Government spending on the elderly continues to climb.  Fueled by rapid growth in the number of Americans over age 65 and increased spending on benefits per person, public expenditures devoted to the elderly continue to edge up. A crucial question for future policy making is whether rising outlays on programs for the aged will squeeze out spending on programs for children, especially investments in their schooling. Many pessimists think this outcome is inevitable, and they urge us to reduce government commitments to the elderly to make room for spending on the young.

Federal spending is especially concentrated on the elderly. The Urban Institute publishes annual estimates of federal outlays on children and adults over 65. The estimates inevitably show a huge imbalance in spending on the two groups. In 2011, federal spending for the elderly amounted to almost $28,000 per person over 65.  In the same year, per capita spending on Americans under 19 amounted to just $4,900 per person. This means aged Americans received $5.72 in federal spending for every $1.00 received by a child 18 or younger.

The Urban Institute’s latest estimates show that federal spending on youngsters has trended down in recent years.  After reaching a peak of about $500 billion in 2010, expenditures on children fell 7 percent by 2012, and they have remained unchanged since then.

Future prospects are not encouraging. Urban Institute analysts predict that from 2014 to 2025, only 2 percent of federal spending growth will go to children. Almost 60 percent will be swallowed up by additional outlays on Social Security, Medicare, and Medicaid. Spending on many federal programs that provide benefits to children are financed out of discretionary programs. In contrast, big public programs for the aged seem to run on automatic pilot, with spending linked to changes in the cost of living and the size of the population past 65. Spending on most domestic discretionary programs is expected to be severely constrained as a result of Congressionally imposed budget caps. This is bad news for many federal programs targeted on children.

Focusing solely on federal government spending gives a misleading picture, however. While federal spending is heavily concentrated on the elderly, state and local spending tilts toward programs that help children, notably, through public school budgets.  Whereas aged Americans receive $5.72 in federal spending for each $1.00 received by someone under 19, those under 19 receive $10.11 in state and local spending for each $1.00 received by someone who is 65 or older. To be sure, total federal spending is considerably greater than that of state and local governments, but the imbalance of public spending on the young and the old is less extreme than federal budget statistics suggest.

Government spending on the aged is high because legislators (and voters) decided to establish government-backed pensions—through Social Security—in the 1930s and government-guaranteed health insurance for the elderly—through Medicare—in the 1960s.  In view of the overwhelming and enduring popularity of these two programs, most voters appear to think this was a sensible choice.  One implication of the policies is that Americans past 65 derive a sizable percentage of their retirement income, and an even bigger share of their health care, from public budgets.

The nation has not made an equivalent commitment to support the incomes or guarantee the health insurance of Americans under 65, except in special circumstances.  Those circumstances include temporary unemployment, a permanent work disability, and low household income.  Families headed by someone under 65 are expected to derive their support mainly from their jobs and from their own savings.  If non-aged families prosper, government spending on them falls.  If instead breadwinners become disabled or lose their jobs, government spending will increase as a result of higher disability payments, unemployment and food stamp benefits, and public assistance rolls.

Nearly all children are raised in families headed by someone under 65.  The government benefits they receive, except for free public schooling, increase in bad times and should decline when the unemployment rate falls.  The Urban Institute’s numbers are instructive.  Between 2007 and 2011, real federal spending on children increased 27 percent, or more than 6 percent a year, as the unemployment rate soared in the Great Recession. Federal spending on children then fell as unemployment—and outlays on government transfer payments—shrank.  For many categories of public spending on children, we cannot assume that lower spending signals a weaker commitment to children’s well-being. Instead it may signal a healthier private economy, a lower unemployment rate, and faster improvement in breadwinner incomes.

Of course, some components of government spending on children do not automatically rise in a slumping economy or shrink when breadwinners’ earnings improve.  Public investments in children’s preschool and K-12 education should be adjusted to reflect the needs of children for compensatory instruction and the expected payoff of added investment in schooling.  Statistics on public school budgets show that spending per pupil has increased considerably faster than inflation and faster than GDP per person over the past seven decades (see Chart 1). Whether spending has increased as fast as warranted is debatable, but rising government spending on the aged has not caused per-pupil spending on K-12 schools to shrink.

Government spending on children’s health has also increased over time as public insurance for children has been expanded.  In 2014 just 6 percent of Americans under age 19 lacked health insurance for the entire year.  The only age group with higher health insurance coverage was the population past 65, which is covered by Medicare (see Chart 2).  The main explanation for rising insurance coverage among children is that federal and state health insurance programs have been expanded to cover most low-income children.  Insurance coverage of children can and should be improved, but a sizeable expansion of public insurance has occurred despite the increase in public spending on the elderly.

The presumption that rising outlays on programs for Americans past 65 must come at the expense of spending on children rests on the unstated assumption that voters will zealously defend programs for the aged while tolerating cuts in programs that fund education, income protection, and health coverage for the young.  The trend toward higher public spending on the elderly has been underway for at least five decades, but the predicted cuts in spending on the young have yet to materialize.

Editor's Note: this op-ed first appeared in Real Clear Markets.

In last Thursday’s GOP debate, Marco Rubio, Ted Cruz, Jeb Bush and Chris Christie managed to avoid mentioning their common proposal to “reform entitlements” by raising the Social Security retirement age from 67 to 70. That was probably a good idea. Their proposal only demonstrates their lack of understanding about the demographics of older Americans, especially the dramatic disparities in their life expectancy associated with education and race. Recent research on life expectancy indicates that their proposed change would effectively nullify Social Security for millions of Americans and sharply limit benefits for many millions more.. While many people in their 30s and 40s today can look forward to living into their 80s, the average life expectancy for the majority of Americans who do not hold a college degree hovers closer to 70, or the average life expectancy for all Americans in 1950.

The Widening Inequalities in the Life Spans of Americans

This research, summarized recently in a study published in Health Affairs, found that life expectancy for various age cohorts of Americans is closely associated with both educational achievement and race. For example, the average life expectancy for college-educated American men who were age 25 in 2008, or age 33 today, is 81.7 years for whites and 78.2 years for blacks. (Table 1, below) By contrast, the projected, average life span of high-school educated males, also age 25 in 2008 or 33 today, is 73.2 years for whites and 69.3 years for blacks. Women on average live longer than men; but similar disparities based on education and race are evident. The average life expectancy of women age 25 in 2008, or 33 years old today, was 79 years for whites and 75.4 years for blacks for those with a high school diploma, and 84.7 years for whites and 81.6 years for blacks with college degrees. Most disturbing, the average life expectancy of Americans age 25 in 2008 without a high school diploma is just 68.6 years for white men, 68.2 years for black men, 74.2 years for white women, and 74.9 years for black women. Surprisingly, the researchers found that Hispanics in this age group have the longest life expectancies, even though they also have the lowest average levels of education. Since these data are anomalous and may reflect sampling problems, we will focus mainly on the life-expectancy gaps between African American and white Americans.

Tallying How Many People Are Adversely Affected

Census data on the distribution by education of people age 25 to 34 in 2010 (ages 30 to 39 in 2015) provide a good estimate of how many Americans are adversely affected by these growing differences. Overall, 56.3 percent of all Americans currently in their 30s fall are high school graduates or left school without a high school diploma, educational groups with much lower average life expectancies. (Table 2, below) More precisely, 10.1 percent or almost 4.8 million Americans in their 30s today lack a high school diploma, and 46.2 percent or 18.9 million thirty-somethings have high school diplomas and no further degrees. All told, they account for 23,702,000 Americans in their 30s; and among older Americans, the numbers and percentages are even higher.

Since race as well as education are major factors associated with differences in life expectancy, we turn next to education by race (Table 3, below). The totals differ modestly from Table 2, because Census data on education by race cover ages 30-39 in 2014, while Table 2 covers age 30-39 in 2015 (ages 25-34 in 2010).

• Among people in their 30s today, 45.4 percent of whites or 10,613,000 Americans have a high school degree or less – and their average life expectancy is 9.4 years less than whites in their 30s with a college or associate degree.
• Among people in their 30s todays, 64.4 percent of blacks or 3,436,000 Americans have a high school degree or less – and their life expectancy is 8.6 years less than blacks in their 30s with an B.A. or associate degree, and 11.6 years less than whites with a college or associate degree..
• Among people in their 30s today, 75.6 percent of Hispanics or 6,243,000 Americans have a high school degree or less – and their life expectancy is 5.0 years less than Hispanics in their 30s with a college or associate degree.

As a policy matter, these data tell us that across all communities—white, black, Hispanic—improvements in secondary education to prepare everyone for higher education, and lower-cost access to higher education, can add years to the lives of millions of Americans.

Preserving Meaningful Access to Social Security Benefits

The widening inequalities in average life expectancy associated with race and education have more direct policy implications for Social Security, because the number of years that people can claim its benefits depends on their life spans. The growing inequalities in life expectancy translate directly into growing disparities in the years people can claim Social Security benefits, based on their education and race. Assuming that Americans in their 30s today retire at age 67 (the age for full benefits for this age group), they can expect to claim retirement benefits, on average, ranging from 1.2 years to 19.3 years, based on their education and race. (Table 4, below)

The most pressing issues of life expectancy and Social Security involve white males, black males, and black females without college degrees: Among Americans age 33 today, white and black men without high school diplomas and black males with high school degrees, on average, can expect to live long enough to collect benefits for less than three years. Similarly, white and black women without high school diplomas and black women with high school degrees, on average, can expect to collect benefits for less than eight years. Together, they account for 25.2 percent of whites and 64.4 percent of blacks in their 30s today. By contrast, male and female white college graduates age 33 today, on average, can expect to collect Social Security for between 14.7 and 17.7 years, respectively; and 33-year old black men and women with college degrees, on average, will claim benefits for 11.2 to 14.6 years, respectively.

These findings dictate that proposals to raise the Social Security retirement age should be rejected as a matter of basic fairness. Among this year’s presidential hopefuls, as noted earlier, Ted Cruz, Marco Rubio, Jeb Bush and Chris Christie all have called for raising the retirement age to 70 years. Under this policy, black men in their 30s today without a college degree and white men now in their 30s without a high school diploma, on average, would not live long enough to collect any Social Security. The change would reduce the average number of years of Social Security for Americans in their 30s today,

• From 1.6 years to -1.4 years for white men with no high school diploma,
• From 1.2 years to -1.8 years for black men with no high school diploma, and
• From 2.3 years to – 0.7 years for black, male high-school graduates.

Furthermore, among Americans in their 30s today, white and black women without a high school diploma, white male high school graduates, and black female high school graduates, would live long enough, on average, to collect Social Security for just 3.2 to 5.4 years. The GOP change reduce the average number of years of Social Security for Americans in their 30s today,

• From 6.2 years to 3.2 years for white, male high school graduates,
• From 7.2 years to 4.2 years for white women with no high school diploma,
• From 7.9 years to 4.9 years for black women with no high school diploma, and
• From 8.4 years to 5.4 years for black, female high-school graduates.

All told, proposals to raise the retirement age to 70 years old would mean, based on the average life expectancy of Americans in their 30s today, that 25.2 percent of whites in their 30s and 64.4 percent of blacks of comparable age, after working for 35 years or more, would receive Social Security benefits for 5.4 years or less.

Social Security faces a financial shortfall. If Congress does nothing about it, current projections indicate that benefits will be cut automatically by 21 percent in 2034. Congress could close the gap by raising revenues, lowering benefits, or doing some of both. If benefits seem generous, Congress is likely to lean toward benefit cuts more than revenue increases. If they seem stingy, then the reverse.

Given the split between the two parties on whether to cut benefits or to raise them, evidence on the adequacy of benefits is central to this key policy debate. Those perceptions will help determine whether Social Security continues to provide basic retirement income for workers with comparatively low earnings histories and a foundation of retirement income for most others or it will become just a minimal safety-net backstop against extreme destitution?

Down-in-the-weeds disagreements among analysts often seem too arcane for anyone other than specialists. But sometimes they are too important to ignore. A current debate about the adequacy of Social Security benefits is an example.

The not-so-simple question is this: are Social Security benefits ‘generous’ or ‘stingy’? To answer this question, people long looked to the Office of the Social Security Actuary. For many years that office published estimates of something called the ‘replacement rate’—that is, how high are benefits paid to retirees and the disabled relative what they earned during their working years. A 2014 retiree with median earnings had average lifetime earnings of about $46,000. That worker qualified for a benefit at age 66 of about $19,000, a replacement rate of about 41%. Replacement rates vary with earnings. Dollar benefits rise with earnings, but they rise less than proportionately. As a result, replacement rates of low earners are higher than replacement rates of high earners.

As you might suppose, there are many ways in which to compute such ‘replacement rates. Because of analytical disputes on which method is best, the Social Security trustees in 2014 decided to stop including replacement rate estimates in their annual reports.

In December 2015, the Congressional Budget Office (CBO) offered what it considered a better measure of the generosity of Social Security. It estimated that replacement rates for middle income recipients were about 60%–dramatically higher than the 41% that the Social Security Trustees had estimated.

The gap between the estimates of CBO and those of Social Security is even larger than it seems. To see why, one needs to recognize that to sustain living standards retirees on average need only about 75% to 80% as much income as they did when working. Retirees need less income because they are spared some work-related expenses, such as transportation to and from work. Those are only average of course; some need more, some less.

If one believed the SSA actuaries, Social Security provides median earners barely more than half of what they need to be as well off as they were when working. Benefit cuts from that modest level would threaten the well-being for the majority of retirees who are entirely or mostly dependent on Social Security benefits—and especially for those with large medical expenses uncovered by Medicare.

On the other hand, if one accepted CBO’s estimates, Social Security provids more than three-quarters of the retirement income target. Against that baseline, benefit cuts would still sting, but they would pose less of a threat, and not much of a threat at all for most retirees who have some income from private pensions or personal savings.

When the CBO estimates came out, conservative commentators welcomed the findings and cited CBO’s well-established and well-earned reputation for objectivity. They correctly noted that many retirees have additional income from private pensions, 401ks, or other personal savings, and asserted that there was no general retirement income shortage. By inference, cutting benefits a bit to help close the long-term funding gap would be no big deal. Social Security advocates were put on the defensive, hard-pressed to challenge the estimates of the widely-respected Congressional Budget Office.

But earlier this year, CBO acknowledged that it had made mistakes in its Decameter estimates and revised them. The new CBO estimate put the replacement rate for middle-level earners at around 42%, almost the same as the estimate of the Social Security actuaries, not the much higher level that had sent ripples through the policy community. One conservative analyst, Andrew Biggs, who had trumpeted the initial CBO finding in The Wall Street Journal, promptly and honorably retracted his article.

Two aspects of this green-eyeshade kerfuffle stand out. The first is that policy debates often depend on obscure technical analyses that are, in turn, remarkably sensitive to ‘black-box’ methods to which few or no outsiders have ready access. The second is that CBO burnished its reputation for honesty by owning up to its own mistakes — in this case, a whopping overestimate of a key number. Such candor is all too rare; it merits notice and praise.

But there is a broader lesson as well. Technical issues of comparable complexity surround numerous current political disputes. Is Bernie Sanders’ single-payer plan affordable? Will Marco Rubio’s tax plan cause deficits to balloon? To vote rationally, people must struggle to see through the rhetorical chaff that surrounds candidates’ favorite claims. There is, alas, no substitute for paying close attention to the data, even if they are ‘down in the weeds.’

Editor's note: This piece originally appeared in Fortune.

Ahead of Tuesday’s primary elections in Ohio, Florida and other states, the 2016 presidential candidates have been talking about the future of Social Security and its funding shortfalls.

Over the next two decades, the money flowing into Social Security will be too little to pay for all promised benefits. The reserve fund will be exhausted soon after 2030, and the only money available to pay for benefits will be from taxes earmarked for the program. Unless Congress and the President change the law before the reserve is depleted, monthly benefits will have to be cut about 21%.

Needless to say, office holders, who must face voters, are unlikely to allow such a cut. Before the Trust Fund is depleted, lawmakers will agree to some combination of revenue increase and future benefit reduction, eliminating the need for a sudden 21% pension cut. The question is: what combination of revenue increases and benefit cuts does each candidate favor?

The candidate offering the most straightforward but least credible answer is Donald Trump. During the GOP presidential debate last week, he pledged to do everything within his power to leave Social Security “the way it is.” He says he can do this by making the nation rich again, by eliminating budget deficits, and by ridding government programs of waste, fraud, and abuse. In other words, he proposed to do nothing specifically to improve Social Security’s finances. Should Trump’s deal-making fail to make us rich again, he offered no back-up plan for funding benefits after 2034.

The other three GOP candidates proposed to repair Social Security by cutting future pensions. No one in the debate, except U.S. Sen. Marco Rubio from Florida, mentioned a specific way to accomplish this. Rubio’s plan is to raise the age for full retirement benefits. For many years, the full retirement age was 65. In a reform passed in 1983, the retirement age was gradually raised to 66 for people nearing retirement today and to 67 for people born after 1960. Rubio proposes to raise the retirement age to 68 for people who are now in their mid-40s and to 70 for workers who are his children’s age (all currently under 18 years old).

In his campaign literature, Rubio also proposes slowing the future rate of increase in monthly pensions for high-income seniors. However, by increasing the full retirement age, Rubio’s plan will cut monthly pensions for any worker who claims benefits at 62 years old. This is the earliest age at which workers can claim a reduced pension. Also, it is by far the most common age at which low-income seniors claim benefits. Recent research suggests that low-income workers have not shared the gains in life expectancy enjoyed by middle- and especially high-income workers, so Rubio’s proposed cut could seriously harm many low-income workers.

Though he didn’t advertise it in the debate, Sen. Ted Cruz favors raising the normal retirement age and trimming the annual cost-of-living adjustment in Social Security. In the long run, the latter reform will disproportionately cut the monthly pensions of the longest-living seniors. Many people, including me, think this is a questionable plan, because the oldest retirees are also the most likely to have used up their non-Social-Security savings. Finally, Cruz favors allowing workers to fund personal-account pensions with part of their Social Security contributions. Although the details of his plan are murky, if it is designed like earlier GOP privatization plans, it will have the effect of depriving Social Security of needed future revenues, making the funding gap even bigger than it is today.

The most revolutionary part of Cruz’s plan is his proposal to eliminate the payroll tax. For many decades, this has been the main source of Social Security revenue. Presumably, Cruz plans to fund pensions out of revenue from his proposed 10% flat tax and 16% value-added tax (VAT). This would represent a revolutionary change because up to now, Social Security has been largely financed out of its own dedicated revenue stream. By eliminating the independent funding stream, Cruz will sever the perceived link between workers’ contributions and the benefits they ultimately receive. Most observers agree with Franklin Roosevelt that the strong link between contributions and benefits is a vital source of the enduring popularity of the program. Social Security is an earned benefit for retirees rather than a welfare check.

Gov. John Kasich does not propose to boost the retirement age, but he does suggest slowing the growth in future pensions by linking workers’ initial pensions to price changes instead of wage changes. He hints he will impose a means test in calculating pensions, reducing the monthly pensions payable to retirees who have high current incomes. Many students of Social Security think this a bad idea, because it can discourage workers from saving for retirement.

All of the Republican candidates, except Trump, think Social Security’s salvation lies in lower benefit payouts. Nobody mentions higher contributions as part of the solution. In contrast, both Democratic candidates propose raising payroll or other taxes on workers who have incomes above the maximum earnings now subject to Social Security contributions. This reform enjoys broad support among voters, most of whom do not expect to pay higher taxes if the income limit on contributions is lifted. Sen. Bernie Sanders would immediately spend some of the extra revenue on benefit increases for current beneficiaries, but his proposed tax hike on high-income contributors would raise enough money to postpone the year of Trust Fund depletion by about 40 years. Hillary Clinton is less specific about the tax increases and benefit improvements she favors. Like Sanders, however, she would vigorously oppose benefit cuts.

None of the candidates has given us a detailed plan to eliminate Social Security’s funding imbalance. At this stage, it’s not obvious such a plan would be helpful, since the legislative debate to overhaul Social Security won’t begin anytime soon. Sanders has provided the most details about his policy intentions, but his actual plan is unlikely to receive much Congressional support without a massive political realignment. Cruz’s proposal, which calls for eliminating the Social Security payroll tax, also seems far outside the range of the politically feasible.

What we have learned from the GOP presidential debates so far is that Republican candidates, with the exception of Trump, favor balancing Social Security through future benefit cuts, possibly targeted on higher income workers, while Democratic candidates want to protect current benefit promises and will do so with tax hikes on high-income workers. There is no overlap in the two parties’ proposals, and this accounts for Washington’s failure to close Social Security’s funding gap.

Editor’s note: This piece originally appeared in Fortune.

Rumor has it that an economist started hitting the gym after finishing two milestone research papers, in expectation of a Nobel Prize, which is only rewarded to a living person. Almost no one denies that greater expectations translate into healthier behaviors, while the converse rarely enters the health policy discussion: expectations of a less-than-desirable future may lead to unhealthy behaviors, including smoking, excessive drinking, sedentary lifestyles, and drug abuse. The health issues of the deprived may have a deeper root in economics.

Professor Zhu Xi from Shanghai Jiao Tong University and I found evidence of this in our working paper “Affordable Care Encourages Healthy Living: Theory and Evidence from China's New Cooperative Medical Scheme”. Standard economic theory predicts that providing medical insurance encourages unhealthy behavior by mitigating economic consequences. We developed a novel theoretical framework in which the opposite is possible because insurance makes longevity more affordable and thus desirable.

We test the theory utilizing a unique experiment of China introducing the New Cooperative Medical Scheme, unique in its long-term credibility necessary for their proposed channel. This scheme reduces cigarette use by around 9% and bolsters subjective perception of the importance of physical exercise and healthy diet. These effects depend significantly on the number of children and the local culture of elderly care. We can rule out alternative explanations of these robust results. The empirical evidence affirms a causal link between concerns about negative bequest and unhealthy behavior, and how to break it.

Breaking the causal link would not be an easy task, because bringing a brighter future to the deprived would not be. But this does not revoke the necessity of considering this “expectation” mechanism in designing health policies. For example, it is trendy to study how smokers may substitute other tobacco products for cigarettes and the ensuing health consequences. According to our analytical framework, the substitution could be broader, that is, a person expecting a miserable future would consciously or unconsciously resort to other means of shortening life. Case and Deaton, in their sensational paper, pinned down drug and alcohol poisonings, suicide, and chronic liver diseases and cirrhosis as the causes of the rising mortality in midlife among white Americans. The war against tobacco use may be complicated by this potential substitution.

In general, recognizing the source of a problem is the first step in solving it. The association between income and life expectancy in the United States is well identified by a Brookings study by Bosworth and Burke and a paper by Chetty et al. The hypothesis that poverty may rationally trigger unhealthy behaviors and thus shorter life expectancy is under-explored.

Our research suggests that constructing a social safety net – by subsidizing health or old-age insurance, for example – brightens the future and thus promotes healthy living. Libertarians who believe in “from each as they choose, to each as they are chosen” may frown upon the idea of expanding the government for the sake of saving people from their own poor choices. As usual, an argument could be made that the positive externality outweighs the cost. In this case, a better social safety net can make a person more forward-looking and thus more beneficial to the society.

Discovering hidden incentives and mechanisms is one of the primal tasks of economists. Our research suggests, surprisingly, that both the Center of Disease Control and Prevention and the Department of the Treasury are important players in promoting healthy living. Let them be.

Editor's note: This piece was written in response to John McDonough's article in the American Journal of Public Health titled "Shorter lives and poorer health on the campaign trail." Read McDonough's article here. 

McDonough is right about two very important things. First, that in America we have quite dismal outcomes for the enormous amount we spend on health care. And second, that there is a real opportunity for a new political dialog between left and right to take root—though perhaps one that is more of a quiet agreement than a high-profile grand bargain. 

McDonough wisely draws attention in Figure 3 of his editorial to the sharp distinction between the United States and other Organisation for Economic Cooperation and Development countries in the relative proportions of gross domestic product spent on health services and social services. The United States is a lonely outlier because we overmedicalize our approach to health conditions and community health. Generally a blend of social, housing, public health, and other preventive strategies would yield better health results than calling an ambulance—and at a fraction of the cost. Even our higher survival rates after age 75 years is a mixed blessing, as Gawande points out, because expensive and frequent medical interventions may extend age but often not the quality of life.¹ 

The good news, both substantively and politically in this election year, is the growing recognition that addressing the social determinants of health is a key—perhaps the key—to improving health outcomes while slowing the growth in health spending as a proportion of gross domestic product and public spending. McDonough and I agree on that, despite his affection for Bernie Sanders’ utopian Medicare-for-all, which likely would do little to address the underlying cost and outcomes problem. 

So how could a new conversation develop, of the kind both we both would like to see? I think on several fronts. 

First, building on existing collaboration, serious analysts and policymakers on both sides of the political spectrum should explain more extensively how resources currently restricted to either health care or social services and housing should and could be more routinely braided together. Despite some interesting experiments and demonstrations that allow certain health and housing money to be mixed and used creatively, budget restrictions and payment systems generally make this dif- ficult. We could seek to agree on a mixture of legislative action on payments and budgets, and using Medicaid (Section 1115) waivers, to permit money currently available only for medical services to be used instead for housing and social services where that could be shown to improve the health of individuals in a community. 

Second, we could agree on bipartisan steps to allow states to experiment with more creative approaches to alter the blend of strategies they have available to achieve improved health outcomes. Section 1332 of the Affordable Care Act (Pub L No. 111–148) is a start, since it will allow states to propose alternatives to some Affordable Care Act provisions to improve coverage and outcomes without increasing federal costs. McDonough and I agree on using 1332 waivers in this way. But a further step would be legislation to allow states to seek even broader waivers to shift money between health and social service programs. For that to happen, conservatives would have to accept increases in total spending on some social service programs. Progressives would have to accept reductions in health programs and reduce their reluctance to granting states more flexibility. Both would have to accept rigorous evaluation to determine what works and what does not. 

And third, there is an opportunity for agreement on empowering intermediary institutions² in neighborhoods, including charter and community schools, as well as health systems,³ to serve as hubs for integrated approaches to achieving health communities. That approach combines the conservative emphasis on the importance of nongovernmental institutions with the progressive emphasis on community action. Again, systematic evaluation is needed. 

Hopefully there can be cross-party congressional support agreement on these themes, as McDonough notes has occurred in alternative sentencing. But it is unlikely in the election season that such themes will be seized upon by presidential candidates. In my view, that is probably good, because presidential elections are about differences, not path-breaking agreements. Better, during this election cycle, to foster positive conversations that cause such themes to be taken out of the election debates, so that they will have broad support for enactment after the Election Day dust has settled. 

1. Gawande A. Being Mortal. New York, NY: Metropolitan Books; 2015. 

2. Singh P, Butler SM. Intermediaries in Integrated Approaches to Health and Economic Mobility. Washington, DC: The Brookings Institution; 2015. 

3. Butler SM, Grabinsky J, Masi D. Hospitals as Hubs to Create Healthy Communities: Lessons From Washington Adventist Hospital. Washington, DC: The Brookings Institution; 2015.

Editor's note: This piece originally appeared in the American Journal of Public Health. 

Much public focus is being given to a broader role for hospitals in improving the health of their communities. This focus parallels a growing interest in addressing the social determinants of health as well as health care policy reforms designed to increase the efficiency and quality of care while improving health outcomes.

This interest in the community role of hospitals has drawn attention to the federal legal standards and requirements for nonprofit hospitals seeking federal tax exemption. Tax-exempt hospitals are required to provide community benefits. And while financial assistance to patients unable to pay for care is a basic requirement of tax-exemption, IRS guidelines define the concept of community benefit to include a range of community health improvement efforts.

At the same time, the IRS draws a distinction between community health improvement spending–which it automatically considers a community benefit–and certain “community-building” activities where additional information is required in order to be compliant with IRS rules. In addition, community benefit obligations are included in the Affordable Care Act (ACA).

Specifically, the ACA requires nonprofit hospitals periodically to complete a community health needs assessment (CHNA), which means the hospital must conduct a review of health conditions in its community and develop a plan to address concerns. While these requirements are causing hospitals to look more closely at their role in the community, challenges remain. For instance, complex language in the rules can mean hospitals are unclear what activities and expenditures count as a “community benefit.” Hospitals must take additional steps in order to report community building as community health improvement.

These policies can discourage creative approaches. Moreover, transparency rules and competing hospital priorities can also weaken hospital-community partnerships. To encourage more effective partnerships in community investments by nonprofit hospitals:

• The IRS needs to clarify the relationship between community spending and the requirements of the CHNA. 
• There needs to be greater transparency in the implementation strategy phase of the CHNA. 
• The IRS needs to broaden the definition of community health improvement to encourage innovation and upstream investment by hospitals.

Download "Hospitals as Community Hubs: Integrating Community Benefit Spending, Community Health Needs Assessment, and Community Health Improvement" »

Downloads

• Download "Hospitals as Community Hubs: Integrating Community Benefit Spending, Community Health Needs Assessment, and Community Health Improvement"

Editor's note: On March 16, Paul B. Ginsburg testified before the California Senate Committee on Health on fostering competition in consolidated markets. Download the full testimony here.

Mr. Chairman, Madame Vice Chairman and Members of the Committee, I am honored to be invited to testify before this committee on this very important topic. I am a professor of health policy at the University of Southern California and director of public policy at the USC Schaeffer Center for Health Policy and Economics. I am also a Senior Fellow and the Leonard D. Schaeffer Chair in Health Policy Studies at The Brookings Institution, where I direct the Center for Health Policy. Much of my time is now devoted to leading the new Schaeffer Initiative for Innovation in Health Policy, which is a partnership between USC and the Brookings Institution. I am best known in California for the numerous community site visits over many years that I led in the state while I was president of the Center for Studying Health System Change; most of those studies were funded by the California HealthCare Foundation.

The key points in my testimony today are:

• Health care markets are becoming more consolidated, causing price increases for purchasers of health services, and this trend will continue for the foreseeable future despite anti-trust enforcement; 
• Government can still play an effective role in addressing higher prices that come from consolidation by pursuing policies that foster increased competition in health care markets. Many of these policies can be effective even in markets with high degrees of concentration, such as in Northern California.

[1] Gaynor, M., and R. Town, The Impact of Hospital Consolidation – Update, Robert Wood Johnson Foundation Synthesis Report (June 2012).

[2] Baker, L. C., M.K Bundorf and D.P. Kessler, “Vertical Integration: Hospital Ownership Of Physician Practices Is Associated With Higher Prices And Spending,” Health Affairs, Vol. 35, No 5 (May 2014).

[3] Chapin White and Megan Egouchi, Reference Pricing: A Small Piece of the Health Care Pricing and Quality Puzzle. National Institute for Health Care Reform, Research Brief No. 18, October 2014.

Downloads

• Download the testimony
• Download the slides

Fee-for-service payments encourage high-volume services rather than high-quality care. Alternative payment models (APMs) aim to realign financing to support high-value services.

The 2 main components of gastroenterologic care, procedures and chronic care management, call for a range of APMs. The first step for gastroenterologists is to identify the most important conditions and opportunities to improve care and reduce waste that do not require financial support.

We describe examples of delivery reforms and emerging APMs to accomplish these care improvements. A bundled payment for an episode of care, in which a provider is given a lump sum payment to cover the cost of services provided during the defined episode, can support better care for a discrete procedure such as a colonoscopy. Improved management of chronic conditions can be supported through a per-member, per-month (PMPM) payment to offer extended services and care coordination.

For complex chronic conditions such as inflammatory bowel disease, in which the gastroenterologist is the principal care coordinator, the PMPM payment could be given to a gastroenterology medical home. For conditions in which the gastroenterologist acts primarily as a consultant for primary care, such as noncomplex gastroesophageal reflux or hepatitis C, a PMPM payment can support effective care coordination in a medical neighborhood delivery model. Each APM can be supplemented with a shared savings component.

Gastroenterologists must engage with and be early leaders of these redesign discussions to be prepared for a time when APMs may be more prevalent and no longer voluntary.

Download "Shifting Away From Fee-For-Service: Alternative Approaches to Payment in Gastroenterology" »

Downloads

• Download "Shifting Away From Fee-For-Service: Alternative Approaches to Payment in Gastroenterology"

According to an August 2015 survey, 72 percent of Americans find drug costs unreasonable, with 83 percent believing that the federal government should be able to negotiate prices for Medicare. Recently, Acting Administrator of the Centers for Medicare and Medicaid Services (CMS) Andy Slavitt commented that spending on medicines increased 13 percent in 2014 while health care spending growth overall was only 5 percent, the highest rate of drug spending growth since 2001.

Some of the most expensive drugs are covered under Medicare’s medical benefit, Part B, because they are administered by a physician. They are often administered in hospital outpatient departments and physician offices, and most commonly used to treat conditions like cancer, rheumatoid arthritis, and macular degeneration. Between 2005 and 2014, spending on Part B drugs has increased annually by 7.7 percent, with the top 20 drugs by total amount of Medicare payments accounting for 57 percent of total Part B drug costs. While overall Part B drug spending is a small portion of Medicare drug spending, the high growth rate is a concern, especially as new expensive breakthrough cancer drugs enter the market and have a negative effect on consumers’ pockets.

Unlike Part D, the prescription drug benefit, there are fewer incentives built in to Part B for providers to consider lower cost treatments for patients even if the lower cost drug may be clinically equivalent to the more expensive drug, because prior to budget sequestration, providers received 6 percent on top of the Average Sales Price (ASP) of the drug. Larger providers and hospitals often receive discounts on these drugs as well, increasing the amount they receive directly on top of the out-of-pocket cost of the drug.

This leads to more out-of-pocket costs for the consumer, as patients usually pay 20 percent of Part B services. The Government Accountability Office (GAO) estimated that in 2013, among new drugs covered under Part B, nearly two-thirds had per beneficiary costs of over $9,000 per year, leading to out-of-pocket costs for consumers of amounts between $1,900 and $107,000 over the year. On top of these high costs, this can lead to problems with medication adherence, even for serious conditions such as cancer.

A New Payment Model

To help change these incentives and control costs, CMS has proposed a new demonstration program, which offers a few different reimbursement methods for Part B drugs. The program includes a geographically stratified design methodology to test and evaluate the different methods. One of the methods garnering a lot of attention is a proposal to lower the administration add-on payment to providers, from current 6 percent of ASP, to 2.5 percent plus a flat fee of $16.80 per administration day.

Policymakers, physician organizations, and patient advocacy organizations have voiced major concerns raising the alarm that this initiative will negatively affect patient access to vital drugs and therefore produce poorer patient outcomes. The sequester will also have a significant impact on the percentage add on, reducing it to closer to an estimated .86 percent plus the flat fee. But we believe the goals of the program and its potential to reduce costs represent an important step in the right direction. We hope the details can be further shaped by the important communities of providers and patients who will deliver and receive medical care.

Geographic Variation

Last year, we wrote a Health Affairs Blog that highlighted some of the uses and limitations of publicly available Part B physician payment data. One major use was to show the geographic variation in practice patterns and drug administration, and we particularly looked at the difference across states in Lucentis v. Avastin usage. As seen in Exhibit 1, variation in administration is wide among states, even though both are drugs used to treat the same condition, age-related macular degeneration, and were proven to have clinically similar outcomes, but the cost of Lucentis was $2,000 per dose, while Avastin was only $50 per dose.

Using the same price estimates from our previous research, which are from 2012, we found that physician reimbursement under the proposed demonstration would potentially change from $120 to $66.80 for Lucentis, and increase from $3 to $18.05 for Avastin. Under the first payment model, providers were receiving 40 times as much to administer Lucentis instead of Avastin, while under the new proposed payment model, they would only receive 3.7 times as much.

While still a formidable gap, this new policy would have decreased financial reimbursement for providers to administer Lucentis, a costly, clinically similar drug to the much cheaper Avastin. As seen in Exhibit 1, a majority of physicians prescribe Avastin, thus this policy will allow for increased reimbursement in those cases, but in states where Lucentis is prescribed in higher proportions, prescribing patterns might start to change as a result of the proposed demonstration.

Source: Author’s estimates using 2012 CMS Cost Data and Sequestration Estimates from DrugAbacus.org

The proposed demonstration program includes much more than the ASP modifications in its second phase, including:

• discounting or eliminating beneficiary copays,
• indication-based pricing that would vary payments based on the clinical effectiveness,
• reference pricing for similar drugs,
• risk-sharing agreements with drug manufacturers based on clinical outcomes of the drug, and
• creating clinical decision tools for providers to help develop best practices.

This is all at the same time that a new model in oncology care (OCM) is being launched, which could help to draw attention to total cost of care. It is important that CMS try to address rising drug costs, but also be sure to consider all relevant considerations during the comment period to fine-tune the proposal to avoid negative effects on beneficiaries’ care.

We believe CMS should consider offering a waiver for organizations already participating in Center for Medicare & Medicaid Innovation (CMMI) models like the OCM, because financial benchmarks are based on past performance and any savings recognized in the future could be artificial, attributable to this demonstration rather than to better care coordination and some of the other practice requirements that are part of the proposed OCM. Furthermore, because this demonstration sets a new research precedent and because it is mandatory in the selected study areas rather than voluntary, CMS must try to anticipate and avoid unintended consequences related to geographic stratification.

For example, it is possible to imagine organizations with multiple locations directing patients to optimal sites for their business. Also, without a control group, some findings may be unreliable. The proposed rule currently lacks much detail, and there does not seem to be enough time for organizations to evaluate the impact of the proposed rule on their operations. Having said that, it will be important for stakeholders of all types to submit comments to the proposed rule in an effort to improve the final rule prior to implementation.

The critical question for the policymakers and stakeholders is whether this model can align with the multitude of other payment model reforms — unintended consequences could mitigate all the positive outcomes that a CMMI model offers to beneficiaries. Helping beneficiaries is and should be CMS’ ultimate obligation.

Editor’s Note: This analysis is part of The Leonard D. Schaeffer Initiative for Innovation in Health Policy, which is a partnership between the Center for Health Policy at Brookings and the USC Schaeffer Center for Health Policy and Economics. The Initiative aims to inform the national health care debate with rigorous, evidence-based analysis leading to practical recommendations using the collaborative strengths of USC and Brookings.

The passage of the Medicare Access and CHIP Reauthorization Act (MACRA) just over a year ago signaled a strong and unique bipartisan agreement to move towards value-based care, but until recently, many of the details surrounding how it would be implemented remained unknown. But last week, the Centers for Medicare and Medicaid Studies (CMS) released roughly 1,000 pages that shed more light on how physician payment will hopefully dramatically change for the better.

Some Historical Context

Prior to MACRA, how doctors were paid for providing care to Medicare patients was subject to a reimbursement formula known as the Sustainable Growth Rate (SGR). Established in 1997 to control the rate of increase in spending on physician services, the SGR pegged total spending among all Medicare-participating physicians to an overall budget target. Yet in this “tragedy of the commons,” no one physician benefitted from her good stewardship of health care resources. Total physician spending often exceeded the overall budget target, triggering reimbursement rate cuts. However, lawmakers chose to push them off into the future through what were called “doc fixes,” deferring the rate cuts temporarily. The pending cut rose to over 21 percent before MACRA’s passage as a result of compounding doc fixes.

Moving Forward with MACRA

When it was signed into law on April 16, 2015, MACRA ended the SGR, its cuts, and many previous payment incentive programs. In their place, MACRA established two overarching payment incentive schemes for providers to choose from:

1. the Merit-Based Incentive Payment System (MIPS) program, which supplants three previous payment incentives and makes positive or negative adjustments to a physician’s payment based on her performance; or


2. the Alternative Payment Model (APM) program, which awards a 5 percent bonus through 2024—with higher annual payment updates thereafter—for having a minimum percentage of Medicare and/or all-payer revenue through eligible APMs. Base physician fee rates for all Medicare providers would be updated 0.5 percent for each of the first four years, followed by no increases until 2026, when base fees would increase at different rates depending on the payment incentive program in which a physician participates.

MIPS addresses providers’ longstanding complaints that reporting that reporting under the existing programs—the Physician Quality Reporting System, the Value-Based Modifier, and Meaningful Use — is duplicative and cumbersome. Under the new MIPS program, physicians report to the government payer directly (CMS) and receive a bonus or penalty based on performance on measures of quality, resource use, meaningful use of electronic health records, and clinical practice improvement activities. The bonus or penalty physicians may see starts at 4 percent of the fee schedule in 2019 (based on their performance two years prior—in this case 2017) and increases successively to 5 percent in 2020, 7 percent in 2021, and 9 percent from 2022 onward. From 2026 onward, MIPS providers would receive an annual increase of 0.25 percent on their base fee schedules rates.

In contrast, the APM incentive program awards qualifying physicians a fixed, annual bonus of 5 percent of their reimbursement from 2019- – 2024, and provides that their fee schedule rates grow 0.5 percentage points faster than those of MIPS in 2026 and beyond, in recognition of the risk they assume in these contracts.

Yet, according to MACRA, not all APMs are created equal. APMs eligible for this track must use quality measures similar to those of MIPS, ensure electronic health records are used, and either be an approved patient-centered medical home (PCMH) or require that the participating entity “bears more than nominal financial risk” for excessive costs. Then, in order to receive the APM track bonus, physicians must have a minimum of 25 percent of their revenue from Medicare come through eligible APMs in 2019, with the minimum increasing through 2023 up to 75 percent. In 2021, a new all-payer Advanced APM option becomes available, allowing providers in APM contracts with other payers to participate in the Advanced APM incentive. To do so, they must meet the same minimum thresholds—50 percent in 2021, 75 percent in 2023—but through all provider contracts, not solely Medicare revenue, while still meeting a significantly lower Medicare-specific threshold. By creating an all-payer option, CMS hopes to enable greater provider participation by allowing all payer revenue to count toward the same minimum threshold. Under the all-payer model in 2021, for example, providers must have no less than 25 percent of Medicare revenue through Advanced APMs and 50 percent of all revenue through Advanced APMs.

MACRA Implementation Details Revealed

The newly released proposed rule provides answers to significant questions that had been left unanswered in the law surrounding the specifics of implementation of MIPS and the APM incentives. At long last, providers are gleaning insight into how CMS intends to implement MIPS and the APM track. Given the fast-approaching MIPS performance period in January 2017, here are three key highlights providers need to know:

1. Qualifying for the APM incentive track—and getting out of MIPS—will be difficult. In order to qualify for the bonus-awarding Advanced APM designation, APMs must meet the “nominal financial risk” criteria, which will be measured in three ways: an APM’s marginal rate sharing for losses, minimum loss ratio (the threshold above which providers would begin sharing in losses), and total potential risk as a percent of expected costs. Clinicians must further have a minimum share of revenue that comes in through the designated APMs.


2. Providers will have fewer opportunities to see and improve their performance on MIPS. Despite calls from provider groups for more frequent reporting and feedback periods, MIPS reporting periods will be annual, not quarterly. This is true for performance feedback from CMS, as well, though they may explore more frequent feedback cycles in the future. Quarterly reporting and feedback periods could have made the incentive programs more “actionable” for providers, alerting them to their performance closer to the time the services were rendered and providing more opportunities to improve performance.


3. MIPS allows greater flexibility than previous programs. Put simply, MIPS is the performance incentive program clinicians will participate in if not on the Advanced APM track. While compelling participation, the proposed MIPS implementation also responds to stakeholder concerns that earlier performance incentive programs were onerous and sometimes irrelevant—MIPS reduces the number of measures required in some categories and allows physicians to select from a set of measures to report on based on relevancy to their practice.

With last week’s release of the proposed rule, the Leonard D. Schaeffer Initiative for Innovation in Health Policy is kicking off a series of work products that will focus dually on further MACRA implementation issues and on translating complex policy into providers’ experience. In the blogs and publications to follow, we will dive into greater detail and discussion of the pieces of MACRA implementation highlighted here, as well as many other emerging physician payment reform issues, as the law’s implementation unfolds.