Dodging the Pension Disaster

The looming pension meltdown is still within our power to avert

by JOSH BARRO

When Dan Liljenquist began his first term as a Utah state senator in January 2009, his financial acumen quickly earned him serious legislative responsibilities. A former management consultant for Bain & Company, Liljenquist was appointed by the Utah senate president, Michael Waddoups, to three budget-related committees; he was also made chairman of the Retirement and Independent Entities Committee. As Liljenquist remembers it, Waddoups pre-empted any concerns the freshman might have had about his new responsibilities: "Don't worry," Waddoups said, "nothing ever happens on the retirement committee."

But then, in the early months of 2009, the stock market went into free fall. Worried about the effects the market crash would have on Utah's public-employee pension plan — which, like most states', is invested heavily in equities — Liljenquist asked the plan's actuaries to project how much taxpayers would have to pay into the pension fund in order to compensate for the stock-market losses. The figures that came back were alarming: Utah was about to drown in red ink. Without reform, the state would see its contributions to government workers' pensions rise by about $420 million a year — an amount equivalent to roughly 10% of Utah's spending from its general and education funds. Moreover, those astronomical pension expenses would continue to grow at 4% a year for the next 25 years, just to pay off the losses the fund had incurred in the stock market.

This scenario alarmed lawmakers, and for good reason. It also alarmed public employees, who feared that rising pension costs would limit the state's ability to pay higher wages. Tapping into these concerns during the 2010 legislative session, Liljenquist built consensus around a cost-saving reform plan: Utah will now require all state employees hired after June 2011 to choose one of two retirement options — either a 401(k)-style benefit plan, or a sharply modified pension plan with costs to taxpayers capped in advance. The reform isn't perfect, of course, but it will be significantly less expensive for Utah's taxpayers, and will leave more room in the state budget for the real business of government.

Utah, it seems, has thus narrowly escaped catastrophe. But what about the other 49 states? The pension-cost explosion is hitting nearly every one of them, too. And unlike Utah's, these states' efforts at pension reform are not being overseen by management consultants. Rather, in most places, state legislators are overmatched by savvy public-employees' unions and by pension-fund managers wedded to the status quo. Their influence explains why, though 18 states enacted some sort of pension reform in 2010, very few will offer real, long-term relief to taxpayers.

Concern about this impending crisis should extend far beyond state capitals, because its consequences will affect much more than state balance sheets. The staggering burden of paying out retirement benefits is increasingly preventing state and local officials from financing all the other services that citizens expect their governments to perform. For example, Camden, New Jersey — one of the most crime-ridden cities in the country — recently had to lay off nearly half its police force because the state's public-sector unions, including those representing police, were unwilling to cut costly benefits provisions from their contracts. Moreover, should the states prove incapable of getting their pension costs under control, they will put the squeeze on taxpayers across the country — forcing them to pay more in exchange for fewer government services — and could precipitate federal bailouts that would effectively transfer money from fiscally responsible states to profligate ones.

Everyone, then, has a stake in understanding just how the states got into this terrible mess, where many states are going wrong in trying to rectify it, and — perhaps most important — just what principles should animate any reform plan capable of both shoring up state pensions and shielding the taxpaying public.

DEFINED BENEFITS

To understand how the states got into their current sorry predicament, it is essential to examine the structural flaws in the state-pension edifice. There are two fundamental problems with the pension plans offered by state and local governments all across America: One is that, in many cases, the benefits are excessively costly, insofar as they are more generous than is necessary to attract qualified talent to government work. The other is that, by guaranteeing annuity-like streams of income in retirement — regardless of whether the pension funds' assets and market performance can support those payouts — such plans expose taxpayers to enormous risk. After all, those taxpayers are the people who will be responsible for making up any shortages.

Both of these problems are driven by the structure of most public-employee retirement plans, which follow what is known as a defined-benefit model. As the name would suggest, a defined-benefit pension plan guarantees some fixed level of income to workers upon their retirement; benefits are determined by a formula that is typically based on the number of years worked and average earnings in several years leading up to retirement. (Under some defined-benefit plans, retired workers are also entitled to annual cost-of-living adjustments.)

In principle, defined-benefit pensions are designed to be pre-funded. Employers are supposed to set aside money during a worker's career to pay for his benefits in retirement; in many cases, the employee is required to make some portion of the total contributions himself. The employer — in the case of public workers, the state — then invests these assets, mostly in equity investments (such as mutual funds or stocks) with a minority in fixed-income vehicles, such as bonds.

The key to defined-benefit plans, however, is that the employee's benefit payments are not affected by the market performance of those assets. In this sense, defined-benefit plans are explicitly designed to shift investment risk from employees to employers. In the case of public pensions, if a plan misses its target investment return, state workers don't see their benefit checks shrink: Instead, taxpayers hand over more of their earnings to the government, so that it can make good on its promises to public-sector retirees.

Over the long term, defined-benefit pension plans can pose serious challenges to an employer's bottom line — as Chrysler and General Motors (and the American taxpayers who had to bail them out) can attest. Perhaps this is why defined-benefit pensions have become increasingly rare in the private sector: As of March 2010, just 30% of private-sector workers at medium and large firms participated in a defined-benefit pension program (down from 80% in 1985). Far more common in today's marketplace is the "defined contribution" plan — such as a 401(k) — through which an employee sets aside earnings in a tax-free account that he can draw from upon his retirement.

But the public sector, characteristically, has been much slower to grasp the problems inherent in the defined-benefit structure — and the model is still used by every state to provide benefits for at least some workers. To be sure, some states are starting to inch away from this approach: Michigan and Alaska abandoned the defined-benefit model for most new state workers in 1997 and 2005, respectively; Alaska's reform also extends to local-government employees. But overall, in 2009, 84% of state and local workers in America were offered defined-benefit plans; today's federal employees also receive relatively small defined-benefit pensions in conjunction with 401(k) plans. (In addition to pensions, many public workers in the United States are eligible for free or heavily subsidized health insurance in retirement — a benefit that states have typically done even less to pre-fund than they have pensions, and one that is rising rapidly in cost.)

Often, taxpayers aren't even fully aware of the degree to which they are on the hook for state workers' generous benefits. Indeed, one of the inherent dangers of defined-benefit pensions is that such schemes allow lawmakers to promise future payments to state workers without having to fund those benefits adequately in the here and now. And because the present costs of state workers' benefits are never transparent to the voting and taxpaying public, politicians enact more expensive benefits provisions than they could get away with otherwise.

ACCOUNTING DOUBLE STANDARDS

In the past year, think tanks and the press have tried to illuminate the opaque accounting practices of public-pension plans and the vast unfunded liabilities they obscure. Though often presented as impossibly complicated, the problem at the heart of the pension crisis is fairly simple: A pension plan holds a pool of assets and owes a stream of payments to government workers and retirees. If the plan doesn't hold enough of the former to cover the latter, it is said to be "underfunded." And it is this underfunding that presents a severe, long-term challenge to state policymakers. Even accepting the accuracy of the plans' own figures, the funding gap is alarming: As the Pew Center on the States noted in a report released last year, state and local governments were $1 trillion short of funding their pension and retiree health-care commitments. And these calculations were based on figures from the end of fiscal year 2008 — meaning that they included, for the most part, financial statements prepared before the stock market took its nosedive.

Yet dire as these figures seem, they are still far too rosy. A consensus has emerged among economists that government plans are fundamentally miscalculating the liabilities they owe, as they understate the cost of benefits to be paid in the future by counting on high investment returns that may not materialize (as discussed below). As a result, these government plans are also understating their funding gaps. Estimates from the Cato Institute and Credit Suisse put states' unfunded liabilities just for health care north of $1 trillion. And economists Robert Novy-Marx (of the University of Chicago) and Joshua Rauh (of Northwestern University) find that pension funds are short by more than $3 trillion.

These numbers are enormous, but their true magnitude becomes more clear when they are placed in fuller context. Consider that the total outstanding bond debt of state and local governments is about $2.4 trillion. If one accounts for pension and health-care debts using the figures supplied by Novy-Marx and Rauh (among others), the total outstanding obligations of the states rises to as much as $6.4 trillion — meaning that our sub-national governments are nearly three times further in the red than they appear to be at first glance.

The difference between these market-value estimates and the official tabulations of pension liabilities has to do with the choice of a discount rate — or rather, the presumptive interest rate one uses to determine precisely how much a sum of money that one will either pay out or receive in the future is worth today. In a sense, discounting works like an interest calculation in reverse: If a person owes $1.05 in one year, is discounting at a 5% rate, and holds $1 today, he can say that his $1.05 payment due in one year is fully funded.

Governmental Accounting Standards Board rules allow public-pension plans to set their liability discount rates equal to the investment returns they expect to achieve on their assets. Among major public-sector pension plans in the United States, that rate ranges from approximately 7.5% to 8.5%, with 8% being the most common choice. These rates reflect the fact that pension funds typically invest most of their assets in equities, which can achieve relatively high returns.

But those higher returns carry a downside: volatility. With the right investment mix, pension funds can, over time, average returns in the neighborhood of these targets; indeed, they have historically achieved such rates, even when one factors in the recent crash. The problem is that they cannot reliably yield such returns in any given year: In some moments, investments will produce windfalls that far exceed expectations; at other times, as in the period from 2008 to 2009, the funds' returns will come in far behind.

State governments expect to exist in perpetuity, which means they tend to take a very long-term view of their pension obligations — longer than, say, a corporation that has to worry about remaining profitable. But that doesn't mean governments can simply wait around for investments to bounce back after a stock-market crash while continuing to pay out benefits at the same level. If they do, they risk allowing a pension-fund balance to run all the way down to zero.

As a result, when pension funds lose money, taxpayers must step in to make up the difference. For example, in a recent report, my Manhattan Institute colleague E. J. McMahon and I estimated that employer — i.e., taxpayer — contributions to the New York State Teachers' Retirement System will more than quadruple over the next five years, principally as a result of recent stock-market declines. In this sense, taxpayers provide valuable insurance to public-employee pension plans, guaranteeing equity-like investments with bond-like certainty.

By contrast, pension plans in the private sector — governed by the separate Financial Accounting Standards Board — are not allowed to choose a discount rate based on expected returns on assets. Instead, they choose a discount rate based on a principle called the "market value of liabilities." Under this principle, a payment due in the future should be discounted at an interest rate consistent with the risk experienced by the creditor (which, in the case of a pension plan, is the worker or retiree). The amount of the liability is unaffected by either the nature or quantity of the assets the pension plan holds. For most private-sector pension plans, the market-value approach produces a discount rate between 5% and 6% — noticeably lower than the 8% presumed by the public sector.

The wisdom of the private-sector approach over that of the public sector is illustrated by an example from Novy-Marx and Rauh. Let's say that a person owns a home that has a mortgage on it, and he also has significant liquid investments that he intends to use to gradually pay off the mortgage. Now imagine that he re-allocates his investment portfolio away from bonds to stocks, increasing his expected return. Would we say that this change in his investment strategy has caused his mortgage balance to fall? Of course not. But that is what public pension plans do, by using expected asset returns as a component of the calculation of liabilities.

THE TRUE COST OF PUBLIC PENSIONS

The degree to which these GASB-approved accounting tricks have led public pension funds to the brink of insolvency is usually the problem that draws analysts' focus. They are concerned — rightly — that pension plans are claiming to be much better funded than they really are. But this means that a related cash-flow issue often goes overlooked: The same error in discount-rate selection means that pension funds are also far more expensive than they claim to be. Clearly, this misrepresentation has major political and fiscal consequences. When lawmakers, and the voters to whom they are accountable, evaluate a state's pension system, the obvious first question is, "What do the benefits cost?" Accounting rules make answering this question less simple than it should be — and that difficulty has allowed lawmakers to hide pensions' true costs to taxpayers.

On the surface, the most obvious way to assess pension costs would be to look at the payments that state and local governments actually make into their pension systems. Every year, pension actuaries essentially send a bill to these governments, telling them what they must pay to cover the accrual of more benefits and to shore up plans that are underfunded. This first component — covering benefits accrued by active employees in the current year — is called "normal cost"; the second part is called "amortization cost." (In the case of an overfunded plan, the amortization cost may be negative.)

The trouble with using pension payments as a stand-in for pension cost — which the Bureau of Labor Statistics does when calculating compensation data — is that the total payment into a pension system in a given year usually does not accurately reflect the cost of paying out that year's benefits. For one thing, amortization cost — a key component of the actuaries' annual bills — is really an assessment to pay for labor that public employees provided in past years. Even if a government had no current workers on the payroll, its pension plan would still have to amortize any unfunded liability.

The complications posed by amortization costs are illustrated through a situation that was common at the peak of the tech bubble. During that period, high investment returns led some pension funds to report that they were significantly overfunded, resulting in large, negative amortization costs. As a result, pension contributions in those years were near zero. But this did not mean that pension benefits in those years were free; by promising pension benefits to workers, governments were still incurring liabilities that they would not otherwise have owed. Rather, the very low fund contributions meant simply that governments were using their investment returns, not tax receipts, to pay for retirement benefits in those years.

Another problem with using annual state contributions to determine a pension fund's cost is that some states disregard their actuaries' recommendations and pay less than they are supposed to when times are tight (Illinois and New Jersey are notable culprits). But failing to properly fund pension benefits does not make them cheaper (indeed, just the opposite); it simply means that the government is delaying the cash payment to a later year.

Rather than looking at annual cash contributions, then, pension cost should be thought of as equivalent to normal cost — the amount by which pension liabilities grow because workers are able to earn pension credits in a particular year.

But while this would be an improvement, there are still problems with a normal-cost approach. And chief among them is yet another issue relating to discount rates. In a pension fund, normal cost represents the present-day expense of promising to make a stream of payments in the future — which means that the cost is calculated using a discount rate. Funds generally use the same rate that they use to calculate their aggregate liabilities — about 8%. Of course, in this case, too, the discount rate is excessively high, leading states to underestimate their annual normal costs.

The effect of using the wrong discount rate is significant. Consider the New York State Teachers' Retirement System, which estimates its normal cost for employees hired in 2009 or earlier at 11.8% of wages and salaries (meaning, on average, that paying out one dollar in wages implies pension benefits that will cost 11.8 cents each year to provide). Adjusting the discount rate from 8% to a more realistic 5% increases that normal cost to approximately 19% — meaning the pension is really about 60% more expensive than its stated normal cost. If lawmakers had to calculate pension funds' normal costs on a market-value basis (as private pension funds do), and also had to increase required taxpayer contributions accordingly, they might think twice about making government workers' pension benefits ever more generous.

Accounting issues like these are but one way in which defined-benefit pensions systematically lead lawmakers to award overly generous pension benefits. Another major flaw is inherent in the very nature of pensions: They allow lawmakers to give valuable benefits to public workers (and to placate unions) without ever having to deal with the ugly future consequences. Handing out a wage increase, after all, generally requires coming up with a significant amount of money in this year's budget, which can pose enormous financial (not to mention political) difficulties. Sweetening pension benefits, on the other hand, achieves much the same political end — and while it does increase a pension system's unfunded liability, that cost is spread across pension payments that will be made for many years. In this way, legislators can please public employees now and leave it to future legislatures to clean up the mess.

This practice was common at the peak of the tech bubble, when many states — particularly New Jersey, New York, California, and Illinois — handed out pension sweeteners that, thanks to the high discount rates permitted under GASB, had the appearance of being "free." Of course, those sweeteners weren't free: They were simply financed with returns on investments that belong to taxpayers — returns that should have been used to reduce taxpayers' contributions to public workers' pensions. Then the stock market performed poorly over the next decade, and those "free" sweeteners ended up being paid for with tax dollars after all. (New York and New Jersey have since taken steps to reverse the increases, but only for newly hired employees.)

Defined-benefit pension plans thus provide lawmakers with both the motive and the means to seriously abuse state finances. All over the country, state lawmakers are tempted to appease government workers now, and let someone else figure out how to pay the bill in the future. At the same time, the complex accounting rules that govern defined-benefit pensions make it easy to cover up the costs of the scheme.

It is not impossible, of course, to construct a defined-benefit plan with reasonable costs. Certain institutional changes — like de-linking liability discount rates from asset returns, or requiring voters to approve any sweetened benefits for government workers (as they must approve the issuance of bonds in many states) — could help make costs somewhat easier to control. But there is another major problem with defined-benefit pensions that is fundamental to their nature — one that cannot be avoided without switching to a defined-contribution system. And that is the excessive risk to which they expose taxpayers.

RISKY BUSINESS

Defined-benefit plans are, by their very design, intended to shift investment risk away from workers to employers. In most cases, both employees and employers make payments into a pension fund, but the employee payments are fixed — so only employer-contribution rates change with the funds' asset performance. Thus these plans carry the risk that, during economic downturns, the employer will be expected to come up with sharply increased payments for pensions. In the case of public-sector pension plans, this means taxpayers will have to pay much more at precisely the moment when they can least afford to.

To see just how painful this can be for taxpayers, consider the case of the New York State Teachers' Retirement System and the New York State and Local Employee Retirement System. Both were battered by the stock-market decline that followed the 2008 financial crisis; as a result, the funds held $212 billion in assets at their reporting dates in 2010, reflecting no increase since the year 2000. Yet because of growing retiree headcount, rising salaries, and a series of pension sweeteners approved by the state legislature, the funds last year were paying out twice as much in annual pension payments as they were a decade earlier. Accordingly, required taxpayer contributions are rising sharply: While employer-contribution rates were near zero a decade ago, and ranged between 6% and 10% of covered payroll over the past several years, rates for both plans are likely to climb above 20% by 2015.

When expressed in dollar terms, the increases appear even more stark. In fiscal year 2011 (which ends June 30), New York state school districts are making $900 million in contributions to the teachers' retirement system. By fiscal year 2016, that annual figure will rise to approximately $4.5 billion — and that's assuming that the fund hits its 8% targeted investment return over the next five years. These figures mean that school property taxes will have to rise 3.5% per year just to pay for the growing expense of providing teachers' pensions — even before addressing cost increases in any other area of the state's education system (including any potential wage increases for current teachers). All of this is the result simply of stock-market declines since 2007.

For this reason, it is incorrect to think of pension-fund assets as "belonging" to retirees. What the pensioners own is a valuable bond-like promise from the government. The taxpayers owe those promised funds to the pensioners — which means they also own the funds' asset pools, and rely on their strong performance to prevent massive tax increases. Essentially, state governments are long in the stock market on taxpayers' behalf.

Placing such a burden on federal taxpayers would be one thing; to limit risk, Washington could just issue Treasury bonds to pay rising costs, and pay them off when times are good (though taxpayers would still eventually be on the hook for interest payments). But almost every state is required to balance its budget each year. This means that swinging pension contributions place added pressure on already strained state budgets during recessions. The result is a harmful de-stimulative effect on the economy — as state program spending must be cut to make up for pension losses, or taxes must be raised on taxpayers who are feeling their own pain from the economic downturn. Both are policies to be avoided during recessions — and yet states' defined-benefit pensions leave lawmakers with few other choices.

Unfortunately, Utah is one of only two states to have seriously attacked the risk problem since the recession hit; Liljenquist's reform plan will shift investment risk away from taxpayers, at least for the retirement benefits of newly hired employees. (Michigan, having moved a large portion of state workers to 401(k) plans in 1997, will start offering new teachers a somewhat less generous defined-benefit pension, plus a 401(k) plan with a 2% match.) Each of the other 24 states that enacted reform plans in the past two years, meanwhile, has worked within the confines of the defined-benefit model. These reforms may provide some cost reductions, but they will not address the enormous risks that today's pension schemes pose to taxpayers.

PRINCIPLES FOR REFORM

Any successful pension reform, then, should address both problems: excessive cost and excessive risk. Achieving these twin aims will clearly not be easy, and each state will need to consider its own unique problems and circumstances. Even so, there are three general principles that states can follow if they want to enact meaningful reforms that stand some chance of staving off pension disaster.

First, pension reforms should include all benefits that will be accrued in the future, not just benefits that will be accrued by new hires. As mentioned earlier, most states are limiting their pension reforms to new employees only — which means they are likely dooming their reforms to failure.

Admittedly, the political logic of their approach is easy to understand. Workers who might be hired in the future do not belong to unions; they have no voice in the political process. Active workers and retirees, however, are strong political forces in state capitals. Moreover, reforms that apply only to new workers can in fact generate significant savings over time: In the case of Michigan, for instance, thanks to the reform plan enacted in 1997, approximately half of the state's current work force is not eligible for the defined-benefit pension plan. For those employees who remain in the defined-benefit plan, Michigan's pension costs per worker have risen sharply in the past few years — as have those for every other state. But for the majority of the work force now in the defined-contribution plan, the state's costs are flat. As a result, Michigan was saving $210 million per year by 2010, with the savings growing each year.

The problem with this approach, however, is that cost savings in the near term are extremely limited. Last year, New Jersey passed a pension-reform law that rolled back a 9.09% pension sweetener enacted in 2001. But the rollback applies only to workers hired after the enactment of the reform law — meaning that it won't have much effect on the size of pension checks until around 2040. Other large states, including Illinois and New York, have recently enacted similar, purely prospective cutbacks.

Reductions in benefit accruals do enable states to cut back somewhat on pension-contribution payments before they actually flow through as smaller pension checks: A reduction in normal cost for new employees means that less money can be deposited each year to keep the pension system actuarially sound. But even this effect takes years to become substantial, because within the first few years after reform, a large majority of active employees will be accruing under the old rules, with the old (higher) normal cost. This does little to alleviate the cash-flow problem that taxpayers face today. For reform to have any meaningful effect on required contributions in the short term, it must also include current workers.

This does not mean that states should take away workers' already vested benefits. Those payments represent promised compensation for labor provided in the past, and states should honor such contracts unless they lack the overall financial wherewithal to honor their debts. But it does mean that states should act to reduce the benefits that current workers can accrue in future years. Essentially, states should adopt the private-sector model of pension-plan termination: Unless firms go bankrupt, they cannot revise pension benefits already accrued, but they can reduce or eliminate the benefits that workers expect to earn in future years of employment.

A few recent proposals have come close to striking the right balance. The Civic Committee of the Commercial Club of Chicago has suggested such a reform for Illinois, after obtaining a legal opinion stating that cuts to future benefits for current workers are allowable under the pension-guarantee provision in the state's constitution. And New Jersey governor Chris Christie's pension-reform proposal includes some reductions of future accruals by existing workers. But the states that have actually touched the benefits of current workers — Minnesota, Colorado, and South Dakota — have done so in the wrong way, by implementing across-the-board reductions that do not differentiate between benefits already accrued and benefits to be accrued in the future. Not only does this involve the abrogation of real promises to public workers, it has also invited litigation. Clearly, more education needs to take place before state lawmakers have a proper understanding of what to touch, and what not to touch, when it comes to trimming back future benefits.

Second, serious pension-reform plans should abandon the defined-benefit model. Three states — Michigan, Alaska, and Utah — have enacted reforms that will move many employees to defined-contribution retirement plans, or at least to sharply modified defined-benefit plans that shift most investment risk away from taxpayers. In most states, however, pension reform has been a matter of tinkering: increasing employee contributions, adjusting benefit formulas, raising retirement ages, and so on.

The problem with tinkering is that it addresses only the matter of pensions' high costs, doing nothing to shield taxpayers from the investment risks discussed above. Moreover, if states don't seize this moment to do away with defined-benefit models once and for all, they are unlikely to have another chance any time soon. And those small fixes that they do enact will likely be undone in future years. Administrations change, legislative seats turn over, and public-employee unions wax and wane in their power. Given a few years of strong stock-market returns and friendly lawmakers, public-worker interests will likely be able to restore any benefits they lose during the current reform cycle — which will put state finances, and the taxpayers affected by them, in the same tough position during the next downturn.

New York taxpayers have learned about these dangers the hard way. There is a reason that the pension fixes enacted in 2009 were called "Tier V" and not "Tier II": There had been three previous attempts to rein in the excessive cost of New York's public-employee pensions by creating less generous pension "tiers" for newly hired employees. These reforms date back to the fiscal crisis of the 1970s, when unsustainably generous contracts with public-employee unions threatened to throw New York City into bankruptcy. Since then, though, New York's public-worker unions have been highly successful in unwinding previously enacted pension reforms. The new Tier V is nearly identical to Tier IV at the time of its enactment in 1983 — but Tier IV has been repeatedly, and retroactively, sweetened through increases in benefit formulas, cuts to employee contributions, and reductions in the retirement age. Similarly, by the time substantial numbers of workers actually start retiring under Tier V around 2040, this plan, too, will probably bear little resemblance to its current form.

Moving to a defined-contribution system would not make such reversals impossible, but it would probably make them much less common. For lawmakers, a key appeal of pension sweeteners in a defined-benefit system is their opacity: The fiscal cost often appears to be financed by windfall returns on pension assets. Similarly sweetening a 401(k) plan, however, would require a direct infusion of cash into workers' accounts, at a significant and immediate cost to taxpayers — making such goodies much more difficult for lawmakers to distribute.

Third, states should consider voluntary buyouts of existing pension benefits. The two reform principles outlined above address only the costs of pension benefits going forward; they do not help resolve the very real problems associated with states' existing pension liabilities — those that were incurred by governments as payment for labor that employees provided in the past. Here, there are no easy policy maneuvers: Short of defaulting on these debts, the only way states can eliminate unfunded pension liabilities is to fund them.

Unless, that is, employees voluntarily agree to sell their pension benefits back to their employers. Even if governments can be trusted to make pension-benefit payments as scheduled — which, given some states' current circumstances, is a big "if" — many employees would probably accept significantly reduced pension payouts if they could get their benefits in one up-front, lump-sum payment. The reason is the difference between cost and value: Part of why pension benefits are so expensive is that it is costly to provide insurance of long-term returns; workers, however, may not place a value on that insurance that is as high as the cost of providing it. Effectively, a pension benefit is similar to a 401(k) plan invested entirely in annuities. And the fact that few individuals choose to invest their retirement accounts entirely in annuities — especially during their working years — suggests that pension benefits may be worth less than they cost to provide.

A working paper by Maria Fitzpatrick, a fellow at the Stanford Institute for Economic Policy Research, attempts to determine just how highly some public employees value their pension benefits. She examined Illinois teachers' choices when, in 1998, they were offered a chance to make a one-time payment up front in exchange for more generous benefits in retirement. The terms of the purchase varied significantly depending on a teacher's salary and years of service. Using reasonable discount rates, the up-front purchase cost was lower than the present value of benefits for nearly all teachers — 99% could expect at least a 7% annual return on investment, with no risk so long as the state did not default. But the deal was sweeter for some teachers than for others, a variation that made it possible to estimate the subjective present value that teachers placed on future benefits.

Fitzpatrick's finding is, in a way, depressing: On average, teachers were willing to pay only 17 cents on the dollar to obtain a pension-benefit increase. This suggests that defined-benefit pensions are a highly inefficient form of compensation, costing taxpayers far more than they are worth to public employees.

But it also suggests an appealing policy solution: Governments can offer to buy back promised pension benefits at a discount, and employees may be inclined to take the deal. Admittedly, the proposal presents a political problem to lawmakers, in that it requires them to produce an immense sum of cash up front in order to eliminate a long-term liability. To alleviate some of that pain, however, governments could responsibly issue bonds to raise the money — since this would mean simply substituting explicit debt for a larger amount of implicit pension debt. Governments would incur an obligation to pay interest on the bonds, but in most cases that amount would be more than offset by the reduction in required employer pension contributions.

So just how much would governments have to offer workers to induce them to participate in such a buyback program? There are reasons to suspect the payment would have to be higher than 17% of their benefits' present value. To begin, people tend to irrationally value assets they already own more highly than assets they could buy — a phenomenon known as the "endowment effect." Workers might therefore demand a higher price when selling their benefits than they would have been willing to pay to buy the benefits in the first place (which is what Fitzpatrick measured in her study of Illinois teachers).

There are also negative federal tax implications attached to taking pension benefits as a one-time buyout. In addition to the income tax, workers under 59 and a half years old would have to pay a 10% early-withdrawal penalty. These payments could be avoided by rolling the payout over into an IRA, but the requirement could make the buyout program less appealing to workers seeking flexibility. (Congress should consider waiving the early-withdrawal penalty for buyouts of public-employee pensions as a way to help states shrink their liabilities.)

Finally, union resistance to a voluntary-buyout program would likely be strong — meaning they would probably demand higher payouts for workers in exchange for supporting the plan. Legislators might argue that a voluntary buyout could only make workers better off: If they didn't like the deal, they could keep the benefits they were already entitled to. Historically, however, unions have strongly resisted any moves that have been perceived as weakening the defined-benefit system (such as voluntary 401(k) options for employees).

Obviously, none of these reforms would be easy to implement. The states are in grave trouble, and do not have easy options. Nor are these principles to be mistaken for complete solutions. But they do offer politicians a basic guide for constructing policies that can help rescue seriously ailing pension plans. And they show that the challenges that any reform will inevitably present are not insurmountable. Indeed, the greatest challenge at this moment may be getting lawmakers to realize that whatever unpleasantness reform might bring now would pale in comparison to the pain they invite by doing nothing.

AGAINST THE CLOCK

Those reformers intent on rescuing states from pension crises will clearly have their work cut out for them. How, then, ought they to get the process moving? To begin, state and local lawmakers will have to borrow a page from Dan Liljenquist's book and clearly explain to their colleagues and voters what will happen to pension plans (and to taxpayers) in the absence of reform. They should get their pension funds' actuaries to disclose the plans' underfunding on a market-value basis, and to reveal the expected trajectory of required taxpayer contributions in upcoming years. Because of the way pension funds delay recognition of abnormal gains and losses, most states can expect a cost explosion similar to the one that is coming in New York state — an increase that is likely to be a powerful motivator for reform.

Congress, too, has a useful role to play. Lawmakers in Washington should mandate, or at least strongly encourage, greater pension transparency by states. For example, the Public Pension Transparency Act — sponsored by Republican congressman Devin Nunes of California — would tie federal subsidies for municipal bonds to states' making certain disclosures about their pension funds, including market valuations and projections.

Even with greater transparency, however, modifying pensions for current employees will remain a tough sell politically; unions especially will fervently resist. In this situation, lawmakers might be well served by a divide-and-conquer approach: They can begin to draw a clear distinction between benefits that have already been earned (and the valid contractual claims held by the people who have earned them) and future benefits to be paid (which, if disbursed, would impose unacceptable and unnecessary costs on future taxpayers). Lawmakers can promise current retirees that their benefits will not be touched at all in any reform plan, emphasizing that those benefits will in fact be safeguarded by changes to benefits not yet accrued — changes that will make the pension system as a whole better funded. Drawing such distinctions might help lawmakers drive a wedge between active and retired government workers, thereby fracturing the anti-reform coalition.

Of course, the best tool reformers have at their disposal is urgency. When Chris Christie talks about why public workers should support reform, he argues that, without meaningful changes, New Jersey's pension funds could run out of money — leaving everyone without benefits in a decade. There are, naturally, some steps short of total default on pension obligations: A state could, for instance, write pension checks on a pay-as-you-go basis once the funds are tapped out. But such a measure would be extremely costly: In the case of New Jersey, it would require a payment of more than $3 billion a year — more than 10% of the state's current general fund, and nearly $400 per resident. The amount would only grow over time, and at a faster clip than the economy. In that event, pensioners would almost certainly end up taking a haircut on their benefits. Lawmakers need to communicate — and public workers and their unions need to understand — that it is much better for everyone to plan and adapt now, in advance, so that cuts are orderly and focus on benefits not yet earned, and thus avoid slashing the fixed incomes of people who are already retired.

Even though we are just starting to recover from the Great Recession, another fiscal crisis lurks around the corner. What we need now are serious reforms — plans that focus on the underlying causes of pensions' excessive costs and excessive risks. The good news is that the looming pension meltdown is still within our power to avert. The question is whether lawmakers and public workers can muster the discipline and political courage to do it.