Thoughts on a Center-Left Entitlements Strategy

In a forum such as this, authors can dispense with many of the usual preliminaries. I take it we agree that suitably structured and regulated markets generate wealth more effectively than other economic systems but do not reliably produce either a reasonable distribution of prosperity's fruits or an adequate level of security against life's physical and financial vicissitudes. It is for that reason, we agree, that since the 1930s the United States has developed a web of programs to assist the poor and vulnerable, to make work pay and provide protection against unemployment, and to ensure older Americans a decent retirement. We agree that these programs reflect necessary and important public purposes that government may legitimately pursue. We agree, therefore, that the task before us is not to disable or dismantle these programs in the name of other purposes but rather to improve them and to do our best to ensure their perpetuation for future generations.

When we move from principles to particulars, the arena for discussion is vast. I will focus my remarks on two of the largest programs, Social Security and Medicare. Means-tested and work-related programs pose different but no less important challenges, and I'm confident that other participants in this symposium will do justice to them. Nor will I deal with creative new ideas such as progressively structured retirement savings programs on top of Social Security, which I support.

A philosopher once remarked that if you truly will an end, you must also will the means to it. So let me begin with the most obvious point: unless economic output, employment, and wages grow at reasonable rates over extended periods, it will be impossible to finance programs that provide adequate income for retirees and protect them against devastating health care costs. For that reason, it is important to structure and finance entitlement programs so as to minimize potential negative impacts on growth and employment. For example, employers compare the marginal cost of adding workers to the gains at the margin that those workers could produce. In making that comparison, they look at total compensation, not just money wages. There are limits, then, as to how high payroll taxes (and health insurance premiums) can rise before they discourage employers from hiring.

There is a further consideration. I am not the only contributor to this forum, I suspect, who believes that economic and productivity growth require robust public as well as private investment. The list of pro-growth public investments is long and familiar, ranging from education and training to infrastructure and research. At some point, competition for public resources between investments and entitlement spending may develop. Indeed, it may be imminent. The Congressional Budget Office's analysis of President Obama's proposed FY2014 budget shows that while mandatory spending would rise slightly as a share of GDP between 2012 and 2023, discretionary spending would fall from 8.3 percent of GDP to 5.0 percent (even with the sequester replaced). CBO's updated baseline projections, which assume current policy, put discretionary spending at 5.3 percent of GDP in 2023-the lowest level of discretionary spending relative to GDP since at least 1962. (The bipartisan budget deal struck in December changes budget authorizations for fiscal years 2014 and 2015 but has no impact on authorized spending in the out-years.) No one believes that we can adequately defend our country and assist low-income Americans and invest in the future with only 5 percent of our nation's output. Something has to give.

Many progressives believe that what should give way is the current limits on government: the United States should do all of the above, raising taxes and expanding government as needed to fund and implement them. Setting aside potential political obstacles, there are some fiscal and economic considerations that warrant caution.

Most Americans believe that the U.S. public sector is substantially smaller than its European counterparts. In fact, the gap is relatively modest. According to the OECD, U.S. government expenditures totaled 41.7 percent of GDP in 2011, compared to 45.3 percent for Germany and 44.5 percent for Norway. To be sure, some OECD countries (France and Sweden, for example) spent substantially more, but others (Japan and Switzerland) spent less. Although the United States is a bit below average in public outlays, it is not a conspicuous outlier. The same is true for public revenues, which now total about 36 percent of GDP.

There are two reasons why we are inclined to believe otherwise. The first is our system of federalism. Compared to most other OECD countries, the United States conducts a high proportion of its public activities-nearly half, judged by outlays and revenues--below the national level. It's a mistake to compare our federal government to the more centralized national governments of Europe. Second, more than half our health care expenditures are private, compared to only one-third for the OECD as a whole. If ours were the same as the rest of the developed world, the gap between the size of our public sector as a share of GDP and theirs would disappear. (To be sure, if more of our health care were in the public sector, it might cost less-as it does everywhere else.) None of this proves that we couldn't increase the size of government and level of taxation without damaging the economy. But it does suggest that our room to maneuver is more constrained than is often thought.

Two other factors point in the same direction. As I argued above, to carry out core functions, discretionary spending in the coming decade would have to be about 2 points of GDP higher than the president's budget calls for, bringing total federal spending to roughly 24 percent of GDP. And second, spending increases in Social Security and health programs are scheduled to accelerate sharply after the current ten year budget window ends. Between 2022 and 2037, according to the long-term budget scenario CBO regards as the most realistic, these large mandatory programs will increase from 12.9 percent of GDP to 16.6 percent, forcing us to choose between unsustainable deficits and a huge increase the federal tax burden-that is, unless we were willing to rein in the rate of increase in these programs.

The CBO analysis allows us to quantify these alternatives. Under their realistic scenario, federal spending before interest payments would be 26.1 percent of GDP in 2037, up from 20.7 percent in 2022. In the absence of tax increases, federal borrowing would surge, boosting federal debt from 93 percent of GDP to 199 percent, and interest on the debt from 3.7 percent of GDP to 9.5 percent. Conversely, we could stabilize the debt/GDP ratio at the 2022 level by enacting, that year, an annual increase in federal taxes that CBO estimates to be more than 6 percent of GDP. (In 2013 terms, that would amount to about $1 trillion.) Avoiding both these unattractive futures would require us to work out some balance between spending restraints and revenue increases.

Most Democrats hope that full implementation of the Affordable Care Act will slow the rate of increase in health care costs, not just for a few years but permanently. I hope so too, and I have no basis for saying that these hopes are futile. But even if the ACA exceeded expectations, it wouldn't materially change the budget picture during the coming generation. Here's why. First, CBO's baseline scenario already assumes no excess cost growth (above the rate of growth in GDP) for Medicare. Second, between now and 2037, the aging of the population will contribute far more to increasing federal health care outlays than will rising health care costs per beneficiary. And finally, the aging of the population accounts for all the increase in Social Security outlays.

In short, the surge in spending for mandatory programs over the next quarter century is primarily driven by predictable, irreversible changes in the age structure of our population. Americans 65 and over now equal about 23 percent of working-age Americans between 20 and 64; by 2037 they will be about 38 percent. And federal spending on programs in which they participate will rise commensurately, even if medical cost increases magically subside to the rate of general inflation and remain there indefinitely.

But so what? Maybe deficits and debt don't matter. I think they do. The problem is that it is impossible to be precise about when they will have an unsustainable impact on interest rates, output, and public confidence. That's why the best metric is a rule of thumb-namely, a debt/GDP ratio that stabilizes over economic cycles at a level consistent with economic growth and the ability to cope with financial and national security crises. Richard Kogan and Paul Van de Water of the Center on Budget and Policy Priorities offer an excellent discussion of this metric and its rationale: "Generally, the debt ratio should rise only during hard times or major emergencies and should decline during good times. . . . The debt ratio cannot rise forever. If it did, national savings available for private investment would shrink, . . . ultimately impairing productivity growth and living standards. . . . In a crisis, international credit markets might refuse to lend to the U.S. public or private sectors at a reasonable price. . . . All else being equal, a lower debt-to-GDP ratio is preferred. It reduces interest costs (allowing more of the budget to be devoted to actual programs) and gives policymakers more flexibility in economic and financial crises."

There's a fallback argument: Yes, deficits and debt matter, but not now. The spending restraints and tax increase enacted during recent years are enough to stabilize our finances over the next decade. So even if we eventually need to do something about Social Security and Medicare, we don't have to act anytime soon, and therefore we shouldn't. Things could change, and we should wait until we can act on the basis of the fullest information possible.

I question the premise of this argument. Current ten-year projections rest on what I regard as unrealistic assumptions-about discretionary spending, payments to physicians under Medicare, and much else. But for the sake of argument, let's accept the premise. The conclusion still doesn't follow.

In the first place, there is broad agreement across party lines that any changes in Social Security and Medicare should honor the legitimate expectations of individuals at or near retirement and that individuals now ages 55 and over should be held harmless. So even if we legislated changes in these programs tomorrow, they would have no effect until 2023. And if we wait to make changes until then, they would have no effect until 2033, much too late to ward off the debacle I described earlier.

Second, there is an issue of generational equity. As the Public Trustees of the Social Security and Medicare trust funds argued in their 2013 report, it is important to enact a solution soon enough to maximize the number of generations who contribute to solving the problem: "Substantial further delay risks further concentrating the burdens of correcting the shortfall on the younger workers who already stand to be treated less favorably . . ."

Third, acting now offers a wide range of programmatic, social, and fiscal advantages. In a recent statement, Robert Greenstein, president of the Center on Budget and Policy Priorities, put the case this way: "Although Social Security faces no imminent crisis, policymakers should act sooner rather than later to restore its long-term solvency. The sooner policymakers act, the more fairly they can spread out the needed adjustments in revenue and benefits formulas, and the more confidently people can plan their work, savings, and retirement. Acting sooner also helps the budget as a whole by modestly reducing federal borrowing in coming years. This will contribute to helping stabilize the ratio of debt to GDP – a key test of fiscal sustainability – and limit the overall interest costs that we must pay."

And finally, the longer we wait, the more abrupt and draconian the necessary changes will be. A CBO analysis suggests that as a share of GDP, the revenue increases and/or spending cuts needed to stabilize the debt/GDP ratio at current levels would roughly double between 2013 and 2025. That is why, in the words of the Social Security and Medicare trustees, "[T]he amount of time remaining to enact a financing solution that is both reasonably balanced and politically plausible is far less than the amount of time projected before final depletion of Social Security's combined trust funds. . . . Each passing year of legislative inaction reduces the likelihood that a solution can be found that is acceptable to lawmakers on both sides of the aisle."

It's important to understand Social Security's dimensions-and the nature of the problems it now faces. As things stand, program outlays will increase from 4.2 percent of GDP to 6.2 percent by 2037, and from 11.3 percent of workers' taxable earnings to 17.0 percent. According to the program's Trustees, the 75-years actuarial deficit amounts to 2.72 percent of taxable payroll, which is 21 percent of the non-interest income the program is scheduled to receive. By historical standards, these gaps are significant. As the public trustees put it in their recent message, "Social Security's long-term income shortfall is now larger than it has been at any point since before the landmark program reforms of 1983. . . . even if a Social Security solution were enacted today and effective immediately, it would require financing solutions that are substantially more severe than those enacted in the 1983 program amendments."

One part of the program-disability insurance-is in much worse shape. While trust fund backing ordinary retirement payments won't be exhausted until 2035, the disability fund faces depletion in 2016. While the two funds are legally distinct, there are compelling reasons to deal with disability in the context of overall reform.

So what should we do? The principles that should guide reform are reasonably clear and uncontroversial (I hope): the measures we adopt should preserve and enhance the program's universality, progressivity, and intergenerational equity while minimizing the negative consequences for growth and employment.

The following are some steps consistent with these principles:

• When Social Security began, it did not include any state and local government workers. Today, despite legal changes, some remain outside the system. As state and local pension systems come under increasing pressure, these workers are vulnerable. Including all new state and local public employees in Social Security and would give them greater long-term security and would modestly bolster the system's finances during the next few decades.

• Since 1977, the financial structure of Social Security has rested on the assumption that wages subject to the payroll tax would represent about 90 percent of the kinds of earnings covered under the law. For various reasons, including the explosion of wages at the very top, taxable wages now amount to only 83 percent of covered earnings. Over time, we should increase the cap on taxable wages to bring 90 percent of such earnings back under the cap.

• Two measures would increase the progressivity of Social Security while reducing financial pressures on the system. For low-income beneficiaries, we should institute a minimum benefit such that no one qualifying for benefits would receive annual payments amounting to less than 125 percent of the poverty line. And for beneficiaries above the median of earnings, we should change the current formula for calculating initial benefits to reflect, on a sliding scale, consumer prices as well as wages. This would mean that initial benefits for workers in the 30^th percentile would continue to be based entirely on wage inflation; for workers in the 60^th percentile, mainly on wage inflation; for those in the 90^th percentile, mainly on price inflation. Formula parameters would be set to ensure that (a) progressive indexation amounts to no more than half the package needed to restore the system's 75-year actuarial balance; and (b) workers at the top continue to receive higher benefits than those in lower earnings percentiles.

• Because employers make hiring decisions at the margin on the basis of total compensation rather than wages and salaries, steep increases in the payroll tax are likely to depress employment growth. While the system needs more revenue, raising the payroll tax rate is not the best way to go. Along with many others, I favor a broad-based carbon tax yielding a revenue stream sufficient at least to keep the payroll tax rate where it is now, and preferably to roll it back. I do not favor eliminating the payroll tax altogether, however, because the link between individual contributions and individual benefits strengthens the moral basis of Social Security-and the program's political support.

​Medicare raises more complicated issues, in part because of the diversity of its funding structures. One program-Hospital Insurance (HI)-includes a trust fund that will be exhausted in 2026. The other big parts of Medicare-for physician and outpatient services and for prescription drugs-are funded in part by premiums but mostly by general revenues. As CBPP's Paul Van de Water points out, these programs aren't and can't go "bankrupt"; the term simply doesn't apply to the way they are funded.

That doesn't mean that they pose no fiscal challenges. As the Trustees put it in their 2013 report, "Concern about the long-range financial outlook for Medicare and Social Security often focuses on the depletion dates for the HI and OASDI trust funds-the times when the projected trust fund balances under current law will be insufficient to pay the full amounts of scheduled benefits. A more immediate issue is the effect the programs have on the unified Federal budget prior to depletion of the trust funds. . . ."

The Trustees proceed to quantify this concern: "In 2013, the projected difference between Social Security's expenditures and dedicated tax income is $79 billion. For HI, the projected difference between expenditures and dedicated tax and premium income is $26 billion. The projected general revenue demands of SMI [Parts B and D of Medicare] are $239 billion. Thus, the total General Fund Requirements for Social Security and Medicare in 2013 are $344 billion, or 2.1 percent of GDP. Redemption of trust fund bonds, interest paid on those bonds, and transfers from the General Fund provide no new net income to the Treasury, which must finance these payments through some combination of increased taxation, reductions in other government spending, or additional borrowing from the public. . . . [T]he difference between cost and revenue . . . will grow rapidly through the 2030s as the baby boom generation reaches retirement age . . ., equaling 4.5 percent of GDP by 2040."

I have no comprehensive cure for this expanding revenue gap. But I do have a radical proposal for a piece of it-and a thought-experiment for the rest. Whether they know (or care to admit it), most highly educated professionals are in the highest earnings decile, and many of them are on the verge of qualifying for membership in the dreaded "1 percent." I can think of no reason why individuals who earn much less should be asked to subsidize high earners' health care in retirement. For those at the top, the benefit of Medicare should be guaranteed issue-period. When they sign up for Part B and Part D, the government should calculate the investment value of the payroll tax contributions they made during their working lives, and then set premium levels to close the gap between those contributions and the actuarial value of the benefits they are expected to receive during their retirement.

During the recent struggle for universal health insurance, "Medicare for all" was a rallying-cry in some quarters. If the Affordable Care Act works well, I wonder whether in the long run, "Obamacare for all" might not be the best way forward-guaranteed issue, a choice among programs, and progressively structured subsidies, backed by strategic regulations. The principled argument for maintaining two health insurance systems-one mainly for retirees, the other for working age families-is not evident, at least to me. But I await instruction from others.

Readers have no doubt noticed that I've focused on entitlements as a policy challenge and have said nothing about political strategy. I'm perfectly willing to engage on the political issues. But I think our first task is to think through the policies that can sustain the programs we care about in a manner consistent with our principles. While it's naïve to imagine that good policies are always good politics, it's even more naïve to imagine that the short-term imperatives of coalition management will be consistent with the policies we need to solve the long-term problems we confront.