FROM ISSUE NUMBER 23 ~ SPRING 2015 GO TO TABLE OF CONTENTS

TEXT SIZE A A A

How to Fix Disability Insurance

SCOTT WINSHIP

In late 2016, right around the time we elect the 45th president of the United States, we are due for our first entitlement crisis in two decades. That's when the actuaries in the Social Security Administration project that the Disability Insurance Trust Fund will be depleted. That trust fund — a pool of U.S. Treasury securities purchased with past excess payroll-tax revenue — helps pay for the benefits provided by the Social Security Disability Insurance program now that current payroll-tax revenues are no longer sufficient to cover the program's cost. SSDI insures most American workers and their families against the risk of a long-term illness or injury that would prevent them from working.

If the Disability Insurance Trust Fund is exhausted, we will be in uncharted territory. The Social Security Administration would be statutorily prevented from paying out full benefits on schedule, meaning that benefits received will, in total, be less than promised. Payroll-tax revenue would cover just 81% of scheduled benefits. At the same time, SSDI beneficiaries would be legally entitled to receive full benefits, setting up a dilemma that would probably end in a rushed legislative fix or legal action by beneficiaries.

Legislatively, it could be a relatively simple matter to extend the solvency of the DI Trust Fund by, for instance, reallocating payroll taxes between SSDI and the Social Security retirement program. But doing so would ignore SSDI's more fundamental problems.

SSDI was intended as a program to support those struck with health conditions or injuries that make working difficult or impossible. It continues to serve that vital role. But it has also become a permanent dole for a rising number of adults with limited earning potential who clearly are physically able to work. Reforming SSDI could both slow the growth of program spending and ensure the responsible stewardship of program funds. It could also boost economic growth by reducing federal budget deficits and increasing employment. But most important, SSDI reform could promote personal responsibility and help those who would benefit in the long run from continuing to work, even as it offered a stronger private safety net for older men and women.

A BRIEF HISTORY OF SSDI

The SSDI program was created in 1956. It promised benefits to workers between the ages of 50 and 65 who were disabled. In 1957, around 0.2% of adults received SSDI benefits. Over the next ten years, eligibility was extended to workers under age 50 (and subsequently loosened for workers age 30 or younger), and temporarily disabled workers became eligible. By 1970, about 1.6% of adults between the ages of 25 and 64 were on SSDI.

The share of adults receiving benefits also grew in the early and mid-1970s, but because of program changes not deliberately designed to increase rolls. First, the Social Security Amendments of 1972 made SSDI recipients eligible for health benefits from the new Medicare program after a 24-month waiting period, making SSDI more attractive. Second, the Supplemental Security Income program, aimed at low-income children and disabled adults, was introduced in 1974. SSI applicants had to first apply for other disability benefits, which tended to expand SSDI.

Third, the deep recession of 1974-75 increased SSDI applications, demonstrating that the program's size could reflect not only the underlying health of the workforce but the health of the economy. Fourth, Social Security Administration oversight of state disability examiners' initial eligibility decisions declined dramatically early in the decade, with the rate of review falling from around 70% to 5% in 1972. This not only allowed more marginal cases into the program, but it likely increased applications as the public came to understand that it had become easier to get benefits. Reviews of existing beneficiaries' disability status by the SSA also declined by half in the early 1970s, to the same effect. Finally, SSDI awards grew more generous through 1977, which increased the incentive to apply for benefits. Between 1970 and 1980, the median beneficiary went from receiving just under 50% of his past earnings to receiving 70%.

By 1977, 2.7% of adults were on SSDI, a rise of nearly 70% since the start of the decade. In response to the increase, Congress called on SSA to rely less on demographic and labor-market considerations in determining eligibility for benefits. It also amended the Social Security Act in 1977 and 1980 to make benefits less generous, to require that SSA administrators review no fewer than 65% of state-examiner eligibility decisions, and to mandate that the status of beneficiaries with only temporary disabilities be reviewed at least every three years.

The seven years between 1977 and 1984 saw the only significant decline in SSDI receipt over the history of the program. A General Accounting Office report in 1981 had revealed SSA estimates that one in five SSDI beneficiaries was ineligible, at an annual cost of $2 billion. As a consequence of the stepped-up reviews and the other policy changes, by 1989 the share of adults on SSDI had dropped to 2.3%.

The tightening of eligibility, however, produced a fierce political backlash after the widely publicized removal of 490,000 beneficiaries from the program — many of them with mental conditions, and some of them dropped in error. Stories in the press dramatized the consequences of the tightening, highlighting cases that resulted in homelessness and even suicide. Eligibility offices in half the states defied SSA and refused to apply the new standards. Nearly two dozen class-action suits were filed against SSA, and courts in all but one federal district reversed or limited the new SSA policies.

The result was passage — unanimously in both Congressional chambers — of the Disability Benefits Reform Act of 1984. The law — along with regulatory changes in response to court rulings — cemented the liberalized SSDI program that we have today. Receipt of SSDI benefits began a long climb, slowly at first, then accelerating by the end of the decade, to the point where 5% of adults in 2012 were on the program.

SSDI TODAY

Not just anyone is eligible for SSDI benefits. To qualify, applicants' work histories are assessed along with their health. Generally, an applicant must have worked in a job covered by Social Security in at least one calendar quarter in every year after turning 21 years old and before the year of disability onset. Alternatively, someone having worked in a covered job for 40 quarters (ten years) is also eligible. Furthermore, the person must have worked at least 20 calendar quarters during the 40 quarters (ten years) preceding the disability. (The requirement is looser for younger workers.) Today, nearly three-quarters of men between the ages of 20 and 64 are covered by SSDI, as are two-thirds of women.

To qualify for benefits, the worker must be "unable to engage in any substantial gainful activity by reason of any medically determinable physical or mental impairment which can be expected to result in death or which has lasted or can be expected to last for a continuous period of not less than twelve months." The "substantial gainful activity" requirement is formally defined by the SSA as employment that generates a specific level of pre-tax monthly earnings ($1,090 in 2015 for non-blind applicants, $1,820 for the blind), net of work expenses. The impairment must be severe enough that applicants are "unable to do [their] past relevant work or any other substantial gainful work that exists in the national economy."

SSDI benefit levels are tied to a worker's past earnings and updated annually using a cost-of-living adjustment tied to wage growth. Benefits begin only after a five-month waiting period, to discourage payment for temporary injuries. In January 2015, the average monthly benefit for disabled workers was $1,262. As of 2003-04, the average initial SSDI payment was 55% of career earnings, but payment is on a sliding scale: the lower the earnings, the higher the percentage awarded. For the poorest beneficiaries (roughly the bottom 15%), monthly SSDI benefits can replace over 90% of earnings.

In addition to this cash benefit, after two years on SSDI, beneficiaries receive Medicare coverage as well. Among beneficiaries in traditional Medicare, with or without disability, average monthly expenditures amounted to $1,017 in 2013. In other words, SSDI enrollees' Medicare benefits are about as valuable to them as the cash payment.

Nearly nine million former workers received SSDI benefits in January 2015, up from fewer than three million in 1980. Adding the spouses and children of former workers raises the total number of beneficiaries to nearly 11 million. In 2012, SSDI cost $140.3 billion, which was 3.2 times the 1980 level after adjusting for inflation. Add in the cost of Medicare benefits for SSDI recipients, and the 2012 total increases to $208.6 billion, or 3.6 times the 1980 level. While the SSDI program as a whole has never been insolvent, program spending has exceeded revenues since 2009, requiring that trust-fund assets be drawn down. That is to say, SSDI has depended on general federal revenues for the past six years.

WHY HAVE THE ROLLS GROWN?

The share of adults age 25 to 64 receiving SSDI benefits tripled between 1970 and 2010, rising from 1.6% to 5%. It is important to understand that much of this increase stems from changes that should not worry anyone. One important reason for the rise is that more women became eligible for benefits as more of them participated in the labor force for longer periods of time. In the past, relatively few women had worked enough quarters in jobs covered by Social Security. Between 1970 and 2010, SSDI receipt among women rose by a factor of five. The share of men receiving SSDI doubled.

There is another unsurprising demographic factor behind the rise in SSDI receipt. As the baby boomers have grown older, they have moved into age categories with higher disability rates, mechanically pushing up the overall rate of SSDI receipt. The share of covered workers receiving SSDI benefits roughly doubles as people move from age 40 to age 50, and it roughly doubles again as they move to age 60.

Nonetheless, disability expert Mark Duggan has shown that, from 1989 to 2011, just 38% of the rise in SSDI receipt per adult male can be accounted for by the aging of the population and increases in eligibility. The remaining 62% is due to increases after holding age constant.

Some analysts have pointed to an increase in the Social Security full-retirement age as contributing to the longer-run rise in SSDI receipt; whereas in the past, older SSDI beneficiaries could have switched to Social Security old-age pensions at age 65, their counterparts today must wait until age 66. But this cannot explain why SSDI receipt has risen so much among younger men and women. For instance, MIT economist David Autor finds that among men between the ages of 50 and 54, the percentage enrolled in SSDI rose from 4.4% to 6.6% between 1989 and 2009 — an increase of 50%.

Indeed, a study by Mary Daly, Brian Lucking, and Jonathan Schwabish found that 44% to 57% of the rise in SSDI receipt per working-age adult between 1980 and 2011 was due to factors unrelated to demographics or the increase in the retirement age. They found that, after adjusting for these factors, SSDI receipt still rose by about 45% from 1980 to 2010, and they found a roughly 70% increase in adjusted SSDI receipt after 1989. The Social Security Administration's Office of the Chief Actuary estimates that the "age- and sex-adjusted" share of SSDI-insured workers receiving benefits rose from 3.1% in 1980 to 4.6% in 2012, an increase of 48%. The increase from 1995 to 2012 was roughly 30%.

In short, the evidence consistently indicates that about 45% to 55% of the rise in SSDI receipt since 1980 (and 60% since 1989, when the rise in receipt began to accelerate rapidly) has been due to a greater likelihood of adults receiving benefits at any given age. That means that, after adjusting for demographic changes, there has been a roughly 50% increase in SSDI receipt since 1980 and a 60% to 70% increase since 1989.

But this sort of accounting may actually be overestimating the impact of demographic changes. After all, in recent decades a number of factors have pulled in the direction of declining SSDI receipt. Among those arguing that demographics — age and sex — are the primary drivers of the increase in SSDI rolls, the implicit assumption is that men and women of a given age today should be no less likely to need disability benefits than their counterparts in the past.

That assumption, however, ignores the fact that the economic shift from manufacturing, agriculture, and mining to service work has reduced the physical demands on workers. Occupational-injury rates have declined. Advances in medicine have lessened the severity of worker impairments. Jagadeesh Gokhale of the Cato Institute has shown that the share of people with health impairments who also use an assistive device has increased dramatically since 1979. Among 55-year-old men, for instance, this fraction tripled. And not only have the economy and medicine changed — society has, too. In the wake of the Americans with Disabilities Act of 1990, workplaces are more likely to accommodate disabilities than they were in the past.

Furthermore, the available evidence indicates that the health status of Americans has, at worst, been stable over the long run. Richard Burkhauser and Mary Daly found that the share of working-age Americans reporting they were in fair or poor health was unchanged or fell between 1995 and 2007 (even within age groups). They also showed that the share of working-age adults reporting a work limitation changed little from 1980 to the late 2000s, both for all 25- to 61-year olds and within age groups. However, while in 1988 about 35% of working-age male adults with a work limitation reported that they were employed in the previous year, by 2010 that rate had fallen to 23%. In contrast, the share receiving SSDI benefits rose from 33% to 51%.

Most of the broader trends, therefore, cannot account for the expanding SSDI rolls. There are only two trends that stand out as possible (partial) explanations. First, there has been a much-noted rise in obesity and the chronic conditions associated with it. Yet that trend does not show up in the data on self-reported health status or work limitations. Nor have disability awards on the basis of heart disease or diabetes contributed much to the rise in SSDI enrollees.

Second, many claim that mental-health conditions have become more prevalent over time, but the evidence on this point is weak, too. Diagnoses of mental disorders have increased substantially, as have the number of medications prescribed for them. But these trends could simply reflect changing norms such as reduced stigma, closer attention and sensitivity to mental health by health-care providers, and improvements in pharmaceuticals.

While mental-health conditions have not necessarily become more prevalent, the way that we treat mental illness has undergone a revolution. Until the mid-1950s, schizophrenics and others with serious mental conditions were cared for exclusively in state and county mental hospitals, and the mentally ill actually constituted a majority of hospital patients. The next 45 years saw a dramatic decline in the rates at which the mentally ill were institutionalized, as advocates successfully fought to integrate them back into communities, aided by advances in treatment. From a high of 560,000 in 1955, the number of mentally ill hospital residents fell to 50,000 by 2005, with most of the decline occurring over the 1960s and 1970s. Admissions to mental hospitals peaked in 1971, primarily falling in the 1980s and 1990s.

Deinstitutionalization has probably swollen the SSDI rolls to some extent, though many people with severe mental illness are today incarcerated or on the Supplemental Security Income program rather than SSDI. Between 1985 and 2005, the number of SSDI awards for which a mental disorder was the qualifying impairment doubled as a fraction of adults eligible for SSDI. But awards based on infectious disease and diseases of the blood, of the digestive system, of the genitourinary system, and of the musculoskeletal system all more than doubled. In fact, awards based on every category of eligibility rose, with the exception of congenital anomalies and diseases of the circulatory system. Had awards based on mental conditions risen at the same rates as those based on other impairments, instead of SSDI awards rising by 48% as a share of eligible adults, the increase would have been 37%. As we will see, moreover, there are strong reasons to think that the increase in SSDI beneficiaries with mental conditions is a direct result of policy changes.

LOOSENING STANDARDS

In fact, changes in SSDI policies and their implementation have, to a substantial degree, driven the growth of the program by making it easier to be awarded disability benefits and more difficult to be removed from SSDI. Perhaps most important, non-medical factors affecting employability have become increasingly significant in determining who gets and keeps SSDI benefits.

Assessing eligibility for SSDI benefits involves a multi-step process. In the first three steps, SSDI examiners confirm that applicants have enough work experience, that they are no longer working significantly, that they have an impairment that is both severe and will last at least 12 months, and that their specific severe impairment is either included in an elaborate listing of medical conditions maintained by the SSA or is equivalent to a listed condition. Meeting these criteria satisfies the definition of "disability." If an impairment is not included in SSA's listings (or equivalent to one that is), then otherwise-eligible applicants are assessed in two steps to see whether they can do any work given their age, education, and work experience; if they cannot, then they are also deemed to be disabled.

Over time, the number of new SSDI beneficiaries who qualify on the basis of the "vocational factors" assessed in the last two steps has soared. In 1960, fewer than a tenth of new awards were based on vocational factors; that fraction rose to 27% in 1975 before falling somewhat to 23% by 1986. Since then, demographic and work-related factors have accounted for more and more new enrollees. Every year since 2003, at least half of new awards have been based on vocational factors. In 2012, 58% of new beneficiaries qualified not on the basis of their medical condition alone, but because of additional consideration of their demographic characteristics and work experience.

The share of new awards based on vocational factors is largely unrelated to labor-market conditions; however, the number of applications expressed as a percentage of the SSDI-eligible population and the rate at which initial applications are successful have both tended to rise and fall with the unemployment rate. Presumably the prevalence of severe medical impairments has little to no relationship to the state of the labor market. But when the job market is lousy, more people apply for SSDI benefits, and SSDI examiners are more likely to let them onto the program.

Apart from the business cycle, wages for workers with few skills have stagnated over the past 45 years, which has made SSDI look more and more appealing. The average monthly SSDI benefit today is almost exactly what a full-time worker making minimum wage earns before taxes. And that doesn't include the value of Medicare benefits, which has grown as employer health coverage has declined and medical costs have risen.

Reflecting the growing appeal of disability benefits, the increase in SSDI receipt is a major factor behind long- and short-term drops in labor-force participation among men. By my calculations, in 2013, four in ten men between the ages of 25 and 54 who were neither working nor looking for work reported themselves as disabled in the federal survey used to track unemployment. The rise in self-reported disability has accounted for one-third of the decline in labor-force participation among these men since 1969.

Chana Joffe-Walt concluded a 2013 episode of This American Life, a popular public-radio show, by summarizing these developments: "Being poorly educated in a rotten place — that in and of itself has become a disability." This sentiment is no doubt shared by many SSDI examiners and appeals judges. Be that as it may, SSDI is home to an ever-rising number of men and women who, while facing difficult employment prospects, could nevertheless work.

SSDI has thus gradually evolved into a long-term unemployment program, and it is particularly badly designed to play that role. Once awarded benefits, most enrollees remain on SSDI until death or retirement. A study of the 1972 cohort of awardees projected that only 11% would leave the program for work or because their condition improved. While SSDI beneficiaries do not live lavishly, the program affords them a degree of economic security that is difficult to give up. One 53-year-old recipient quoted in a Wall Street Journal piece on SSDI compared being on the rolls to "a blanket covering you," confessing that, "to walk out from it...at my age, it's a little intimidating."

Another indication that eligibility decisions have become more lenient is the increasing share of awards accounted for by medical conditions that are difficult to assess objectively. In 1961, 42% of new SSDI cases involved cardiovascular conditions or neurological disorders, while just 18% involved musculoskeletal disorders (including back pain) or mental conditions. By 2011, those figures had flipped — 19% of new awards involved cardiovascular or neurological conditions while 53% comprised musculoskeletal or mental conditions. While many medical impairments are easily verifiable through diagnostic tests or other objective procedures, problems such as muscle pain or anxiety are much less amenable to examination.

The 1984 SSDI reforms substantially liberalized screening for mental conditions in particular. The eligibility reviews that had been conducted in the preceding years disproportionately affected beneficiaries with such conditions, many of whom were thrown off of SSDI. The new legislation imposed a moratorium on reviews of mentally impaired beneficiaries and required the SSA to develop new standards for determining whether mental conditions prevent work. Regulations implemented in response by SSA the following year expanded the types of mental impairments included in the SSA listings. In addition, the 1984 law forbade eligibility examiners from denying benefits to applicants with mental impairments before a state-employed psychiatrist or psychologist could evaluate them.

Still more policy changes liberalized SSDI eligibility. Since the 1984 legislation, the opinion of an applicant's physician carries far more weight in eligibility decisions. The reforms of that year required SSA to attempt to get evidence from a claimant's physician before using a state-contracted medical examiner to assess his condition. It also required that SSA establish standards around the use of such consulting medical examiners. In response to the law and a series of circuit-court decisions, the SSA adopted the "treating physician rule," in which the opinion of the claimant's doctor is given "controlling weight" when the medical evidence is in question.

The importance of the weight given the treating physician's opinion was well illustrated in the episode of This American Life. In Hale County, Alabama, almost one in four adults receives federal disability payments. Joffe-Walt, after noting the prevalence of back pain there, acknowledged,

But it's confusing. I have back pain. My editor has a herniated disc, and he works harder than anyone I know. So who gets to decide [who cannot work]? Who decides if the story, the condition, is bad enough? Who makes that determination? Well, in Hale County, it basically seems like there's one guy, a man whose name was mentioned in almost every story people told me about becoming disabled, Dr. Perry Timberlake.

Dr. Timberlake, and other physicians like him, may have the best of intentions. But their patient advocacy can have the effect of subverting the SSA definition of "disability." The treating-physician rule has become particularly important in the appeals process, which itself is a major factor behind the bulging SSDI rolls.

ELIGIBLE ON APPEAL

If denied benefits by the initial examiner, applicants may have a second examiner reconsider their eligibility independently. Denied a second time, the claimant may obtain a hearing before an administrative-law judge. If they still don't qualify, claimants can have an Appeals Council review their case again, and if all else fails they can file suit against the Social Security Administration in U.S. district court.

In 2010, 36% of applicants were awarded SSDI benefits at the initial determination, accounting for roughly 74% of all eventual awards. Of those denied at the first stage, 50% appealed the decision. Just 11% of decisions by a second examiner went the applicants' way, making up almost 6% of eventual awards. Three in four applicants denied by the second examiner appealed again. In all, 38% of those denied at the first stage appealed all the way to an administrative-law judge.

The administrative-law judges, in 2010, overturned 54% of the earlier denials, accounting for 20% of all eventual awards. In fact, because many appeals are dismissed without a hearing being held (for example, because of noncompliance or tardy requests), the real rate is much higher; for cases on which they actually ruled, judges overturned 79% of denials.

Of those who still were denied benefits, 37% tried again, this time turning to the Appeals Council. One percent of the time, the Appeals Council reversed the denial of an administrative-law judge, and 18% of the time it remanded the case back to the judge. When all was said and done, around half of SSDI applicants were eventually awarded benefits, including more than 50% of applicants who appealed after initially being turned away.

Many reapply after trying and failing to get SSDI benefits, even after going through the appeals process. Eric French, an economist at the Federal Reserve Bank of Chicago, and Jae Song of the Social Security Administration found that, of those not receiving benefits three years after being assigned to an administrative-law judge, 40% were on SSDI within the next decade.

It is striking that administrative-law judges overturn half of the denials first made upon application, especially since the decisions have (in most states) been reinforced independently a second time. But it is not surprising when you understand key features of this stage of appeals. Administrative-law judges afford no deference to the state examiners who previously denied benefits to claimants. They are under pressure from the SSA to clear cases, and approving benefits is much easier than denying them. In part, that is because the hearings are "non-adversarial," meaning that the government's (and taxpayers') interests are not represented except by the judge who also must represent the claimant's best interests. Claimants are not subject to cross-examination, nor are treating physicians, whose opinions are afforded controlling weight. Physician reports are not sworn under penalties of perjury.

What's more, claimants appealing before administrative-law judges are usually represented by an attorney or someone from a nonprofit organization or state welfare agency. Their representative can receive fees out of the SSDI back-pay they successfully win for the claimant. Search for "disability benefits" on the web, and a flurry of legal and advocacy sites shows up among the SSA pages. A sizable industry has developed to help people obtain SSDI benefits. About 38% of federal disability cases before administrative-law judges in 1977 involved claimants represented by an attorney. In 2010, three in four did.

The result is that SSDI benefits are awarded to many applicants who should not be receiving them. Just last year, to cite an especially egregious example, an SSA investigation found that 44 judges around the country had improperly awarded benefits to 25,000 claimants over seven years, costing the federal government $2 billion. Meanwhile, those who are unsuccessful in their appeals experience months or years outside the labor force and see their remaining skills erode over time. If they faced employability challenges when they submitted an application, they face a far greater problem upon emerging at the end with no SSDI benefits to show for their efforts.

The evidence that a significant number of SSDI applicants and beneficiaries are actually able to work is overwhelming. Among adults with a self-reported work limitation, fewer and fewer are working, and more and more are receiving SSDI benefits. One study compared the 23% of SSDI applicants who had a similar set of characteristics, but did or did not get benefits because of the particular examiner to whom they were assigned. It estimated that, within this group, employment would have been 28 percentage points higher if no one had received benefits compared with employment if everyone had. The more the comparison is targeted toward SSDI applicants with less-severe impairments, the greater the disemployment effect. Another study found that denied SSDI applicants would be 35 percentage points more likely to work if there were no appeal option. Yet another estimated that up to 40% of beneficiaries are "work-oriented."

An issue brief from the U.S. Department of Health and Human Services notes that "research consistently demonstrates that, with the right supports, 40-60% of people with serious mental illness (SMI) can work." And another paper concluded that "rejected male applicants to the Disability Insurance (DI) program who are younger or have low-mortality impairments such as back pain and mental health problems exhibit substantial labor-force attachment" after they are denied benefits.

Sociologist Rourke O'Brien recently published a paper showing that self-reported disability is more common in countries with more generous welfare states, even controlling for demographics and health. Research has found that changes in the strength of the coal industry during the 1970s and 1980s "provide clear evidence that as the value of labor-market participation increases, disability program participation falls." SSDI applications among unemployed workers rise when unemployment benefits end and fall when they are extended. The share of workers applying for and receiving SSDI rises more during recessions than in other periods.

THE AIMS OF REFORM

Against this backdrop, SSDI reform should be driven by six goals. First, there is the immediate imperative of shoring up the DI Trust Fund. Many liberals are resistant to fundamental SSDI reforms that would seek to reduce the long-term costs of the program; they would prefer to simply reallocate the two Social Security payroll-tax rates (for retirement and disability benefits) and be done with reform. SSDI's long-term solvency could be assured by permanently raising the DI payroll tax by 0.4 percentage points. The Obama administration's fiscal year 2016 budget has proposed increasing that tax from 1.8% to 2.7% from 2016 through 2020 and reducing the Old Age and Survivors Insurance payroll tax (which funds Social Security retirement benefits) by the same 0.9 percentage points. According to the SSA Office of the Chief Actuary, this reallocation would leave both trust funds solvent until 2033 (which moves the OASI insolvency date up by one year).

Some proponents of this approach argue that the SSA's chief actuary foresaw in 1994, when the last reallocation between the DI and OASI trust funds occurred, that the DI trust fund would be depleted in 2016, and that therefore the rise in SSDI costs and beneficiaries is simply due to well-understood demographic changes. But this cherry-picks the SSA's projections, as it relies on the one time out of five projections between 1988 and 2005 that the office accurately predicted growth rather than underestimating it. Moreover, SSA projections are revised based on past changes in enrollment, so the earlier unanticipated swelling of the SSDI rolls is incorporated into projections as future anticipated growth. If SSDI rolls should have grown by less in the past, then this methodology papers over the question of whether accurate projections are worrisome.

Recognizing liberal resistance to reform, House Republicans successfully adopted a rule early in 2015 that prevents a payroll-tax reallocation unless the long-term solvency of the combined OASDI trust fund is strengthened. Given the urgency of addressing the DI Trust Fund's insolvency, a reallocation is the simplest way to ensure the short-term balance of the trust fund. One option would be a smaller temporary reallocation than that proposed by the administration. Another would be to adopt the President's proposal but to use much of the revenue from the reallocation to fund the program integrity initiatives and demonstration projects proposed below, which involve front-loaded outlays that will reduce long-term SSDI costs.

A second goal of SSDI reform should be to expand the support options available to able-bodied men and women who can work, harnessing the benefits of markets and demanding personal responsibility. SSDI should not be an early-retirement program for people who don't want to work or even a long-term-unemployment program for those who have fallen on hard times.

The proposals below would make qualifying for SSDI more difficult for those who are not impaired or whose ailments do not impede working significant hours. To accompany these proposals, federal policy changes should facilitate and encourage the creation of markets for private long-term-unemployment insurance (PLUI). These markets would allow individuals to voluntarily purchase policies that pay out to qualifying policyholders a pre-specified monthly benefit for a predetermined duration upon the exhaustion of public unemployment benefits and unsuccessful application for SSDI benefits. Insurers would be able to specify eligibility criteria similar to those used by public unemployment compensation programs — in particular requiring that claimants have paid premiums for a minimum duration before becoming eligible for benefits. They could also structure payouts to offer a lump-sum benefit for moving, educational, or training expenses.

PLUI would not be tax-favored or otherwise subsidized by the federal government, and no one would be forced to purchase it. It would be intended as a way for workers, beginning when they are younger, to responsibly self-insure against the risk that they will become unemployed for periods longer than those covered by public unemployment-compensation programs. Of course, reform of those public programs to facilitate independence would complement PLUI in helping low-skilled workers who experience joblessness. The SSDI proposals below should be viewed with this PLUI offering as a backdrop.

A third goal of SSDI reform should be to reduce the number of Americans applying for benefits. One way to do this would be to increase employer incentives to take two kinds of steps: adopt safety measures that would prevent workers from becoming disabled and to accommodate and rehabilitate disabled workers rather than directing them to SSDI. When employers pay for insurance covering risks faced by their employees, it may be rational for them to expend less effort preventing covered events or deterring employees from filing claims. These perverse incentives can be thwarted by making insurance rates reflect the cost of claims filed by those the employer insures. For this reason, private disability insurance, state workers' compensation programs, and the state-federal unemployment-compensation program all rely on "experience rating" for determining the premiums or tax rates levied on employers. Through experience rating, employers whose workers have more disabilities, accidents, or layoffs pay higher rates than employers with lower insurance payouts.

In contrast, as economists Richard Burkhauser and Mary Daly note, all employers pay the same SSDI payroll-tax rates. That means that employers have lower-than-optimal incentives to prevent their workers from becoming disabled. Perhaps more important, they — and their insurance agents when firms offer private disability insurance — have strong incentives to encourage a disabled worker to apply for SSDI benefits, rather than spending the money to rehabilitate him or accommodate his disability. By experience rating the employer's share of the SSDI payroll tax, so that firms sending an outsized share of workers to the SSDI rolls pay higher rates, more employers could be incentivized to prevent disabilities and to help retain disabled workers, keeping them in the workforce. And it could be done without raising the total payroll taxes paid by employers.

Experience rating would also give employers incentives to make greater use of claims management. Since they would bear some of the cost when their former workers fraudulently obtain SSDI benefits, employers would have a strong financial interest in the eligibility determination process. Under experience rating, they should be able to obtain and submit evidence to be considered by state examiners, and they should be able to contest their former employees in hearings before the administrative-law judge. In other words, the hearings would become adversarial, with evidence submitted by the former employer and cross-examination of the claimant by the firm's representatives. As a result, the number of able-bodied SSDI applicants awarded benefits would fall, and because appeals would have lower odds of success than they do in today's non-adversarial hearings, fewer applicants would appeal initial denials.

Obviously, this would be a major change to SSDI: Beyond the financing change and the impact on eligibility, it would require SSA to hire more examiners, judges, and legal staff to accommodate the extra work that employer evidence would create. It would also strongly affect employers' operating decisions. For these reasons alone, experience rating along these lines should be piloted in a demonstration project rather than rolled out nationally. But there are two further problems with experience rating that should encourage reformers to move slowly. By making it more costly to hire and retain workers who are disabled or at high risk of a disability, experience rating could make finding work more difficult for disabled workers. In setting tax rates, it would be important to assess firms using formulas that took into account workers' ages, occupations, and health statuses. Alternatively, employers could have the option of excluding a small number of their employees from the pool used to determine tax rates. Another potential problem is that employers would shift the costs of prevention, accommodation, and rehabilitation onto their workers, and since lower-paid workers are likely to be disproportionately represented in workplaces with high disability risk, any effects would particularly affect them.

DEFINING DISABILITY

Another way to reduce the number of SSDI applications is to make it less likely that applicants will be awarded benefits if they do not meet a strict definition of disability. That should be the fourth goal of SSDI reform. There are a number of changes at the initial stage of determination that would serve this goal.

One sensible approach would be to incorporate hours worked into the determination of whether an applicant's impairment prevents him from being able to work significantly. Under current law, an applicant or beneficiary can earn up to $1,090 per month without penalty. This translates into greater or fewer hours worked depending on the hourly wage a person earns. At the minimum wage, someone could work 34 hours per week, 52 weeks a year, and still be under the limit. At $10 an hour, he could still work 25 hours per week year-round. On the other hand, someone working for $20 an hour would be able to work only 12 hours per week. Since many SSDI beneficiaries and applicants have low educational attainment — 16% of beneficiaries in 2010 had fewer than 12 years of schooling and 52% had no more than twelve years — they command relatively low wages, meaning that for any given monthly earnings limit, they are allowed to work more hours and still retain SSDI eligibility. An additional requirement for SSDI eligibility, then, might be for applicants and beneficiaries to work no more than 20 hours per week.

The statutory definition of disability emphasizes that an SSDI beneficiary should have an impairment severe enough that it prevents him from working much. It is, therefore, appropriate to apply tougher scrutiny to applicants who have recently received unemployment compensation benefits — which are only allowed to those who are actively looking for work. Having received unemployment compensation after the onset of a disability should automatically bar a person from qualifying for SSDI on the basis of vocational factors unless he can prove that his disability worsened substantially. Otherwise, he would have to qualify because his impairment is included in the SSA listings or is equivalent to one that is.

An additional change that would tighten eligibility is considering vocational factors only if an applicant worked two-thirds of the quarters in the six years before becoming disabled, instead of half of the quarters in the preceding ten years, as is currently required. Applicants being considered because of vocational factors (and not because their impairments are recognized in the SSA listings) would thus be required to show a significantly stronger attachment to the labor force than other SSDI beneficiaries. (Exceptions could be granted for applicants who can prove they spent time out of the labor force taking care of dependents.) Having such explicit rules about when vocational factors may be taken into account also offers the advantage of preemptively limiting the discretion of administrative-law judges hearing appeals.

For applicants or beneficiaries seeking or having eligibility on the basis of mental conditions, such as depression or anxiety, or on the basis of musculoskeletal conditions, such as back pain, greater consideration should be given in eligibility decisions to whether the conditions are amenable to treatment. If they are, examiners should assess whether applicants or beneficiaries have sought out treatment and complied with therapeutic regimens. Many mental conditions are amenable to psychiatric or psychological therapy or medication, and musculoskeletal conditions can sometimes be treated through physical therapy or surgery.

Reform should also require improvements to the "Occupational Information System" used to assess whether an applicant is able to work in at least some occupation despite his medical condition. It is badly out of date; the last update was in 1991, and much of the system was last updated in 1977. Refreshing it would result in more applicants being disqualified on the basis of vocational factors because the number of less-demanding service occupations has increased, and many of them are not even included in the current listing.

Other reforms could reduce the number of awards stemming from appeals before administrative-law judges. Attorneys and other claimant representatives should be held more accountable for the evidence they present at hearings. Physicians and other health-care providers assisting SSDI applicants should also be held more accountable for the opinions they give and the evidence they provide, and the weight accorded that evidence should be lessened.

Former-senator Tom Coburn developed SSDI reform legislation last year that included a number of measures to serve these objectives. Under his bill, "The Protecting Social Security Disability Act of 2014," claimants and their representatives would be unable to withhold any relevant medical evidence at hearings, and claimants would have to certify that they have not done so, subject to criminal and civil penalties. Health-care providers would have to sign a "Medical Consultant Acknowledgement Form" that spells out their legal and ethical obligations, and they would have to certify that they have read and understood it. Along the same lines, the Obama administration's fiscal year 2016 budget proposes holding treating physicians and attorney representatives liable for overpayments to beneficiaries whom they assisted in obtaining fraudulent SSDI payments.

In addition to these measures, the Coburn bill would make it more difficult for disreputable attorneys and physicians to represent a claimant in hearings. And in an important move sure to be controversial, it would drop the "treating physician rule" that gives the views of claimants' doctors controlling weight when evidence conflicts.

A final way to restrict eligibility to SSDI would be to increase funding for program integrity. Additional state disability examiners, administrative-law judges, and legal staff would allow for more time in deciding cases and appeals. This would have the effect of increasing denials because it is faster to simply award benefits than it is to challenge the evidence presented by claimants, their physicians, and their attorneys. Particularly among state examiners, staffing levels are lower than ideal. There is significant turnover, and the recession and its aftermath resulted in state furloughs and hiring freezes.

ENCOURAGING WORK

The fifth goal of SSDI reform should be to make it less attractive and more difficult to receive benefits from the program for those who, while impaired, are able to work. Former senator Coburn's bill would add a time limit for certain beneficiaries. Currently, upon being awarded benefits, new claimants are assigned a designation based on how likely it is that their medical condition will improve. These categories determine how often beneficiaries are reviewed for eligibility. Coburn would modify the categories and require beneficiaries in the least-severe one — those with a condition expected to improve within two years — to reapply for SSDI after three years. (They could initiate the new application before the three years are up, so that there would be no gap in payment if they were reapproved.)

This group is fairly small; in mid-2013, just 4% of disability cases (including SSI beneficiaries) were expected to improve so quickly. Beneficiaries in the second-least-severe category — where improvement is likely within three to five years — would be subject to a full medical review in their fifth year (in contrast to the more cursory mail-in reviews often conducted by the SSA now), and those in the second-most severe category would face full reviews in their seventh year.

A second policy change to discourage SSDI receipt among the able-bodied would be to cut off SSDI benefits once people turn 62. Today, older adults can receive SSDI benefits between the ages of 62 and 66 instead of taking their Social Security retirement benefits early. Full-retirement age is 66, so if not for SSDI, these beneficiaries would have to take early retirement at 62, which would permanently lower their monthly retirement benefits. The SSDI payments from age 62 to 66, however, equal the full retirement benefits, and at age 66 SSDI beneficiaries can switch to retirement benefits and receive the full retirement payments thereafter. Note, too, that these problematic incentives will worsen as the Social Security full-retirement age continues to rise.

Former-senator Coburn proposed that all SSDI beneficiaries eligible for both programs and not expected to see their medical condition improve be moved onto the Social Security retirement rolls. Note, however, that making adults age 62 and older ineligible for SSDI benefits would mean they would not receive Medicare coverage until age 65. Under current policy, they would be required to purchase coverage on the Obamacare exchanges if they did not otherwise have coverage. A less-disruptive alternative to disallowing SSDI for adults older than 61 but younger than federal retirement age would be to reduce SSDI beneficiaries' monthly payments to the amounts they would receive if they were receiving early retirement benefits (and then permanently reduce their monthly retirement payments once they switch to Social Security). SSDI beneficiaries would still qualify for Medicare benefits as in current law.

Another vital step would be to eliminate the 900,000-case backlog of beneficiary reviews plaguing the SSA. The agency estimates that $9 is saved over ten years for every $1 spent on reviews. CBO estimated in 2011 that increasing the SSA budget for program integrity by $4 billion over ten years would save $12 billion over the same period in reduced SSDI, SSI, Medicare, and Medicaid benefits. Skimping on funding for these reviews is penny-wise and pound-foolish. The Obama administration has proposed a dedicated stream of nondiscretionary funding for program integrity beginning in fiscal year 2017 that it estimates would eliminate the backlog by the end of 2019.

A final reform to discourage SSDI receipt among those who can work is to reduce the rate at which benefits increase each year. Annual increases are currently pegged to growth in an index known as the "CPI-W" that measures how much prices have increased for "urban wage earners and clerical workers." Switching the index to the "chained CPI," which tracks prices for a more general population ("urban consumers") and corrects for the overstatement of inflation in the CPI-W, would reduce SSDI outlays by 2% in 2037, according to CBO. A complication is that, because Social Security retirement benefits are updated using the CPI-W, changing the adjustment for SSDI would create a disjunction at retirement because Social Security pensions would grow at a faster pace. SSI benefits would also become somewhat more attractive relative to SSDI. The solution would be to switch to the chained CPI for those programs too, which would lower federal spending by $22 billion in 2022 instead of just $3 billion.

The sixth and final goal of SSDI reform should be to facilitate work among those with serious impairments who nonetheless would like to work. SSDI currently has several features that promote work. Beneficiaries can receive earnings above the statutory limit in nine months over a 60-month period without losing any benefits. After that, they enter a 36-month period during which they can receive SSDI benefits in months where their earnings do not exceed the statutory limit, and they can receive Medicare benefits for 93 months regardless of their earnings. They can purchase Medicare at a reduced premium thereafter. If they end up out of work within five years of their SSDI benefits ending, they enjoy an expedited reapplication process and time-limited benefits upon approval. In addition, the SSA has, over the years, tested several approaches to encouraging work and independence through demonstration projects.

The results of all of these efforts have been disappointing. Only one in 200 SSDI beneficiaries earns above the statutory limit. Cato Institute economist Jagadeesh Gokhale has convincingly argued that a fundamental problem with all the incentives tried so far is that beneficiaries fear losing their SSDI and Medicare benefits if they work while in the program and fear that they will need SSDI again in the future if they work enough to be removed from the program.

Gokhale has proposed a "generalized benefit offset" program to replace current work incentives. SSDI beneficiaries who work would see their SSDI check reduced proportionally, but not dollar-for-dollar, so their total income would rise. Above a set earnings threshold, working SSDI beneficiaries would also receive a wage subsidy to further incentivize working. The subsidy would phase out with higher earnings until it disappeared. So that SSDI beneficiaries are not threatened by the eventual loss of SSDI eligibility from working, they would always be entitled to receive SSDI benefits in the event that their earnings fall or they become unemployed. (They could still be removed from the program if their medical condition improves.) Gokhale's proposal is a hybrid including elements of SSA's Benefit Offset National Demonstration and its proposed Work Incentives Simplification Pilot demonstration.

A problem with Gokhale's proposal is that it would be much more attractive than the current SSDI program to adults who have relatively minor impairments and could work independently; it would offer them the safety net of SSDI and subsidize their work. There are a number of other uncertainties around the proposal as well, so it should be tested first as a demonstration project. Different configurations of the policy should be tested, with or without the wage subsidy and including or excluding SSDI beneficiaries who qualify on the basis of vocational factors. Additional configurations could include or exclude early intervention services such as those provided in the United Kingdom's cost-saving "Pathways to Work" program. This incremental approach is preferable to former-senator Coburn's proposal to simply replace the current SSDI work incentives with Gokhale's proposed reconfiguration.

AN OPPORTUNITY FOR IMPROVEMENT

The SSDI program is a crucial component of the federal safety net. It protects workers from unforeseeable and catastrophic health shocks that prevent them from working. If we were starting with a blank slate and could choose any option for insuring against long-term disability, we might eschew social insurance for more market-based solutions. As it is, however, the politics of disability reform means we are stuck with the current behemoth that is SSDI, which costs the federal government about as much at this point (when including its Medicare costs) as public housing, food and nutrition assistance, the school-lunch program, and cash welfare for non-disabled families combined.

The looming insolvency of the DI Trust Fund offers a rare opportunity to fundamentally reform SSDI in order to ensure its financial sustainability, better target assistance to those with debilitating impairments, help them if they want to work, arrest excess growth in program spending, and encourage able-bodied workers to self-insure against the risk of long-term unemployment. The welfare reforms of the 1990s constitute a remarkable success for conservative reform of the means-tested safety net, and SSDI presents a similar opportunity for reform of social-insurance programs.

As with the earlier welfare reforms, we have a chance to demonstrate that a fiscally viable, smaller safety net that promotes personal responsibility and reflects the value of work and independence can be more beneficial to 21st-century Americans experiencing hardship than the unaffordable and inefficient 20th-century approaches that are the legacy of the New Deal and the Great Society.

Scott Winship is the Walter B. Wriston Fellow at the Manhattan Institute for Policy Research.