April 16, 2014

Looking Behind The Colossal Liability Numbers

For the first time ever in 2004, the annual reports of the Social Security and Medicare Trustees contained a complete accounting of the programs' unfunded benefit liabilities. In addition to the usual calculations of Social Security's and Medicare's “actuarial deficit” ($4.0 trillion and $8.5 trillion, respectively), the reports included measures of the programs' “closed group” liabilities ($12.7 trillion and $29.5 trillion) and “infinite horizon” liabilities ($11.9 trillion and $61.7 trillion). The OMB and Treasury have also published similar liability estimates.

The new focus on unfunded liabilities is welcome. The Concord Coalition has long stressed the crucial importance of honest accounting for the government's entitlement commitments. Unfortunately, there is considerable confusion about what the numbers mean. Many policymakers assume that the “official” (and smaller) actuarial deficit figures must be more important than the alternative (and larger) liability estimates, whereas just the opposite is true. Few understand that all of the liability measures are present value figures, and that as such even the largest numbers may understate the long-term challenge.

The federal government's unfunded benefit liabilities are increasingly reported in the press and discussed in Congress. There is even a bill, sponsored by Senator Lieberman, that would make monitoring one special measure--the government's overall “fiscal imbalance”--a formal part of the budget process. It's time to review what the different measures mean, why they are important, and what their limitations are.

A Vast Sum

Let's start with the basics: Unfunded benefit liabilities summarize the gap between projected program revenues and projected expenditures in a single “present value” number. In other words, they tell us the amount of extra money that we would have to have on hand and set aside today in an interest-earning account in order to precisely cover future costs.

Until recently, the only widely reported liability measure was “actuarial balance,” the official indicator of Social Security's and Medicare's trust-fund solvency. The Trustees define actuarial balance as the present value of trust-fund revenue over the next seventy-five years plus current trust-fund assets minus the present value of trust-fund expenditures over the same period. As of the beginning of 2004, Social Security's actuarial balance was a deficit of $4.0 trillion--or the equivalent of an extra 1.9 percent of U.S. worker payroll each year starting today.

Defenders of the entitlements status quo like to cite Social Security's actuarial deficit because it makes the long-term problem seem modest. Yes, the defenders say, an extra two percent of payroll may be a significant levy, but surely we cannot regard it as a catastrophic imposition on future workers.

Actuarial balance, however, is a highly misleading measure of the long-term cost challenge. To begin with, it counts trust-fund assets as genuine savings, even though these assets aren't going to reduce future tax liabilities by one dime. The CBO, GAO, and OMB all concur: The trust funds represent a claim on future tax revenues, not real savings that can be drawn down to defray future benefit costs. If we exclude trust-fund assets from the calculation, Social Security's unfunded liability leaps to $5.5 trillion.

There's another problem with actuarial balance: the arbitrary seventy-five-year time horizon. This limit assumes that future Americans will allow Social Security to run off a financial cliff in 2079. In other words, it assumes that our children, come the year 2025 or 2050, will be utterly unconcerned with the fate of their own children. To assess Social Security's true financial status, we need to look at its liabilities over a longer--indeed infinite--time horizon. While this may strike some readers as unusual, it is actually the seventy-five-year cut-off that is odd. Private financial markets routinely value income and expenditure in perpetuity. Perhaps taking a cue, the Trustees now calculate and publish an “infinite horizon” liability estimate for Social Security. It is $11.9 trillion excluding trust-fund assets--or three times Social Security's official actuarial deficit.

Social Security of course isn't the only program for which the government is accumulating long-term liabilities. There is also Medicare. According to the Trustees, Medicare has an actuarial deficit of $8.5 trillion. This measure, however, is even more misleading in the case of Medicare than of Social Security. Not only does it count illusory trust-fund assets, it ignores trillions of dollars in very real Medicare spending, including the new prescription drug benefit, because it is paid for outside the Hospital Insurance program, the trust-fund financed portion of Medicare. If we include all Medicare spending--and extend the calculation over an infinite time horizon--Medicare's true liability turns out to be $62 trillion, or seven times its official actuarial deficit. The combined liability for the two major senior entitlements thus comes to $74 trillion.

All of the numbers discussed so far are “open group” liabilities--that is, they assume that new generations of participants will join the Social Security and Medicare systems. They measure the extra burden of continuing to pay current-law Social Security and Medicare benefit promises, and thus the programs' fiscal sustainability. They say nothing directly, however, about the programs' generational equity--in other words, about how much of the extra burden will be borne by ourselves and how much by our children. The Trustees also calculate another liability measure, known as a “closed group” liability, that answers this question. The closed group measure assumes that Social Security and Medicare will be closed to all new entrants. It then determines what today's workers and retirees are due to receive in future benefits over and above what those same workers and retirees are due to pay in future contributions. Private pension plans calculate something similar called a “termination liability.”  Indeed, federal law requires them to do so.

As of the beginning of 2004, Social Security and Medicare had a combined closed group liability of $42 trillion. This number represents the subsidy that today's adults expect from future generations, which is another way of saying that it measures the extent to which future generations will fail to get a fair return on their contributions. It also tells us the cost of transitioning from today's pay-as-you-go entitlement system to a new funded system. It is the sunk debt that future generations would have to liquidate before they can invest their own contributions free and clear.

Whichever measure you pick, we are talking about a vast sum. The $42 trillion closed group liability for Social Security and Medicare is ten times greater than our publicly held national debt ($4 trillion).  The $74 trillion infinite future liability exceeds our nation's total net worth ($42 trillion, which includes all property owned by U.S. residents--real and financial). Paying it off would require taxing away an extra 18 percent of workers' earnings forever--again, starting today.

Unfunded Benefit Liability Measures* in Trillions of Present Value Dollars


Social Security



75-year Actuarial Balance




...minus trust fund assets




...plus Medicare SMI 



Infinite Horizon Liability




Closed Group Liability




 *Actuarial balance figures include contingency fund; infinite horizon and closed group liabilities exclude trust fund assets.

Wading Out to the Sandbar

A close look at the unfunded liability numbers leaves no doubt that the nation faces a huge cost challenge. No single present value number, however, can give a complete and accurate picture of the magnitude of the entitlement challenge--or the required response.

One problem is that unfunded liabilities, as present value numbers, implicitly assume that any savings generated by entitlement reforms will be set aside, earn interest, and finance future benefits. They may thus understate the size of the future tax hikes required to sustain today's benefit policies--or conversely, of the future benefit cuts required to sustain today's tax policies. After all, the historical record on budgetary “lockboxes” isn't very encouraging. Three years ago, many hoped that mountainous budget surpluses would pay down the national debt and finance a painless reform of Social Security, but it hasn't worked out that way.

Another problem is that unfunded liability calculations can conceal a roller coaster tax and spending path--and thus obscure the pros and cons of different reform approaches. Consider two reform plans: one that first allows Social Security and Medicare spending to double, then cuts it in half, and a second that keeps spending from rising in the first place. The impact on the programs' unfunded liabilities will be similar, but not the impact on the budget or the economy. It's a bit like wading out to a sand bar. The present value calculation says that you'll be safe and dry once you get there. You need annual budget projections to tell you whether you're going to drown on the way.

A related issue is that the present value accounting framework makes no distinction between formal and informal indebtedness. Some economists argue that this is an advantage. The focus on the annual deficit, they say, creates an irrational bias against reforms that increase the public debt, even if it is projected that the debt will be more than offset by future benefit reductions. In an economist's world, that bias may indeed seem irrational. But in the world of budget politics, it serves a vital purpose. Once public debt is incurred, it must either be paid off or permanently serviced. Future benefit savings is contingent on the willingness of future Congresses to inflict pain on future voters. The one is certain, while the other is not.

There's a final issue policymakers need to consider --namely, that unfunded liability calculations make more sense for some types of spending than others. Economists Kent Smetters and Jagadeesh Gokhale recently calculated an unfunded liability estimate for the entire federal government.   Senator Lieberman's bill would make monitoring this overall “fiscal imbalance” measure part of the formal budget process. The Concord Coalition applauds the Senator for putting honest accounting for the government's entitlement commitments on the Congressional agenda. However, Concord is skeptical about the practicality of the fiscal imbalance measure.

Long-term liability calculations are only meaningful for programs that embody long-term promises. Social Security and Medicare are classic examples. Yes, they are legislated entitlements, not contracts. Still, so long as their authorizing legislation remains in force, accrued benefits are paid out automatically. The whole terminology of social insurance, from “trust funds” and “insured status” to “earned benefits,” implies that participation confers property rights. Much of the electorate believes this to be the case and votes accordingly. As a consequence, no one seriously doubts that Social Security's and Medicare's unfunded benefit promises constitute some sort of government obligation.

Once we move beyond social insurance programs, the political dynamic changes. Means-tested programs like Medicaid may constitute a major cost challenge, but nobody believes that they confer property rights. Spending levels in discretionary programs are determined year to year in the appropriations process. As for general revenues, tax hikes and tax cuts come and go along with short-term political priorities and changes in the fiscal and economic environment.

A single measure of future fiscal imbalance implies that all forms of taxing and spending are equally certain. This can lead to meaningless conclusions, such as the often-heard claim that the seventy-five-year savings from repealing the Bush tax cut would be more than sufficient to close Social Security's seventy-five-year actuarial deficit. The comparison is not so much wrong as irrelevant. Congress routinely changes tax policy year to year. Social Security spending is locked in by promises spanning generations.

Real Americans

In the end, the great value of the unfunded liability numbers may simply be to wake America up to the challenge. They are useful as indicators of the fiscal unsustainability and generational inequity of the entitlements status quo, but are less helpful as guides to reform. They say nothing about annual spending levels, and hence when the entitlement burden will become acute. Nor do they tell us the government's annual borrowing needs, and hence its impact on U.S. savings, investment, and living standard growth.

Solving the entitlement problem will take real Americans making real sacrifices. They will want to know what they have to give up--not just in present value terms, but in terms of their pay stubs and benefit checks.