic enrollment rather than gamble on the test results. The QACA broke the deadlock by changing the terms of the bargain: adopt the escalating default, fund the required employer contributions, give the notices, and the plan is deemed to satisfy the tests. The statute traded certainty for generosity. The sponsor gave workers an escalating default plus employer money; the code gave the sponsor freedom from the annual testing lottery.

The top-heavy relief worked the same way on a second test. The top-heavy rules impose minimum contributions when a plan’s benefits concentrate among key employees, a backstop against the most lopsided designs. A plan consisting solely of QACA contributions is exempt, which removed the last testing obstacle for the sponsor that went all in on the escalating default. Notice what the statute did not do: it did not lower the testing standards or excuse discrimination. It offered a different way to satisfy them, one measured in default rates and employer contributions rather than in annual comparisons. A sponsor that wanted the old bargain could keep running the tests under an EACA or a plain automatic contribution arrangement. A sponsor that wanted certainty could buy it with the QACA’s escalating default. The testing title of the act is thus the clearest statement of its method: where the defined benefit half commanded, the defined contribution half bargained, and the bargaining proved the more durable strategy.

Making the Saver’s Credit Permanent

The fourth step in the defined contribution half was the least behavioral and the most straightforward, and it addressed the population that defaults alone could not reach. The retirement savings contributions credit, commonly called the saver’s credit, had been created by the Economic Growth and Tax Relief Reconciliation Act of 2001 as a temporary provision: a nonrefundable tax credit for low- and moderate-income taxpayers who contributed to retirement accounts, effective for tax years 2002 through 2006, scheduled to expire just as the rest of the act’s automatic savings machinery was switching on. Section 812 of the Pension Protection Act made the credit permanent, and section 833 indexed its income thresholds to inflation in 500 dollar increments beginning in 2007, after five years in which the brackets had been fixed.

The credit’s mechanics explain both its purpose and its limits. Eligible taxpayers, at least 18 years old, not claimed as dependents, not full-time students, and within the income thresholds, could claim a credit of up to 1,000 dollars per person, 2,000 dollars for a married couple filing jointly, calculated as a percentage of up to 2,000 dollars in contributions to IRAs and employer plans. The credit rate declined as income rose, from 50 percent at the lowest tier down through 20 percent and 10 percent to zero. To see the cliff structure concretely, the Congressional Research Service’s illustration for 2022 joint returns put the 50 percent credit at adjusted gross income of $41,000 or less, the 20 percent credit above $41,000, the 10 percent credit above $44,000, and no credit above $68,000. The Pension Protection Act also indexed the income thresholds to inflation beginning in 2007, after five years in which the brackets had been fixed, which kept the credit’s reach from eroding as nominal incomes rose. The Congressional Research Service’s later summaries put the design’s central tension plainly: the credit is nonrefundable, meaning it cannot exceed the taxpayer’s income tax liability, and low-income taxpayers typically owe little or no income tax. The very workers the credit was meant to encourage, the ones with the least capacity to save and the least tax liability to offset, were the ones least able to use it. Automatic enrollment reached those workers through the payroll; the credit reached them, if at all, through the tax return, and the two instruments never quite met. How private saving interacts with the public floor is traced in the series companion on Social Security’s poverty impact (/2012/09/15/social-security-poverty-impact/). In 2022, the SECURE 2.0 Act replaced the credit with a saver’s match scheduled to begin in 2027, an explicit acknowledgment that a nonrefundable credit was the wrong tool for the lowest earners. That later correction belongs to a later Congress, but it is worth noting here because it marks the boundary of what the 2006 act’s half could do: it could change defaults for everyone on a payroll, and it could preserve a tax incentive, but it could not make a nonrefundable credit refundable, and the lowest-paid workers remained the hardest to reach.

What the Act Did Not Do

The misconceptions about the Pension Protection Act are as instructive as its provisions, because each one reveals a plausible but wrong theory of how the statute worked. The first and most persistent is that the act mandated automatic enrollment. It did not. The statute’s automatic enrollment provisions were optional throughout: an employer could adopt them, decline them, or adopt them in the simpler EACA form without the QACA’s testing relief. President Bush’s signing statement said the act “encourages” automatic enrollment, and the encouragement framing was deliberate. The federal mandate for automatic enrollment in newly established 401(k) plans did not arrive until SECURE 2.0 in 2022, sixteen years later, and it arrived precisely because the 2006 act’s invitation had proven the concept without compelling it. Anyone describing the 2006 act as a mandate is confusing the invitation with the later requirement.

The second misconception is that the act required target-date funds. It required nothing of the kind. The QDIA regulation blessed three categories of long-term investment, target-date or life-cycle funds, balanced funds, and professionally managed accounts, and deliberately excluded money market and stable value funds from the protected default. Target-date funds became the dominant QDIA because they required the least ongoing decision-making from the sponsor, not because the statute named them. A plan that defaulted silent participants into a balanced fund or a managed account earned the same fiduciary shield. The market chose target-date funds; the law merely made the choice safe.

The third misconception runs in the opposite direction: that automatic enrollment was already lawful everywhere before 2006, so the preemption was theater. The legal doubt was real. Benefits advisors had asserted that state wage-withholding statutes could forbid automatic deductions outright, and sponsors with multistate workforces faced a patchwork in which the same enrollment feature might have been lawful for one group of employees and a wage-law violation for another. The Internal Revenue Service’s favorable guidance addressed the federal tax treatment but could not silence the state-law argument, and the Joint Committee on Taxation’s technical explanation pointedly added that no inference was intended about the effect of conflicting state regulations before the enactment date. A legal risk does not have to materialize in a lawsuit to shape behavior; it only has to be plausible enough that general counsel says no. The preemption mattered because it turned a plausible risk into a dead letter.

The fourth misconception is the premium misattribution, and it is the one this article has corrected at every turn because it is the easiest to get wrong. The flat-rate premium increase from $19 to $30 per participant and the creation of the $1,250 termination premium belonged to the Deficit Reduction Act of 2005, effective for the 2006 plan year. The Pension Protection Act made the termination premium permanent, redefined the variable-rate base, and capped small-employer premiums. A premium history that assigns the 2006 increases to the 2006 act misattributes costs to this statute and, worse, misreads its method: the act priced the insurance more accurately rather than raising its headline rate, which is a subtler and more characteristic move.

The Compliance Calendar the Act Created

A statute built on conditional safe harbors creates a compliance calendar, because every shield in the act is earned annually rather than granted once. The QDIA conditions require the plan to furnish the default-investment notice in advance of the first investment and annually thereafter, which means the notice is not a one-time disclosure but a recurring obligation with a deadline the plan must hit every year. The automatic contribution arrangements require their own annual notices describing the default rate, the escalation schedule, and the worker’s rights. The EACA’s 90-day withdrawal valve requires administration: contributions must be tracked from the first automatic withholding, withdrawal requests honored within the window, and the amounts distributed with earnings. The quarterly exit rights from the QDIA must be operational, not merely promised, which means the recordkeeper’s systems must actually permit the transfers. And on the defined benefit side, the funding target must be certified, the shortfall amortized, the thresholds monitored, and the benefit restrictions applied the moment a plan crosses below 80 or 60 percent funded.

This calendar is the hidden cost of the act’s method, and it falls unevenly. A large employer with a benefits department absorbs the notice cycle as routine; a small employer without one experiences it as a recurring burden that the safe harbors must be valuable enough to justify. The statute’s designers understood this, which is why the EACA exists as the simpler rung on the ladder and why the small-employer premium cap shields the smallest plans. But the general point stands: a default architecture is not free to operate. It requires the plan to keep telling workers what the default is, keep giving them the chance to leave, and keep certifying that the conditions hold. The notices are the price of the shield, and the shield is only as good as the calendar behind it. On the defined benefit side, the calendar is actuarial rather than administrative: the funding target must be certified each year, the seven-year amortization schedule recalculated as markets move, and the 80 and 60 percent thresholds monitored continuously, because the benefit restrictions bite the moment a valuation crosses them. A plan that drifts below 80 percent funded between annual valuations can discover the restriction only when the next certification lands, which is why sponsors learned to watch their funded status the way they watch their cash position. The act made funding a living number rather than an annual ritual.

The Worker’s Journey Through the Defaults

The provisions interlock, and the interlock is easiest to see by following a single worker through them. She is hired and, because her employer adopted a qualified automatic contribution arrangement, she is enrolled at a 3 percent default deferral rate unless she opts out. The preemption of state wage laws is what makes the payroll deduction lawful; without it, the enrollment could not have happened. She receives the required notices describing the default rate, the default investment, and her right to choose differently; without them, the safe harbors would not protect the plan. She does nothing, as most workers do, and her contributions flow into the plan’s qualified default investment alternative, a target-date fund keyed to her expected retirement year. The QDIA safe harbor is what lets the fiduciary place her money there without fearing liability for the market’s verdict. Each year, her default rate steps upward by at least a point, the Save More Tomorrow mechanism rendered as a plan term, and she does not opt out, as the research predicted most workers would not. Her employer’s matching contributions vest within two years, and the plan skips the annual nondiscrimination testing that the QACA’s bargain bought off. If she is a lower earner with tax liability, the saver’s credit, now permanent, improves the economics of her contributions at filing time.

Every step of that journey is a default, and every default is a provision of the act. The enrollment is the preemption plus the arrangement. The investment is the QDIA. The rising rate is the escalator. The notices are the consent architecture. The testing relief is the employer’s compensation. The credit is the low-income supplement. Remove any one provision and the journey breaks at that step: without preemption, the deduction is doubtful; without the QDIA, the money sits in cash; without the escalator, the rate anchors at 3 percent; without the notices, the shield fails; without the testing relief, the employer never adopts the arrangement. The statute’s genius is not in any single provision but in the completeness of the chain, each link addressing the exact failure the research had documented at that point in the worker’s experience. This is what it means to write behavioral findings into law: not a nudge here and a disclosure there, but an end-to-end redesign of what happens when a worker does nothing, tested against the evidence at every joint.

The Adoption Arc: What the Defaults Built

The statute’s consequence can be measured, and the measurement replaces rhetoric with records. Start with the plan level. An analysis of 2023 Form 5500 filings covering roughly 720,000 plans found that among large 401(k) plans, the share offering target-date funds rose from 32 percent in 2006 to 90 percent in 2023, while the share of large-plan assets held in target-date funds rose from 3 percent to 33 percent over the same period. The default the regulation blessed became the default the market adopted, on a timetable that begins the year the act was signed.

Move to the participant level and the scale sharpens. The Investment Company Institute’s 2023 Fact Book counted 60 million active 401(k) participants holding $6.6 trillion in assets at the end of 2022, and 59 percent of them were invested in target-date funds. These are not projections or claims about what defaults might do. They are counts of what the defaults did.

The enrollment side of the arc shows the same pattern. The same Form 5500 analysis found that 64.1 percent of plans with $1 billion or more in assets used automatic enrollment. A field study by Thaler and Benartzi published in Science in 2013 found that adoption of both automatic enrollment and automatic escalation had climbed steeply since 2005. The Department of Labor’s own preamble to the QDIA rulemaking had projected that plans implementing automatic enrollment could raise participation rates on average from approximately 70 percent to perhaps 90 percent, with the largest gains among the lower-paid, younger, and shorter-tenure workers whose participation rates were otherwise the lowest. The projections and the outcomes point the same direction.

Two cautions keep the arc honest. First, none of this was required. The act never mandated automatic enrollment and never required target-date funds; every figure above describes choices that sponsors made once the law made those choices safe. The adoption arc is therefore evidence for the statute’s method, not for compulsion: remove the legal risk, and private actors will walk through the door that the behavioral research pointed to. Second, the arc measures participation and allocation, not adequacy. A worker defaulted into a plan at 3 percent and left there is participating and still under-saving, which is why the escalating default exists and why the adequacy question outlives the participation victory. The statute solved the problem the research had fully specified and left the harder problem, getting every defaulted worker to a sufficient rate, to the escalator, the match, and the worker’s own attention.

There is a third limitation the numbers imply but do not state. The adoption arc describes the workers inside the system, the 60 million participants in plans that adopted the defaults. It says nothing about the workers outside it: employees of firms that never adopted automatic enrollment, part-time and contingent workers excluded from plan eligibility, and the self-employed, for whom no employer default can operate. The statute’s method, changing what happens when a worker does nothing, presupposes a worker with a plan and a payroll. For everyone beyond that boundary, the defaults do not reach, which is why the act’s achievement, large as it is, is bounded by the reach of employer-sponsored coverage itself. The later mandate in SECURE 2.0 extended the defaults to newly established plans; it did not extend them to workers without plans.

What did the research on 401(k) defaults actually find before Congress acted?

Madrian and Shea’s 2001 study of one employer found that automatic enrollment raised participation sharply and that the default contribution rate and default fund choice anchored most workers’ behavior, even when better options existed. Thaler and Benartzi’s 2004 Save More Tomorrow experiment then showed that commitments to escalate contributions with future raises lifted savings rates without reducing take-home pay.

The two studies entered the legislative process the way good empirical work sometimes does: not as decoration but as design specifications. Brigitte Madrian and Dennis Shea’s paper, “The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior,” published in the Quarterly Journal of Economics in 2001, analyzed the 401(k) behavior of employees at a large American corporation before and after the company switched from requiring an affirmative election to automatic enrollment. Nothing about the plan’s economics changed. The match, the investment menu, the tax treatment, all identical. What changed was the default, and the authors reported two findings that traveled directly into the statute. First, participation was significantly higher under automatic enrollment, rising by as much as 40 percentage points in the studied firm, a gain large enough to make the default itself the single biggest determinant of whether a worker participated at all. The magnitude of that gain is the number that traveled from the journal article into the statute’s design: a default is not a suggestion that workers weigh alongside the alternatives but the outcome that most workers simply keep, which is why the act’s drafters treated the choice of default as the central policy decision rather than a detail. Second, the default contribution rate and the default investment allocation chosen by the company exerted a strong influence on the savings behavior of participants, with a substantial fraction of workers hired under automatic enrollment sticking to both the default rate and the default fund even though almost nobody hired before the change had chosen that combination. The authors attributed the pattern to inertia and to workers treating the default as implicit investment advice from the employer. Their conclusion named the mechanism the act would later codify: large changes in savings behavior could be motivated, in their phrase, by the power of suggestion. The Department of Labor’s own literature review added the distributional detail that mattered most for policy: the participation gains were largest among the groups that had participated least under the opt-in regime, younger workers, lower-paid workers, and workers with shorter tenure, with participation differences still around 30 percentage points at 48 months of tenure. Defaults did not merely raise the average. They reached the workers the old system had missed.

Richard Thaler and Shlomo Benartzi’s contribution addressed the inadequacy problem the first study exposed. Their paper, “Save More Tomorrow: Using Behavioral Economics to Increase Employee Saving,” published in the Journal of Political Economy in 2004, reported on the SMarT plan, under which workers committed in advance to allocate part of each future pay raise to their 401(k) contributions. The design’s psychological logic was explicit: by committing future raises rather than current pay, workers never experienced a reduction in nominal take-home pay, which neutralized loss aversion, and by making the commitment in advance, the plan sidestepped the procrastination that kept workers from raising their rates on their own. At the first company to adopt it, a midsize manufacturer struggling to get lower-paid workers to participate at meaningful levels, the workers who joined the plan saw their savings rates rise from 3.5 percent of income to 13.6 percent by the fourth pay raise, and the site’s average savings rate rose from 4.4 percent to 10.6 percent. Thaler described the program in 2004 Senate testimony as a complement to automatic enrollment, one that used the same inertia in the opposite direction. The statutory fingerprints are visible in the QACA’s escalating schedule: the 3 percent start, the annual step-ups, the 10 percent ceiling, all translate the SMarT insight that defaults should rise over time into a safe harbor an employer can adopt off the shelf.

A fair account of this foundation includes the complication the research itself supplied. Madrian and Shea’s data showed that automatic enrollment, for all its participation gains, reduced the average contribution rate in the studied firm, because workers who would have elected 6 percent or more under the opt-in regime settled at the 3 percent default. Defaults dominate, and domination cuts both ways. The act’s drafters evidently read that finding, because the escalating default exists to solve precisely the problem the first generation of automatic enrollment created. This is the honest version of the behavioral economics story: not that nudges are magic, but that the statute’s designers had evidence about how defaults fail as well as how they succeed, and built the failure mode into the remedy.

The consequence the brief names followed from this architecture with a logic close to inevitability. Once the QDIA regulation excluded money market and stable value funds from the protected default while blessing target-date funds, balanced funds, and managed accounts, plan sponsors seeking the fiduciary shield had a short list, and the target-date fund was the item on it that required the least ongoing decision-making from the sponsor. A target-date fund adjusts its own asset allocation as the participant ages, which made it the natural answer to the question the QDIA posed: what single investment can a fiduciary defend as appropriate for a silent participant of any age? The result, over the years after the act, was the target-date fund’s ascent to the default holding of American retirement saving, the place where the money of workers who never chose anything ended up. The act never required target-date funds, and no plan was obligated to offer one. It made them the safest default to select, and in a system where most automatically enrolled participants never override the default, the safest default becomes the most common holding. The larger account-based shift that this act accelerated, from employer promises to individual accounts, is traced in the series companion on that transition (/2012/10/15/pension-vs-401k-law-shift/).

Translating Research into Statutory Parameters

It is worth pausing on how literally the statute translated the research, because the translation is where the behavioral economics claim earns its precision. Each parameter of the qualified automatic contribution arrangement maps to a finding. The 3 percent starting default is low enough that few workers opt out on impact, which answers loss aversion at the moment of enrollment. The annual step-up of at least one percentage point implements the Save More Tomorrow commitment device, raising the rate out of future compensation growth rather than current take-home pay. The 6 percent floor on the escalated rate and the 10 percent ceiling bracket the mechanism: high enough to matter for adequacy, low enough that inertia cannot become confiscation. The required employer match or nonelective contribution is the price Congress charged sponsors for testing relief, and the two-year vesting schedule is faster than many plans had previously required, which tilts the bargain toward the worker. The vesting term deserves notice because it is the one QACA parameter that is not behavioral at all. Faster vesting does not change what any worker does; it changes what the worker keeps when she leaves, which addresses the oldest complaint about employer contributions, that they function as golden handcuffs whose value evaporates with turnover. By capping the forfeiture period at two years, the statute ensured that the employer money buying the testing relief actually reached the workers whose inertia justified the relief. None of these numbers fell from the sky. Each one is a research finding with a section number, except the vesting rule, which is a fairness finding with a constituency.

The translation extended to the consent architecture as well, and this is the part of the design that answers the paternalism objection before it is raised. A default that cannot be escaped is not a nudge but a mandate wearing a costume, and the statute keeps the costume honest through exits at every stage. The QDIA conditions require advance notice and quarterly exit rights, so the silent participant is never locked into the default. The eligible automatic contribution arrangement gives newly enrolled workers a 90-day window to withdraw automatic contributions with earnings and without the usual early-withdrawal penalty, a pressure valve for the worker who discovers the enrollment on a pay stub. The Department of Labor’s guidance stressed that selecting the QDIA remained a fiduciary act requiring an objective process with attention to fees, so the liability shield never became a license for indifference. The research said defaults dominate because workers treat them as implicit advice and because inertia does the rest; the statute’s answer was to make the advice good, the inertia productive, and the door out clearly marked.

The distributional finding deserves its own emphasis because it is the moral center of the design. The workers least served by the opt-in regime, younger, lower-paid, shorter-tenured, were the workers most transformed by the default regime. Automatic enrollment did not merely lift participation in the aggregate; it compressed the participation gap between the workers who had always saved and the workers who had never gotten around to it. The saver’s credit, made permanent in the same act, was supposed to reinforce that compression from the tax side, and its nonrefundable structure limited how far it could reach. But the default architecture needed no tax liability to work. It operated through the payroll, which every worker touches, rather than through the tax return, which the lowest earners barely do. That asymmetry, payroll over tax return, is the quiet reason the defined contribution half of the act reached further than its own tax incentive.

Why Defaults Dominate: The Psychology the Statute Used

The statute’s behavioral claim rests on four psychological mechanisms, and naming them shows why the drafters were confident enough to write them into law. The first is inertia, the simple tendency to leave things as they are. Madrian and Shea found that a substantial fraction of automatically enrolled workers stuck with both the default contribution rate and the default fund, not because they had evaluated the options and endorsed the default, but because evaluating options takes effort and the default required none. Inertia is the engine of every default in the act: enrollment, investment, and escalation all assume that most workers will not act, and design the outcome of inaction to be the right one.

The second mechanism is implicit advice. Workers treat the employer’s default as a recommendation, reasoning, sensibly enough, that the company that chose the default probably chose it well. This is why the default contribution rate anchored even workers who would have chosen higher rates under the opt-in regime: the 3 percent default read as the employer’s suggestion of the right number. The mechanism cuts both ways, which is why the QDIA conditions require the default to be a diversified, age-appropriate investment selected through a prudent process. If workers will treat the default as advice, the law must ensure the advice is good.

The third mechanism is loss aversion, the finding that losses loom larger than equivalent gains. A worker asked to raise her contribution rate today experiences the increase as a cut to take-home pay and refuses; the same worker asked to commit a future raise she has not yet received experiences no loss at all. The escalating default is loss aversion rendered as a plan term: the rate rises, but nominal take-home pay never falls, because each increase is timed to a raise. Thaler and Benartzi’s insight was that the timing of the increase matters more than its size, and the QACA’s annual step-ups are that insight with a section number.

The fourth mechanism is procrastination, the gap between intention and action. Surveys consistently show that workers intend to save more than they do; the intention is real and the follow-through is not. Automatic enrollment closes the gap by making the intended outcome the default outcome, and the 90-day withdrawal valve in the EACA acknowledges the mechanism’s shadow: a worker who discovers the enrollment on a pay stub and wants out gets a clean exit, because a default that trapped the procrastinator would be a mandate. The statute’s designers understood that each mechanism needed both an application and a safeguard, which is why every default in the act arrives with a notice, an exit, or a ceiling attached.

The Name Against the Effect

There is a final complication worth stating plainly, because the statute’s name invites a simpler story than the statute’s effects support. It is called the Pension Protection Act, and its most durable achievement was to entrench the account-based system that replaced pensions. The defined benefit half of the act, the half that actually protected pensions in the traditional sense, produced stricter funding, changes to the Pension Benefit Guaranty Corporation premium structure, and a plausible acceleration of the freeze. The defined contribution half, the half that made automatic saving safe, produced the default architecture under which some 60 million active 401(k) participants, the Investment Company Institute’s end-2022 count, 59 percent of them in target-date funds, accumulate retirement wealth in individual accounts, in target-date funds they never chose, at contribution rates they never set.

This is worth stating without treating it as hypocrisy, because hypocrisy implies that the drafters said one thing and meant another, and there is no evidence of that. The drafters faced a world in which the traditional pension was already receding and the account-based system was already dominant, and they wrote a statute for the world they had rather than the world the name evoked. Protecting pensions, as 2006 understood the term, meant two things at once: making the remaining pension promises harder to break, and making the replacement system function for workers who would never see a pension at all. The first task was defensive and partial. The second was constructive and transformative. That the constructive half outlasted the defensive half is not a betrayal of the name. It is a measure of which half of American retirement was still growing.

There is also a less cynical reading available, and it deserves its sentence. A statute named for pension protection that succeeds in lifting participation and saving among the 60 million active 401(k) participants counted at the end of 2022 has protected pensions in the only sense that ultimately matters, which is the sense of retirement security rather than the sense of a particular benefit formula. The formula was a means. The security was the end. If the end is served better by defaults than by defined benefit promises, then the name is vindicated by the outcome even as the formula fades.

The Experiment the Statute Ran on Itself

Step back far enough and the Pension Protection Act looks like a controlled experiment that no researcher could have designed and no Congress would have run on purpose. Take two halves of one statute, enacted by the same Congress on the same day, under the same public law number. Give one half the method of command: fund to 100 percent, amortize over seven years, restrict the underfunded, price the insurance. Give the other half the method of defaults: preempt the legal doubt, shield the fiduciary, escalate the rate, bargain away the testing. Then watch for a decade and compare the outcomes.

The defined contribution half won decisively on its own terms. Participation rose, the default investment migrated from cash to diversified funds, the escalator lifted contribution rates, and the adoption arc in plan and participant data bent in the direction the research had predicted. The mechanism worked through private choice rather than against it: sponsors adopted the designs because the statute made them safe and rewarding, and workers stayed because inertia did the rest. No enforcement apparatus was required to produce these outcomes. The statute changed the path of least resistance and let gravity do the work.

The defined benefit half achieved its stated aim and paid the predicted price. It ended the forbearance: plans could no longer smooth, stretch, and sweeten their way around underfunding, and the funding measures became honest. But honesty had a cost, and the cost was counted in freezes. Whether the statute caused the freeze wave or merely priced a retreat already underway, the half that commanded presided over contraction while the half that defaulted presided over expansion. The asymmetry is the series’ recurring finding in its purest form, because here the comparison is not across different statutes in different eras but within a single enactment, holding the Congress, the year, and the political constraints constant.

There is a subtlety worth preserving, because the experiment does not prove that commands never work. The funding half’s commands worked exactly as designed at their primary task: they stopped plans from promising benefits they had not funded, which was the evil the title named. A command that ends forbearance is doing its job even if the activity it disciplines shrinks. The lesson is narrower and more useful: commands are good at stopping bad behavior and bad at producing good behavior, while defaults are good at producing good behavior and useless at stopping bad behavior. The 2006 act needed both because it faced both problems, a funding system that required stopping and a savings system that required starting. The statute’s enduring reputation rests on the starting, because starting is the harder problem and the one the law had never solved before.

Studying the Act: Defaults as Substantive Policy

Return to the statute’s method with the full picture in view, because the defined contribution half is the clearest illustration in this series of the amendment stage doing its characteristic work. The 2006 act did not create the 401(k), did not invent automatic enrollment, and did not discover the behavioral research. It amended an existing regime at the points where private ordering had stalled: the state-law doubt that kept sponsors from enrolling, the fiduciary exposure that kept defaults in cash, the testing rules that made enrollment expensive, the expiring credit that left low earners out. Each amendment changed what happened when the worker did nothing, and the worker doing nothing turned out to be the modal case. This is the recurring finding across the series stated in its sharpest form: changing procedure or defaults often outperforms changing commands, because commands must be obeyed and defaults merely need to be left alone.

The act’s namable claim states the principle in a single sentence: The default is the policy: the 2006 act changed retirement outcomes for tens of millions of workers without changing a single contribution requirement or tax rate, purely by changing what happens when a worker does nothing, which makes it the clearest proof in American law that default rules are substantive policy.

One caution closes the assessment, and it is owed to the reader who has followed the argument this far. Defaults are policy made without consent in the usual sense, and the statute’s protections, notices, quarterly exit rights, the 90-day withdrawal valve, exist precisely because the designers understood that. A default that cannot be escaped is not a nudge but a mandate wearing a costume, and the act’s architecture keeps the costume honest only so long as the exits stay open.

A later Congress eventually went further than 2006 dared. The SECURE 2.0 Act of 2022 added an automatic enrollment mandate for most newly established 401(k) plans, the step the Pension Protection Act had declined to take, converting the 2006 act’s invitation into a requirement for new plans. That evolution, and the further changes it set in motion, is taken up in the series companion on later retirement legislation (/2012/11/01/secure-act-retirement-law-changes/). For the 2006 act itself, the study questions are the ones the statute forces on any reader of legislation: which outcomes in this law came from commands, which came from defaults, and which came from removing the legal risks that had kept private actors from moving at all. Work through those questions against the statute’s two halves in the legislation study notebook, and the amendment stage of a statute’s life will never look like a footnote again.

Frequently Asked Questions

Q: What did the Pension Protection Act of 2006 do?

Public Law 109-280, signed August 17, 2006, is an amendment package that rewrote parts of both ERISA and the Internal Revenue Code. It has two halves that point in opposite directions. The defined benefit half tightened pension funding: a full funding target, shorter amortization of funding shortfalls, restrictions on benefit increases and lump-sum payouts by underfunded plans, and revised premiums paid to the Pension Benefit Guaranty Corporation, the insurer of last resort. The defined contribution half made automatic saving legally safe: it preempted state wage-withholding laws that had cast doubt on automatic enrollment, created a fiduciary safe harbor for qualified default investment alternatives, created a safe harbor for automatic contribution arrangements with escalating default rates, and made the low-income saver’s credit permanent. The statute’s intellectual foundation is behavioral research on defaults by Brigitte Madrian and Dennis Shea and on escalation by Richard Thaler and Shlomo Benartzi.

Q: How did the Pension Protection Act change 401(k) plans?

It made the 401(k)’s most important feature, automatic enrollment, practical instead of merely possible. Before 2006, enrolling workers who had never signed up was legally doubtful because state wage-withholding laws could be read to forbid deducting pay without an affirmative election; the act preempted those state laws for plans with automatic features. It then created two safe harbors: one protecting fiduciaries who default silent participants into a qualified default investment alternative rather than cash, and one for automatic contribution arrangements with escalating default contribution rates. It also made the saver’s credit permanent, improving the economics of saving for low-income workers. The lasting consequence is visible in 401(k) plans designed after the act: the target-date fund as the default holding for workers who never make an investment election. The act did not mandate automatic enrollment and did not require target-date funds; it removed the legal risk from using them.

Q: What is automatic enrollment under the Pension Protection Act?

Automatic enrollment is a plan design in which a worker is enrolled in the retirement plan and contributing at a default rate unless the worker affirmatively opts out or chooses a different rate. The Pension Protection Act did not invent or require it; it made it legally usable. The act preempted state wage-withholding laws that had made automatic payroll deductions legally doubtful, and it created safe harbors protecting plan fiduciaries that default workers into a qualified default investment alternative or use an automatic contribution arrangement with escalating rates. Research by Brigitte Madrian and Dennis Shea had shown that defaults dominate participation outcomes: when enrollment is automatic, participation rates are dramatically higher than when workers must sign up on their own. The statute wrote that finding into law by changing what happens when a worker does nothing.

Q: What is a Pension Protection Act qualified default investment alternative?

A qualified default investment alternative is a default investment that meets conditions set out in the statute and the Department of Labor regulation at 29 CFR 2550.404c-5, finalized October 24, 2007 and effective December 24, 2007, and that triggers a fiduciary safe harbor: when a plan defaults a silent participant into a QDIA, the fiduciary is protected from liability for the market performance of that investment. The logic is behavioral. A fiduciary who defaults an unengaged worker into cash protects the worker from market losses but condemns the account to near-zero real growth; the QDIA safe harbor lets the fiduciary choose a diversified investment instead. The act did not require target-date funds, but target-date funds became the dominant QDIA in practice, which is why they became the default holding of American retirement saving. A QDIA is therefore a legal category with liability consequences, not a product name, and the safe harbor covers the default decision rather than every aspect of plan management.

Q: Did the Pension Protection Act kill defined benefit pensions?

Not exactly, and the causal question is contested. The act’s defined benefit half raised the cost and volatility of sponsoring a traditional pension: a full funding target, shorter amortization of shortfalls, restrictions on underfunded plans, and revised premiums paid to the insurance corporation. Employers facing those rules accelerated the freeze and termination of defined benefit plans, and critics argue the statute hastened the disappearance of the pensions it was named for. Defenders respond that the freeze trend was already underway, driven by accounting changes and the structural shift toward account-based saving, and that stricter funding rules protected the benefits of workers whose plans remained. The brief’s own framing is that the most durable achievement of a law called a pension protection act was to entrench the account-based system that replaced pensions. The funding rules, the freeze trend and the competing explanations are facts; the causal claim is open.

Q: How did the Pension Protection Act tighten pension funding rules?

Before the act, employers could fund pension promises over long horizons and keep benefits growing even as funding deteriorated. The act moved plans to a full funding target, meaning the plan was expected to be funded against its full liability, and shortened the amortization period for funding shortfalls to seven years. It restricted benefit increases and lump-sum payouts by plans that were significantly underfunded, so a struggling plan could not dig a deeper hole while paying out large sums. It also revised the premiums plans pay to the Pension Benefit Guaranty Corporation, making permanent the termination premium created by the Deficit Reduction Act of 2005 and redefining the variable-rate premium base, strengthening the insurer that takes over failed plans. The stated aim was to stop plans from promising benefits they had not funded. The trade-off, which is the subject of the freeze debate, is that the same rules made sponsoring a defined benefit plan more expensive and its required contributions more volatile from year to year.

Q: Is the Pension Protection Act behavioral economics law?

Yes, in the most literal sense: it is the largest enactment of behavioral economics findings into American law. The statute’s design rests on two research programs named in the legislative record. Brigitte Madrian and Dennis Shea showed that automatic enrollment, enrolling workers unless they opt out, produces far higher participation than requiring an affirmative election. Richard Thaler and Shlomo Benartzi showed that escalating contribution rates, starting low and rising automatically, raise saving without triggering the loss aversion that makes workers reject large immediate deductions. The act translated both findings into safe harbors: legal protection for plans that default workers in and for arrangements that escalate their contributions. The result is the act’s central claim: the default is the policy. Outcomes changed for the 60 million active 401(k) participants counted by the Investment Company Institute at the end of 2022, 59 percent of them invested in target-date funds, without any change to contribution requirements or tax rates, purely by changing what happens when a worker does nothing.

Q: What did the Pension Protection Act do to PBGC premiums?

Not in the way the question assumes. The act made permanent the $1,250-per-participant termination premium created by the Deficit Reduction Act of 2005, redefined the variable-rate premium’s “unfunded vested benefits” base to match the new funding target, and capped variable premiums for small employers, while the flat-rate increase to $30 per participant (from $19) was enacted earlier by the Deficit Reduction Act of 2005, not this act. The logic was part of the funding half’s purpose: plans that promise benefits should pay for the insurance that backstops those promises, and a better-funded insurer is part of stopping plans from making promises they had not funded. The premium changes are also one of the cost increases that critics cite in the freeze debate, alongside the funding target and shorter amortization. Premium changes under this act are distinct from premium increases enacted in the Deficit Reduction Act of 2005 and in later budget legislation, so a premium history that mixes them together misattributes costs to this statute.

Q: Does the Pension Protection Act require employers to offer automatic enrollment?

No. This is one of the most common errors about the statute. The Pension Protection Act never required any employer to offer automatic enrollment. What it did was remove the legal risk from offering it: it preempted state wage-withholding laws that had made automatic payroll deductions legally doubtful, and it created safe harbors protecting fiduciaries that default workers into a qualified default investment alternative or use an automatic contribution arrangement with escalating rates. The act made automatic enrollment safe, not mandatory. A federal mandate for automatic enrollment in newly established 401(k) plans came later, with SECURE 2.0 (2022). The distinction matters for plan sponsors evaluating their obligations and for anyone describing the 2006 act’s role: it was an enabling statute built on behavioral research, not a command.

Q: What is the difference between an EACA and a QACA?

Both are automatic contribution arrangements created by the act, and they differ in what they buy the employer. An eligible automatic contribution arrangement (EACA) is the basic form: uniform automatic enrollment with notice requirements and a 90-day window in which a newly enrolled worker can withdraw automatic contributions without the usual 10 percent early-withdrawal penalty. A qualified automatic contribution arrangement (QACA) is the premium form: it satisfies the act’s safe harbor from annual nondiscrimination testing, but only if the plan uses minimum default rates that escalate over time (starting at no less than 3 percent, rising at least one point a year to no less than 6 percent, capped at 10 percent) and provides required employer matching or nonelective contributions. An employer that wants automatic enrollment without the cost and complexity of yearly testing chooses the QACA; an employer that wants flexibility and is willing to keep testing chooses the EACA. Both rest on the same behavioral finding: defaults dominate participation outcomes.

Q: Can an employee opt out of automatic enrollment?

Yes. Opting out is the defining feature of automatic enrollment: the worker is enrolled by default but may decline, change the contribution rate, or redirect the investment at any time. The act’s safe harbors are built around that choice; the fiduciary protection for defaulting a worker into a qualified default investment alternative assumes the worker was given notice and the opportunity to choose differently. Workers who do nothing end up contributing at the plan’s default rate into the plan’s default investment, which is exactly the behavioral mechanism the statute exploits. The practical advice is straightforward: check the enrollment notice when starting a job, because the default rate may be lower than what retirement planning requires, and an automatic escalation feature only helps if it is actually turned on.

Q: What default rate and escalation schedule does the safe harbor use?

The statute created the safe harbor for automatic contribution arrangements with escalating default rates but did not set a single mandatory rate; the detailed schedules live in the qualified automatic contribution arrangement rules. A QACA must start workers at a minimum default rate of 3 percent and escalate it at least one percentage point a year to no less than 6 percent, with a ceiling of 10 percent. The behavioral logic, drawn from the work of Richard Thaler and Shlomo Benartzi, is to start low so workers do not opt out and then raise the rate automatically, because workers accept gradual increases they would reject as a single large deduction. Employers designing a plan use the statutory minimums as a floor rather than a target: many plans set higher defaults because the research the act codified shows that low defaults can anchor workers at inadequate saving rates.

Q: Can a worker sue if a QDIA default loses money?

Not if the plan followed the rules. That is the point of the qualified default investment alternative safe harbor: when a plan defaults a silent participant into a QDIA and meets the conditions, including notice to the worker, the fiduciary is protected from liability for the investment’s market performance. The protection covers the decision to use the default, not everything else. It does not shield a fiduciary who imprudently selected the QDIA in the first place, ignored fees, or failed to monitor it. It also does not remove market risk; a diversified default can still lose money in a downturn. The safe harbor was designed to solve a specific dilemma: fiduciaries who defaulted unengaged workers into cash avoided market losses but guaranteed near-zero growth, while those who chose diversified investments faced lawsuits. The act resolved that dilemma in favor of diversification, provided the conditions are met.

Q: What notices must a plan give for automatic enrollment?

Plans using automatic features must give workers advance notice describing the default contribution rate, the default investment, and the right to opt out or choose differently. The notice is not paperwork for its own sake; it is a condition of the safe harbors. A fiduciary that defaults workers without proper notice may lose the liability protection the act created. For workers, the notice is the practical document: it states the default rate, whether escalation applies, and how to change the election. For plan sponsors, the notice calendar is one of the recurring compliance tasks the act created. Anyone evaluating whether a plan’s automatic enrollment is operating lawfully should start with whether timely, complete notices went out, because the statute makes notice the price of the safe harbor.

Q: Is a QDIA the same as a target-date fund?

No. A QDIA is a legal category; a target-date fund is a product that commonly fills it. The statute created a fiduciary safe harbor for defaulting silent participants into a qualified default investment alternative, defined by conditions such as diversification and professional management. Target-date funds, which shift from stocks to bonds as a target retirement year approaches, became the dominant QDIA in practice, and the act’s consequence is visible in their role as the default holding of American retirement saving. But a plan can satisfy the safe harbor with other diversified defaults such as balanced funds or managed accounts. The recurring error is believing the act required target-date funds; it required nothing of the kind. It made diversified defaults legally safe, and the market then chose target-date funds as the standard answer.

Q: Why did the Pension Protection Act preempt state wage laws?

Because automatic enrollment was legally doubtful without it. Before 2006, many state wage-payment laws could be read to forbid deducting retirement contributions from pay without the worker’s affirmative written election. An employer that enrolled workers automatically risked violating state law, which is why automatic enrollment was rare even though the research of Brigitte Madrian and Dennis Shea had shown it transformed participation. The act preempted those state laws for plans with automatic contribution features, making the payroll deduction lawful nationwide. Preemption is one of the three safe harbors in the act’s design story: it removed the legal risk, the QDIA safe harbor removed the fiduciary risk of the default investment, and the automatic contribution arrangement safe harbor removed the testing risk. Each safe harbor maps to a specific legal obstacle that had kept a behavioral finding from being used.

Q: What made automatic enrollment legally risky before 2006?

Two legal obstacles kept it rare. First, state wage-withholding laws in many states could be read to require an affirmative employee election before any payroll deduction, which made automatic deductions legally doubtful; the act solved this with federal preemption. Second, fiduciary risk: a plan that defaulted an unengaged worker into anything other than cash exposed itself to liability if the investment lost money, so the safe default was the money market fund, which protected the fiduciary while guaranteeing the worker near-zero real growth. The act solved this with the qualified default investment alternative safe harbor, protecting fiduciaries that default workers into diversified investments. The research case was already made: automatic enrollment studies had shown defaults dominate participation. What was missing was not evidence but legal safety, and the statute supplied it in the form of three safe harbors.

Q: What did the Pension Protection Act do to the saver’s credit?

It made it permanent. The saver’s credit, a nonrefundable tax credit for retirement contributions by low- and moderate-income workers, had been enacted as a temporary provision; the Pension Protection Act made it a permanent part of the Internal Revenue Code (IRC section 25B). The credit sits on the defined contribution side of the act’s design: automatic enrollment and escalation get workers into plans, the QDIA safe harbor protects the default investment, and the saver’s credit improves the economics of contributing for the workers least likely to save. Its reach is limited in practice because it is nonrefundable, meaning workers with no income tax liability cannot use it. The permanent credit is the act’s most direct subsidy for low-income saving, distinct from the safe harbors that operate through plan design.

Q: How did the SECURE Act build on the Pension Protection Act?

Later legislation turned the Pension Protection Act’s enabling framework into a partial mandate. The 2006 act made automatic enrollment legally safe through preemption and safe harbors but required nothing; SECURE 2.0 (2022) required most newly established 401(k) and 403(b) plans to include automatic enrollment with escalation. The relationship between the statutes is the series’ recurring finding about the amendment stage of a statute’s life: the 2006 act changed procedure and defaults, the later act layered a mandate on top of the infrastructure the earlier one built. Anyone studying the pair should read them as a sequence, with the behavioral design of 2006 as the foundation and the later mandate as the second story.

Q: What did the Pension Protection Act do to lump-sum payouts from underfunded pensions?

It restricted them. One of the defined benefit funding reforms barred plans below 80 percent funded from paying lump sums and restricted benefit increases by underfunded plans. The logic was protective: a struggling plan should not pay out large lump sums or promise richer benefits while its funding deteriorates, because each payout deepens the hole for the workers who remain. For workers, this is one of the act’s most concrete effects: an employee of a deeply underfunded plan could find the lump-sum option limited or unavailable even when the plan document had promised it. The restriction sits inside the funding half of the act, alongside the full funding target, shorter amortization of shortfalls, and revised premiums paid to the insurance corporation. It is also part of the freeze debate, since limits on payouts and benefit growth made sponsoring traditional pensions less attractive to employers.