Administrative Burden

any context in which the state regulates private behavior or structures how individuals seek public services is a venue in which the state may impose burdens on its citizens.

We focus on the costs that people encounter when they search for information about public services (learning costs), comply with rules and requirements (compliance costs), and experience the stresses, loss of autonomy, or stigma that come from such encounters (psychological costs). All policies that require citizens to engage with the state will, to varying degrees, create such frictions.

the experiences can be minimized in a number of ways. One is to simply reduce burdens, such as trusting someone’s word that they are a citizen rather than requiring a document to prove it. Another is to shift burdens away from the individual, and onto the state, by, for example, requiring eligibility workers to tap into administrative databases to establish whether someone is a citizen.

Many have pushed hard for the expansion of programs and policies that reduce growing inequality. But if those programs bury people in paperwork and fill them with frustration, it undermines not only policy outcomes, but also people’s faith in the capacity of government to do anything right.

The Department of Treasury, by itself, creates 6.7 billion hours of work for taxpayers each year, accounting for nearly 75 percent of the time citizens spend completing federal paperwork. 4 A recent analysis estimated the compliance costs of filing to be about $ 1.2 billion, or 1.2 percent of GDP.

Policies targeted toward the poor are more likely to be burdensome relative to universal policies that all use. Relative to the near 100 percent take-up for more universal programs, such as Social Security and Medicare, take-up rates by eligible beneficiaries of means-tested programs typically aimed at poor people in the United States are much lower: 40 to 60 percent for Supplemental Social Insurance, about 65 percent for the Supplemental Nutrition Assistance Program (SNAP, frequently referred to as food stamps), 30 to 60 percent of Unemployment Insurance benefits, about 50 to 70 percent for Medicaid. 19 For the Earned Income Tax Credit (EITC), a reimbursable tax credit tied to work for low-income earners, the take-up rate is about 80 percent. 20 Aid to Families with Dependent Children (AFDC) had an estimated take-up rate of between 77 to 86 percent. Participation rates declined dramatically after 1990s welfare reform. Its successor, Temporary Assistance for Needy Families (TANF), has a much lower take-up rate, between 42 to 52 percent.

Conservative opposition to simplifying the process of paying taxes is not new. “Taxes should hurt,” declared Governor Ronald Reagan when he opposed the adoption of withholding of state taxes in California. 36 The logic behind this opposition is that when the process of paying taxes generates a more memorable and negative friction, people are more aware of and less supportive of taxation itself.

First, citizens are better off when public programs are designed to be simple, accessible, and respectful of the citizens they encounter. Most people would agree that if the public sector provides a service, it should be one that is visible enough to be seen, simple enough to comply with, and not psychologically taxing. Second, burdens should be minimized to the greatest extent consistent with protecting important public values, such as cost and program integrity. Third, considerations of burden should be evidence based, identifying the multiple values involved and the likely effects of burdens on those values, and informed by logic and empirical evidence rather than by political rhetoric. Fourth, because burdens can affect some groups more than others, we should be especially attentive to costs on those with limited resources or in programs that are specifically to help those with limited means.

a broad framework to understand administrative burdens, built on three types of costs and the individuals’ experiences when they come to interact with government. Learning costs arise from engaging in search processes to collect information about public services: Are there services that can fulfill unmet needs? Would one qualify for them? What are the requirements for the application process? Psychological costs include the stigma of applying for or participating in a program with negative characterizations, a sense of loss of personal power or autonomy in interactions with the state, or the stresses and frustrations of dealing with administrative processes. Compliance costs are the material burdens of following administrative rules and requirements. This is the time lost waiting in line, completing forms or providing documentation of status. It is the money spent on hiring an accountant to do your taxes, or the fees immigrants pay to process paperwork.

The use of information technology in bureaucracies may restrict the capacity of administrators to use their discretion to minimize the effects of burdens. For example, the automation of Food Stamp benefits in Indiana sidelined caseworkers, replacing them with unwieldy technological processes that many claimants struggled with, resulting in a decline in take-up of benefits, even as take-up was increasing in the rest of the country. 15

A more specific definition is that administrative burdens are the learning, psychological, and compliance costs that citizens experience in their interactions with government.

even for prominent programs such as job training and SNAP, about half of eligible nonparticipants believe that they are not eligible. 17 Surveys of nonparticipants suggest that they would apply for programs if they knew for certain they were eligible.

Learning costs also help to explain why those already in one program become more likely to access other services because applying to one program can generate knowledge about others.

Individuals have a basic need for autonomy over themselves and their actions. 33 Processes under which the state imposes burdens act as a source of external direction over individual autonomy. The more forceful that direction and the more at odds with the individual’s intrinsic preferences, the greater the sense of loss of autonomy, which in turn will lower willingness to participate in and satisfaction with the process. Social psychology also points to the importance of procedural justice. 34 Individuals care as much or more about the process of their interactions with the state as they do about the outcome. Procedures perceived as consistent, fair, and equitable are fundamentally important to citizens. Administrative procedures perceived as arbitrary, unfair, and discriminatory leave us unhappy with our interactions with government.

Positive interactions with the state can increase citizen confidence and knowledge of opportunities, and allow them to develop participatory skills. 61 For example, the development of Social Security played a crucial role in converting older Americans from being the least to the most politically active demographic group in U.S. society.

Congress has consistently cut the Internal Revenue Service (IRS) budget since 2010, despite increased demands on the agency. As a result, citizens calling the IRS to get help filing might be greeted with a “courtesy disconnect”—8.8 million callers were disconnected in 2015. In 2010, about 75 percent of calls got through and the average wait time was eleven minutes. By 2015, only 37 percent got through and the average wait time was twenty-three minutes.

the creation or reduction of burdens is a venue where political values—and therefore political processes—play out.

Federal quality control guidelines offered states stronger incentives to avoid overpayment rather than to enroll eligible participants. 77 Contemporary performance evaluations of welfare programs maintain this tradition by rewarding reductions in fraud measures but neglecting beneficiary take-up.

When overt changes are unpopular (such as reducing benefits) or not publicly defensible (such as limiting the franchise), or when law constrains desired policy changes (such as Supreme Court decisions on abortion or federal policies in intergovernmental programs), administrative burdens offer a low-profile alternative, minimizing the need for political and legal processes of consultation and deliberation. 86 For example, directly changing social benefits requires legislative battles that will often be relitigated in court. By contrast, constructing complex, confusing, and time-consuming application procedures is a less visible form of policymaking that can effectively thwart people from accessing benefits, even if eligible by law. Because of these qualities, burdens are especially attractive policy instruments if they achieve goals that political actors are reluctant to explicitly acknowledge; they can also operate unobtrusively in policy areas mired by gridlock.

The unobtrusive nature of administrative burdens, relative to other policy alternatives, results from the combination of three qualities: opacity, controllability, and neutrality. Although issues such as eligibility levels for welfare programs are decided in high-profile debates played out in front of legislators and the media, the seemingly prosaic details such as the length of an application form are more hidden and complex, and their effects less likely to be observed or understood by outsiders. This is the quality of opacity. Controllability means that such details are under the control of administrative actors—the details of programs generally considered to fall into the domain of administrative execution and are delegated to the executive branch. The controllability of rules relates directly to their opacity—as the nature and effect of burdens becomes clear, they become a matter of interest to other political actors, such as the legislature or higher levels of government. A third quality of administrative burdens is their apparent neutrality—changes in burdens can be presented as technical fixes without any specific policy intent, or as facilitating widely accepted political goals, such as the reduction of fraud.

Some populations may find themselves facing burdens with limited third-party supports, reinforcing inequality across groups. For example, those who struggle to manage newly imposed voter ID laws because of difficulties with documentation cannot turn to an obvious nonprofit organization that would offer specialized support. This creates a challenge—and opportunity—for the philanthropic community: identify communities facing significant administrative burdens but without strong networks of support to help minimize the costs they experience.

Higher voter participation tends to increase the share of low-income voters, resulting in more generous welfare services, greater equality, more progressive taxation, and less restrictive welfare participation rules. 1 If policymakers and administrators use their authority to influence who votes in elections, they are also influencing the policies that will be adopted after the election, and the burdens that accompany those policies.

But voting is just the final stage in the election process. As the brief history of the pre–Voting Rights Act era makes clear, registration is the more consequential part of the election process when it comes to burdens. About 10 percent of citizens view registration as difficult, but these numbers are higher among minorities and the young.

How much does registration matter to turnout? One study that analyzed the historical introduction of registration requirements in New York and Ohio counties estimated a 3 to 4 percentage point long-term decline in participation. 19 Another approach is to compare turnout rates for states that allow voters to register on Election Day, versus those that do not, controlling for other factors. Such analyses generate a higher range of estimated effects, from 3 to 7 percentage points.

An alternative to allowing Election Day registration is to let people register in routine interactions with the state, thereby removing the compliance costs that come with a separate trip to a registration site. That was the goal of the NVRA, which was modeled on state laws. The NVRA allowed all citizens to register to vote when they received or renewed a driver’s license, and hence is popularly known as the motor-voter law. A less well-known NVRA requirement is that agencies providing federal welfare benefits—such as the Supplemental Nutrition Assistance Program, Medicaid, and disability benefits—share voter registration forms and provide help in completing them, verifying applicants’ identification and helping them enroll in benefits. However, the requirement to encourage registration through the provision of social benefits has been implemented haphazardly and unevenly across states. 24 Consistent with these two records of implementation, evidence shows that historically the motor-voter program increased turnout but that agency-based registration did not. 25 However, recent legal efforts to compel states to fulfill their statutory responsibilities for agency-based registration have seen millions of new voters added to the rolls, pointing to the untapped potential of this approach.

a relatively small increase in distance to the polling place is enough to lead to a significant decline in turnout, especially among households without a car.

An analysis of the 2004, 2008, and 2012 found that counties with Election Day registration had consistently higher Democratic turnout.

Early voting is an example: early voting tends to be used by wealthier voters, increasing Republican vote share. 63 Early voting sites are also less frequent in locations with high black populations. 64 Those who favor early voting as a means to increase access therefore should focus on the implementation of the law, specifically the location of polling places.

In 2015, Oregon became the first state to put in place automated registration by sharing state DMV records with the secretary of state. Oregon estimated that three hundred thousand new individuals would be added to the rolls. Other states, including California and Vermont, soon followed.

William Faulkner’s observation that “The past is never dead. It isn’t even past.” The history of burdens in the post-Reconstruction South holds striking parallels with the present. Laws and procedures were adopted with clear discriminatory intent and effects.

Pro-life policymakers know they cannot get away with promising to violate constitutional rights, so they offer an alternate value-based justification—protecting women’s health—even as medical research suggests that such claims are unjustified. Abortions are safer than other common and less regulated forms of medical procedures, including colonoscopies, the removal of wisdom teeth, and even childbirth itself. 4 Similar to the administrative burdens placed on voting, the stated goal differs from the actual goal.

Private insurers in eleven states are required to treat abortion as a separate form of medical insurance, offered as a supplemental rider rather than included in primary coverage. Because federal guidelines do not treat abortion as a common health-care event, insurance providers are not required to disclose that they do not cover it.

The lack of education is sort of a lack of power for a lot of people.” 14 A survey of eighteen-to twenty-nine-year-olds confirmed this impression—the majority of those surveyed said they know little or nothing about birth control pills or implants. Half of unmarried women believed that birth control pills could cause cancer, making them less likely to use this method of contraception.

Abortion is the only health procedure where the patient can expect to face a phalanx of protestors at the point of care. The rights of protestors are guaranteed by law, regardless of the psychological burdens it places on women or providers, illustrating that the experience of stigma is partly structured by laws.

The crowding of more appointments and women into fewer clinics creates another problem: it becomes harder to have an early abortion.

Texas has the highest maternal mortality rate in the developed world,

The passage of the ACA, frequently referred to as Obamacare, led to the largest reduction in the uninsured population since the implementation of Medicare in 1965. Between 2013 and 2016, the proportion of the uninsured under age sixty-five dropped from 20 percent to 11.5 percent.

The ACA was designed with the assumption that states would run their own exchanges. The Department of Health and Human Services (HHS) expected that only eight states would participate in the federal marketplace. 2 Instead, only sixteen states chose to run their own exchanges, seven more partnering with the federal government, leaving 59 percent of the uninsured living in states that did not run their own exchange.

Refusing to adopt the state exchanges became a way for those opposed to the ACA to express that opposition. The absence of state exchanges created complications because of the intergovernmental dynamics of Medicaid. Although Medicaid is subsidized by federal dollars, it is run primarily by state governments. Eligibility rules—and insurance regulations—are determined by the state. Thus, having states administer enrollment was consistent with established practice for Medicaid. State-run exchanges could reduce learning and compliance costs for participants by allowing for a one-stop shopping experience. The purchaser could consider private individual insurance plans, subsidies to purchase those plans, and eligibility and enrollment in Medicaid, all in a single venue. In the federal exchange, individuals would be able to buy private insurance, apply for the federal subsidy based on their income, and see whether they might qualify for Medicaid. But there was no clear plan about how Medicaid enrollment, which is administered by states and counties, would work for those using the federal exchange.

Users of the federal exchange experienced frustrating problems and delays but lacked an understanding of what was causing these problems, making it easy for them to point the finger at the Obama administration.

In the days preceding the launch, HHS officials scrambled to find ways to meet unanticipated capacity demands for the federal exchange; they had assumed state exchanges would be managing far more of the demand. HHS also prioritized pushing the website out on schedule, despite evident problems, concerned that congressional critics would use any delay as a basis for pursuing a repeal of the law. Compounding these issues, the day the website launched was the first of a sixteen-day federal government shutdown, which was partly motivated by House-Senate disagreements about delaying the ACA. Only a fraction of pre-shutdown staff was available to identify and address problems. Relatively few pretests and lack of a soft website rollout increased the demands on staff to identify and resolve problems during the rollout.

More than one-third of uninsured adults reported the process of enrolling via the health-care exchanges to be difficult or confusing.

Personal assistance is especially important if online marketplaces provide information written in legalese. The average citizen reads comfortably at an eighth grade level, yet health plans typically assume college-level reading skills, numeric literacy, and familiarity with specialized terms such as deductible, the practical meaning of which varies from one plan to another

the actions of the Trump administration have likely led to enrollment declines. Between 2016 and 2018, enrollment declined from 12.7 to 11.8 million. Although states that ran their own exchanges saw a total decline of 3 percent, representing one hundred thousand people, the decline in federally run exchanges was 9 percent, or nine hundred thousand. Indeed, states that ran their own exchanges saw no decline in enrollment between 2017 and 2018, but those that participated in the federal exchanges saw a drop of five hundred thousand. 54 In short, states found ways to stem the decline of enrollees despite the opposition of the federal government.

More than half of the previously uninsured Americans who received health insurance coverage under the ACA enrolled in Medicaid.

Some online systems are easier to negotiate than others, however. For example, although nearly all states allow individuals to start and stop the application process without losing information already entered, thirty-three states allow individuals to actually upload required documentation and thirty-five allow for the renewal of coverage online. To facilitate outreach, twenty-four states allow advocates or third-party assisters to complete online applications for beneficiaries.

States were given the option to eliminate documentation for some eligibility criteria, including birth dates, state residency, and household composition. 60 For other criteria, including income and citizenship, verification was required. States, however, could access existing administrative data for verification. A federal database was created for this purpose, featuring data from the Social Security Administration, the Internal Revenue Service (IRS), and the Department of Homeland Security. States could also draw on their own sources. Nearly all states use their wage and unemployment compensation data to verify status. About half of states use vital data services and a smaller fraction draw on records from the Department of Motor Vehicles.

Three other options—express lane eligibility, fast-track enrollment using Supplemental Nutrition Assistance Program (SNAP) data, and fast-track enrollment for eligible parents of children already enrolled—have further eased enrollment experiences in some states. 65 Express lane eligibility allows states to automatically enroll children already enrolled in programs like SNAP. For example, a child with an already accepted SNAP application can be enrolled in Medicaid without providing any further information or application materials. Starting in 2013, CMS also allowed states to enroll parents into Medicaid using SNAP data. In states that have expanded Medicaid eligibility, eligibility guidelines for SNAP and Medicaid are nearly identical, showing only slight differences in how income and household size is calculated. States contact individuals based on their SNAP eligibility, ask them if they wish to enroll in Medicaid, and then follow up with a few questions. If eligible, they are immediately enrolled without additional paperwork or procedures, thus reducing duplicative compliance costs. A similar process is in effect for eligible parents of enrolled children. These parents are identified and then contacted, and after providing consent, enrolled. Fast-track enrollment both increases participation among those eligible, as well as reducing administrative financial costs. 66

Individuals are required to work, train for a job, or volunteer at least twenty hours a week. The challenge is documenting this participation. Ten regional workforce boards are tasked with monitoring compliance. 73 People will be required to upload the necessary documentation to verify their activities. But 30 percent of low-income adults most likely to be on Medicaid report they never use a computer, 28 percent say they do not use the internet, and 41 percent do not use email. 74 It is unrealistic to expect that such a population will have the technological literacy to navigate online documentation processes.

Medicare Advantage, which was touted as increasing private-sector participation and competition, thus leading to reduced health-care costs, has produced the opposite effect, costing U.S. taxpayers far more than the traditional, single payer, fee-for-service portion of the program.

The hope that a consumer will use information to choose wisely becomes more distant as more information must be processed across a wider array of choices. A range of studies show that more choices can actually make Medicare consumers worse off. Fewer choices increase older adults’ ability to pick the plans that best meets their needs for the lowest cost. 52 This is not relevant only for entry into the program. A large number of choices increase “stickiness”—that is, the tendency to stay with the default option. 53 Indeed, fewer than one in ten Medicare beneficiaries actually change plans during open enrollment. 54 As fewer people switch, the benefits of market-based competition erode.

A second approach is to expand outreach to assist individuals in picking plans. Currently, support services include online tools on the Medicare website and assistance provided by nonprofit organizations. The online tools are good at taking into account particular health-care needs and filtering the options based on those needs. The problem, however, is that four in ten people age sixty-five and older do not use the internet. 61 Even if they do, online tools are challenging to use, especially for those with cognitive issues. The large majority of older adults need more direct help.

The Supplemental Nutrition Assistance Program (SNAP), colloquially known as food stamps, subsidizes food purchases for half of all Americans over the course of their lives and an estimated one in seven Americans in any given month. Many Americans receive benefits, but generally only for a relatively short time, two-thirds of participants for three or fewer years.

by 2014, approximately 83 percent of eligible individuals received food stamps.

the relative ease of access to SNAP helped the federal government cushion the effects of the Great Recession, not just for the long-term poor but for a much broader swath of Americans as well. As the Great Recession pushed more people into poverty, they became eligible for, and received, SNAP benefits. Recipients spend their benefits quickly. Every SNAP dollar is estimated to generate $ 1.70 in economic activity. 3 Beyond providing income and food for poor households, SNAP also generates long-term positive impacts on children’s health and women’s economic self-sufficiency and has even been found to reduce criminal recidivism.

Because food stamps bring federal support directly into a state economy (there is no requirement for the state to match funding), state policymakers generally have an incentive to expand access

One in four potentially eligible beneficiaries identified a reluctance to answer personal questions as a reason for not applying,

Qualitative studies find that potential applicants see the application process itself to be a greater source of stigma than receiving the benefits:

If you have ever had to deal with the bureaucracy of poverty, of having to prove over and over again to those in charge how fundamentally unworthy you are, you understand that forms are not sacred. . . . There are government agencies that use their forms to try to help you. And there are those that seem to have designed their forms to remind you of the audacity of expecting your government to help you with anything.

Older adults stand out for their relatively low use of SNAP: only about one-third of those eligible actually receive benefits.

An alternative way to reduce learning costs about SNAP among older adults already exists but is not being used. The Social Security Administration (SSA) has interactions with older adults on a regular basis and could offer a more systematic effort to inform those who might be eligible. SSA has ready access to benefit information for individuals and married couples, allowing them to providing reliable case-by-case information. If a total family benefit was below the food stamp eligibility criteria, the SSA could mail information about the SNAP program to potentially eligible older adults. Medicare could also play a more active role. Because older adults have such high out-of-pocket medical care costs and Medicare has information on these costs, they could selectively mail out information about SNAP to individuals with significant health-care expenditures. Even more effectively, administrative linkages between Medicaid and SNAP databases could be more clear to assist caseworkers in verifying medical expenses. 73 More broadly, public-private partnerships could be enlisted to raise awareness of the program among older adults. For example, SSA could partner with the AARP

Ongoing efforts to expand SNAP access via internet applications work less well for older adults. Only four in ten older adults use the internet, and rates are even lower among low-income groups. 74 Most need hands-on assistance in the actual application process.

state governments could simply auto-reenroll individuals with very low incomes who are receiving SSI. The closest they have come to this is the Combined Application Program piloted in eighteen states, which allows SSI beneficiaries in single households to apply for SNAP using a shortened application when they apply for SSI. It also does not require a separate interview or any documentation. These applicants also benefit from a forty-eight-month recertification window. In South Carolina, this change increased the percentage of SSI beneficiaries on SNAP by 12 percentage points. 80 States could also more closely track older adults enrolled in Medicaid with very high medical expenses, and target information and support to these individuals to see whether they are eligible for SNAP—and then provide support for that application process.

a welfare program with low burdens inevitably becomes a target for conservatives generally opposed to the welfare state. The attraction of adding new burdens increases as opponents to SNAP struggle to alter it via traditional legislation that alters eligibility criteria. The federal government could decide to impose a new round of administrative burdens, or to delegate the capacity to do so to states.

We do not demand drug tests of homeowners taking a mortgage deduction, students receiving subsidized loans, or corporate employees provided with tax subsidies.

Medicaid is not just important for providing access to health care. It also matters for inequality. People who lack access to health care are more likely to struggle to break cycles of poverty. As young people gain Medicaid early in life, they are more likely to see long-term benefits to their health and educational attainment. 3 Medicaid can even facilitate intergenerational mobility: because Medicaid is made more available to low-income pregnant women, their children enjoy better economic outcomes as adults.

Medicaid covers about one in five Americans at any given time,

BadgerCare Plus incorporated many family Medicaid programs under one program name and one set of rules, making it easier for people to understand whether they were eligible and reducing the need to apply for multiple programs with multiple applications, thereby reducing learning and compliance costs. This change was a response to the complexity that resulted from decades of adding programs on top of one another—from AFDC to Medicaid to Healthy Start to BadgerCare.

An innovative aspect of ACCESS is that it allowed potential applicants to enter basic information to do a preliminary check of their eligibility. In a 2004 operations memo, the Department of Health Services wrote that the screener tool would give applicants greater confidence of their eligibility: “They may choose not to apply because of incorrect assumptions about their potential level of benefits or specific policies—such as the vehicle asset rule—that have changed over time. Many people believe the application process would involve too much time and effort unless they feel reasonably confident that they will be eligible for benefits.” 32 In September 2005, ACCESS added a Check My Benefits feature to address this problem.

states that have increased enrollment of Medicaid beneficiaries typically engaged in active outreach with partners in the nonprofit and health-care community.

Friedman first proposed a negative income tax in Capitalism and Freedom in 1962, seeing it as a tool to displace other welfare programs while encouraging labor market participation. 1 The negative income tax became the inspiration for the first attempt to pass a wage subsidy, President Richard Nixon’s failed Family Assistance Plan (FAP). From the ashes of this failure, the Earned Income Tax Credit (EITC) was born.

In 2015, almost thirty-two million people benefited from the EITC, claiming $ 67 billion. The EITC achieves an estimated take-up rate of 80 percent of eligible beneficiaries, higher than the 30 to 60 percent typical of means-tested programs.

By tying aid to work, the EITC largely avoids traditional conservative critiques of welfare. 3 Another common explanation is that the EITC is concealed within the tax system, which reduces its visibility—the public simply does not understand it as well as other policies and does not equate tax reimbursements with direct cash payments to poor people. 4 Although useful, these arguments are incomplete, failing to appreciate the degree to which path dependency and the political influence of third parties have shaped the program.

From an employer perspective, the EITC is a subsidy for low wages paid for by taxpayers, whereas increases in the minimum wage are borne directly by business owners. As business interests realized that the EITC offered an attractive political alternative to minimum wage increases, they became strong supporters, helping ensure a level of bipartisan support that eluded other targeted programs.

Another vital interest group is the tax preparation industry, which has structured its business model to take advantage of the EITC and has argued for its expansion. Tax preparers shift EITC burdens away from recipients and onto themselves in exchange for a fee. At the same time, the tax preparation industry has lobbied to maintain enough documentation requirements to ensure that the EITC remains burdensome enough that low-income workers seek professional tax preparation help.

The EITC has reduced income insecurity for many of the poorest Americans. In 2014, the national poverty rate was 15 percent. Without the EITC (along with the refundable portion of the child tax credit), the poverty rate would have been more than 18 percent. The effect on children is even higher—without these tax credits, the child poverty rate would be a third higher, 24 percent rather than 18 percent. 13 Every $ 1,000 increase in the EITC benefits leads to 9 percentage point reduction in poverty among the poorest families. 14 A growing literature suggests that by supplementing the income of the poor, the EITC also has measurable effects on child health, cognitive abilities, and educational performance. 15

The EITC form itself is only a page long, minimizing compliance costs, but the explanation of how to determine eligibility is thirty-seven pages.

About one in three EITC claimants are new each year—meaning that they face greater learning costs than those already familiar with the program.

The IRS sends reminders about the EITC to those who appear to be eligible, which has been shown to generate a 41 percent jump in take-up among initial nonclaimants. 24 It has worked to improve those reminders, learning from experiments that show that reminders featuring simplified language, with estimated benefits, generate greater take-up

Policies are often path dependent—as they are implemented, beneficiaries fight to defend and expand them.

Policymakers first began to see the EITC as an alternative to raising the minimum wage, which—though immensely popular among the general public and cheap for government to implement—was fiercely opposed by business. Businesses, and particularly their lobbying organizations, came to understand that the EITC could be used as a way to fend off minimum wage increases. In short, as path dependency predicts, the policy created its own supporter.

The incentives for employers and their interest groups are clear. Employers must cover the costs of minimum wage increases, but citizens bear all the costs of the EITC. Further, the EITC generates some distinct benefits for businesses. Nearly 20 percent of beneficiaries use their EITC refunds to facilitate their labor force participation—such as buying clothes and cars for employment. 70 To the extent that the effects of the EITC on wages has been studied, they confirm a pattern of restraining real wages, effectively subsidizing employer’s contributions. A 10 percent increase in EITC generosity is associated with a 5 percent fall in wages for high school dropouts, a 2 percent loss for those with just a high school diploma, but negligible effects for those with higher education.

Another profit opportunity for tax preparers is to broker loans from banks that they can sell to filers based on their anticipated return. The IRS initially shared filer debt information to tax preparation firms. This allowed these firms to offer refund anticipation loans to customers without the risk that other creditors could claim the refund. Although refund anticipation loans might be repaid in just a couple of weeks, the associated fees made them highly profitable. At one point, such quick-refund products accounted for 30 percent of the total revenue for Jackson Hewitt Tax Service.

A basic truism of public administration is that different organizations with different histories, different expertise, and different professional backgrounds will approach the same problem in different ways. The IRS did not approach EITC compliance in the same way that traditional welfare agencies would, seeking more documentation before providing benefits. Instead, it approached EITC compliance as another form of tax compliance, which could be best addressed postpayment via cross-checks with other sources of information. The professional culture of the IRS led it to deal with a large population of tax filers by placing much of the verification burden on the government rather than on the individual. As compliance problems were identified, the IRS continued with this approach, with more aggressive cross-checks and use of administrative data to verify participant eligibility. The IRS also sought to change the rules in ways that favored simplification rather than complexity. Such rules both made verification checks easier and reflected an assumption that errors in reporting were at least partly because of confusion arising from the complexity of reporting requirements. Simplified child qualifying criteria were adopted in 1990, and the IRS was given authority by Congress to use Social Security numbers to cross-check dependent claims.

Old-Age and Survivors Insurance (OASI)—more colloquially known as Social Security.

Uncertainty over the constitutionality of the Social Security Act also clouded the implementation process. Some employers simply refused to report earnings, at least partly on the basis of ideological objections, and their resistance was encouraged by the prospect of Roosevelt facing a Supreme Court defeat on the constitutionality of Social Security. 16 The Court had already struck down much of the policy architecture of the New Deal. Indeed, Roosevelt was able to draw staff to support Social Security from the National Recovery Administration only because it had been declared unconstitutional and needed to be disbanded. Most ominously for Roosevelt, the Agricultural Adjustment Act was ruled unconstitutional on the basis of a revenue-raising and spending structure quite similar to Social Security. It was not until May 1937, just seven months before all beneficiaries needed to be enrolled, that the Supreme Court ruled the program constitutional. 17 The favorable judgment was part of the “switch in time that saved nine” when Justice Owen Roberts switched from conservative to the liberal bloc of the Court in the aftermath of Roosevelt’s court-packing proposal.

The spirit of the new organization was built on three broad principles. First, as articulated in the 1941 Social Security Board’s Annual Report, the relationship between the agency and program participants was cooperative: “Social Security represents the collaboration of workers and their employers and the Government of the people.” Collaboration reflected the idea that beneficiaries were not supplicants, but had earned their standing as a partner to government. A second key principle was that participants were deserving of benefits, that “the sacrifices made by these workers in their country’s behalf are manifestly adequate social contributions.” 32 Finally, the principle that Social Security was an insurance program was critical. Although the bureaucrat who was administering traditional public assistance might view the claim of the recipient through the lens of deservingness, insurance suggests compensating a broad population that had collectively contributed to minimize a shared risk.

Enrolling twenty-six million individuals meant keeping track of their earnings over their entire working lifetimes. Newspapers at the time called the new bureau the “biggest bookkeeping organization in the world.”

A major logistical challenge to enrolling millions of people was that thousands of planned regional Social Security offices were not yet in place. The Bureau of Federal Old-Age Benefits improvised. In November of 1936 it employed the Postal Service to enroll employers and employees. With forty-five thousand offices on the ground, daily contact with the public, and a generally positive reputation, the Postal Service offered an efficient and trusted choice that reduced both learning and compliance costs. 50 The Postal Service circulated and collected applications and then distributed Social Security cards. The eligibility process was brought to beneficiaries and employers via a familiar face—their local mailman. Within twenty-eight days of starting, the Postal Service had collected more than twenty-two million of the twenty-six million applications. 51 The success of the Postal Service gave the Bureau of Federal Old-Age Benefits some breathing room to train field office staff that were being hired as the initial enrollments took place.

One group dealt with the general media. Another focused on labor organizations and the labor press. Another division was focused on the business press, as well employers more broadly. An entire publications division existed to generate promotional and educational materials. These were done so well that the New Yorker praised the materials as “a triumph of clarity and simplicity.” 65 The educational division focused on colleges and the educational sector more broadly. An inquiry division answered letters and questions from individuals all over the country. There was even a motion picture division devoted to creating films about the program. Board members themselves were active on the radio giving speeches and doing interviews about the program. 66 The positive messages conveyed by this onslaught of messaging helped to communicate the notion of Social Security as an earned right, thereby reducing psychosocial costs.

SSA staff help individuals enroll in Medicare, its associated programs to help participants pay for out-of-pocket health-care costs, and in the Supplemental Nutrition Assistance Program. Approximately 1,200 field offices are spread across the United States, staffed by eighty thousand employees. They vary significantly in size, smaller offices in rural areas serving fewer than twenty people a day to larger offices in cities serving over five hundred. 71 In aggregate, demand for help either in-person or via phone inquiries remains strong. In 2015, field offices served forty million visitors.

Declining budgets and administrative capacities need to be understood in the context of two big trends. First, SSA’s workload is growing dramatically. In 2002, the SSA served fifty-one million OASI beneficiaries. That increased to sixty-two million by 2015 and is anticipated to jump even higher—to seventy-seven million in 2020 and eighty-five million in 2025.84 The second trend is that the SSA work force, like the clients they serve, are aging, the majority of supervisors eligible for retirement. In contrast to the early SSA investment in training, budget cuts have prevented leadership development programs that could offer a pipeline for future management personnel or succession planning. The SSA has been under a hiring freeze since 2010.85 Given that it takes nearly two years to fully train new staff in the broad array of responsibilities ranging from eligibility determinations to claims reviews, long-term hiring freezes place the agency in a challenging position.

The poverty rate among older adults would be more than 40 percent, versus 9 percent, if Social Security were not included in their income. 89 Moreover, nearly one in three beneficiaries rely on it for 90 percent or more of their entire income.

Social Security’s success—both in terms of poverty reduction and political popularity—is in part a function of the program’s low levels of administrative burden. The program could not be successful if it could not get off the ground. Many experts believed the program was impossible to administer effectively. The rapid speed—and high level of success—of the initial administrative structures that were put in place after the initial 1935 legislation paved the way for the 1939 amendments, which in turn were critical to ensuring the program’s long-term viability by dramatically expanding the number of beneficiaries.

Burdens matter to basic values of governance, such as transparency and rationality. Because the creation, nature, and effects of burdens are often opaque, they undermine transparency. Because debates about burdens usually fail to fully consider value trade-offs, they result in outcomes that are not just less than rational but also sometimes irrational. Put together, such basic failings undermine people’s trust in government.

Shifting burdens away from citizens requires a well-functioning administrative state.

Republicans figured out how to build a politics of burdens that Democrats have been unable or unwilling to develop. Republicans have done so despite the fact that in other areas—such as the regulation of businesses, campaign donations, or firearms—they have also been successful in opposing governmentally imposed burdens in the name of liberty. Their success is not built on an ideological consistency, but in a situational capacity to connect burdens to political values shared by their supporters. Christian conservatives who oppose abortion might support mandatory ultrasounds for women seeking abortions, but bristle at government requirements that employers cover birth control in their health insurance plans.

The policies we identify where burdens are designed to not send negative messages to citizens—the EITC and Social Security—demonstrate two ways to break this general pattern. In the case of Social Security, nearly all Americans benefit from a near universal design, a prioritization of access in implementation, a general conception that the benefits are earned, and an emphasis on rights. 7 In the case of the EITC, burdens were limited partly because the benefit supported the working poor who are classified as deserving, but also because they offered a subsidy to those that employed them, and a market to the tax preparation industry. The “who” in “who gets what, when, where, and how” matters a good deal. Alliances that expand the “who” to include all citizens or powerful business actors facilitate making the “when, where, and how” less burdensome.

To return to the SNAP example, an evidence-based approach should compel policymakers to understand that rules to reduce fraud may also reduce access. Perhaps policymakers already understand this trade-off, but must now formally acknowledge it and consider the scale of the trade-off.

We offer positive examples of how a mixture of administrative data and technology can be used to reduce burdens, but there is also a dark side, in that automated systems might be used to intrusively target people and rob administrators of flexibility to ameliorate burdens when the circumstances require it. 27 Technology can offer both great promise and great risk; it can be used to reduce burdens but can also enhance them. Technology, or any other innovation, will only reduce burdens if that is the goal of those controlling it.