Abstract
Administrative burdens are the frictions that people face in their encounters with public services, leading to meaningful costs that include learning, compliance, and psychological costs. We offer evidence that burdens are a key source and consequence of inequality, resulting in disparate outcomes in people’s access to basic rights. We also detail how these outcomes are patterned by targeting, federalism, bureaucratic pathologies, and the growing use of the private sector and tax system to deliver social welfare benefits. Throughout, we highlight recent and novel contributions, including empirical research in this double issue, that have helped clarify how and why administrative burdens shape inequality. Burdens have not received the political, policy, or research priority that is commensurate with their magnitude or impact on individuals. We conclude by arguing that we need a coherent language and framework to recognize and, where appropriate, reduce burdens across a wide array of policy domains.
Policy debates often (and rightly) center on what people need. What resources, protections, or supports are most critical in their daily lives? We have seen successful policy changes to broaden access to health insurance, college, childcare, special education supports, affordable housing, and other safety net supports. What good are these policies, however, if we cannot access the benefits associated with them?
The gap between people’s needs and the policies that are supposed to provide for them is filled with administrative burdens. Administrative burdens are rooted in laws, organizational rules, and everyday implementation practices. Burdens are a barrier to limit access to everything from formal citizenship to voting rights to the resources required to enjoy social rights, such as education, housing, and health care. Burdens also emerge in the state use of coercive power in contexts such as the criminal legal system, or child protective services (Edwards et al. 2023, this volume, issue 5). Across these and other venues, burdens tend to fall more heavily on more marginalized groups, undermining their claim to citizenship rights. Further, burdens are often not just a result of inattention or lack of capacity, but also the product of deliberate design.
To take one example, encouraged by the Donald Trump administration, Arkansas added a (now defunct) work reporting requirement to its Medicaid program with claims that it would boost employment. Such gains never emerged. Instead, more than eighteen thousand people lost health insurance coverage. One study estimated that an astonishing 95 percent of those who lost coverage were still eligible because they were working or should have been exempted from the work requirement (Sommers et al. 2019; Sommers et al. 2020; Wagner and Schubel 2020). Many were unaware of the new requirements. Those who attempted to report work faced a cumbersome process, including an online-only documentation system that proved difficult to navigate. The impact of these changes was both substantial and unevenly distributed. Racial disproportionalities in the Arkansas Medicaid program made Black program beneficiaries particularly vulnerable (Sommers et al. 2019). Black people make up roughly 15 percent of the state population but 26 percent of Medicaid beneficiaries. More than 47 percent of Black Arkansans rely on Medicaid relative to 25 percent of their White counterparts. As a result, the large-scale loss of Medicaid coverage sparked by work reporting requirements had implications for racial inequity (Michener 2020).
This example illustrates the relevance of our central argument: we need to shift attention from policy potential and promises to concrete access. Understanding the origins, experience, and consequences of administrative burden allows us to make this shift. In doing so, we focus on people’s experience of government as well as the administrative processes that shape those experiences. A fuller understanding of the administrative state and inequality means seeing and measuring burdens as well as their consequences.
To “see” burdens, we need to better examine where they come from and how they operate. This double issue shows us key points of demarcation. One is the distinction between universal and means-tested programs. In the 1980s and 1990s, attention to administrative burdens in policies and programs were largely focused on targeted poverty-based programs such as Aid to Families with Dependent Children. The general view was that targeted policies had high levels of burdens, whereas universal social insurance policies, such as unemployment insurance and Medicare, were relatively more accessible. As we detail, however, the sources of burdens as well as the types of programs they are embedded within cannot be explained by simply focusing on whether policies are targeted or universal. In some cases, more universal programs can have large burdens. Indeed, as Zachary Parolin, Christina Cross, and Rourke O’Brien (2023, this volume, issue 5) demonstrate, increases in burdens in both Temporary Assistance to Needy Families (TANF), a classic targeted policy, and unemployment insurance (UI), a New Deal–era universalist social insurance policy, explain accelerating racial inequalities in take-up of U.S. social welfare benefits.
The presence of multiple actors and conflicting incentives in service delivery also contribute to burdens. Programs managed by a mix of federal, state, and local partnerships are typically more complicated and less accessible. For-profit providers that deliver social welfare benefits further fragment the map of service delivery and sometimes have incentives to make policy benefits inaccessible, whether private insurers participating in the Medicare program or private agencies providing job training to welfare participants.
Another key demarcation in the functioning of burdens include those between rights-granting and rights-depriving programs (Brown 2023, this volume, issue 5; Edwards et al. 2023, this volume, issue 5) as well between state administered and third-party administered programs (DeLuca, Katz, and Oppenheimer 2023, this volume, issue 5; Yu 2023, this volume, issue 4). Even further, comparing the experiences of individual programs is not always the best analytical approach; doing so can obscure the cumulative consequences of having simultaneous experiences with multiple programs (Sackett and Lareau 2023, this volume, issue 4).
If burdens are so important, why have they historically received relatively little attention? After all, they are hardly an unfamiliar topic. Ask anyone about their interactions with government, and chances are you will get an earful about a seemingly Kafka-esque experience that they or a family member has faced trying to access vitally important social rights such as health care, income support, unemployment, and food assistance, or a fundamental political right, such as voting (Lowrey 2021).
Administrative burdens have the odd combination of being both grindingly familiar to us as individuals, and largely unattended as a matter of policy analysis, design, and practice. They are a widely observed fact of life, but not a widely used conceptual tool to analyze life. One explanation for this failing is the fragmented discourse around them, siloed both across and within academic disciplines and policy areas. We are social scientists who share an interest in policy, but who assess it from different perspectives: as an economist, a political scientist, a sociologist, and a public administration scholar. The interdisciplinary nature of our team reflects both an acknowledgment of the benefit of a diversity of approaches and the need for cross-disciplinary communication. The economist and sociologist uncovering burdens in a social welfare program are often not talking to each other. Similarly, the policy analyst thinking about hassles in education does not share their insights with the Supplemental Nutrition Assistance Program (SNAP) expert. In large part, this is because they do not conceive as burdens as a single analytical concept but are more apt to think of them as specific to a policy domain or area of study. Rather than a general toolbox for reducing burdens, we are left instead with lots of little toolboxes, bereft of enough instruments to comprehensively address the problem.
This article seeks to advance a standard way to talk about burdens, whether in research, applied practice, or in policy debates. There is power in legibility and coherence. Naming something allows us to see it more clearly. It is difficult to accumulate knowledge when researchers use different terms and miss entire bodies of research. Framing something allows us to understand its consequences and to consider how to address it. These basic points are true for different but related domains: social science, politics, and policymaking.
In the following sections, we break down the types of costs in administrative burdens and offer some examples of their effect in different policy domains and the ways they reinforce patterns of inequality. We consider the sources of burdens before looking at potential ways to reduce them. In each section, we highlight recent innovations that have helped build and deepen our understanding of burdens and their consequences, including many from this collection. Table 1 summarizes the key insights we draw from the emerging body of research on administrative burden and provides the roadmap for our discussion in remaining sections.
ADMINISTRATIVE BURDEN AS A KEY PART OF POLICY IMPLEMENTATION
We start by providing a conceptual framework to understand people’s experience of onerous policy administration. Crucially, a focus on costs shifts attention to the individual’s experiences and beyond the perspective of state actors or institutions (Michener, SoRelle, and Thurston 2022). This draws and builds on efforts to articulate the basic idea and consequences of burdens (Bozeman 2000; Currie 2004; Moffitt 1983). Most burdens can be categorized in terms of one of three costs: learning, compliance, and psychological costs (see table 2) (Herd and Moynihan 2018). In practice, such costs may be correlated and overlap, but they are distinct enough to consider separately (Madsen, Mikkelsen, and Moynihan 2022). Doing so allows us to describe, diagnose, and communicate the nature of the burdens at play. We cannot fix problems we do not clearly identify.
LEARNING COSTS
Learning costs are the challenges that people face finding out about a program’s existence and benefits, determining whether they are eligible for the program and what benefits they might receive, as well as understanding how to apply for, retain, and redeem benefits.
You cannot access a program or benefit if you do not know that it exists or you do not know you are eligible for it. The Earned Income Tax Credit (EITC) is an important example. The EITC is the second largest income support policy, only lagging behind Social Security. It is the largest antipoverty program for children and reduces the number of children living in poverty by 25 percent (CBPP 2019). A single parent, with two children, earning just under $48,000 can receive a benefit up to $5,980 (2021 tax law). Approximately 20 percent of those eligible for the benefit do not receive it, however. That some are unaware of the program is part of the problem. One issue is that many eligible people do not file taxes because their incomes are low enough that they are not required to file. Nearly two-thirds of those eligible, but who fail to claim the EITC, do not file a tax return (Census Bureau 2013). A study among childless tax filers eligible for the EITC but who did not apply for it found that simply notifying them and explaining the potential benefit increased take-up by 31 percentage points (Bhargava and Manoli 2015).
Learning costs are likely to be especially high for benefits that are less familiar, such as new programs or rarely used programs. A key barrier to housing assistance during the pandemic was simply not knowing that housing assistance was available (Aiken, Ellen, and Reina 2023, this volume, issue 5). Even knowledge that the program exists is not necessarily enough, though, because individuals then need to learn how to negotiate a new set of administrative processes. A substantial barrier to accessing emergency assistance after Hurricanes Katrina and Rita was not knowing what was needed to access that aid (Raker and Woods 2023, this volume, issue 5). Learning costs can also extend beyond the administrative processes that occur directly within bureaucracies. The process of redeeming benefits for the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC)—the third-largest food assistance program in the United States—presented program participants with distinct challenges as they figured out what they could buy, given the limits of WIC food packages, and where they could buy it, given variability in the selection of WIC-approved foods across retailers (Barnes 2021; Barnes, Halpern-Meekin, and Hoiting 2023, this volume, issue 5).
COMPLIANCE COSTS
Compliance costs involve the time and effort spent on the often tedious task of filling out forms, documenting status, or responding to bureaucratic directives. It includes the time spent waiting on hold, or driving to an administrative office. Jennifer Bouek (2023, this volume, issue 5) highlights an underexplored time cost—the time spent on wait lists trying to access benefits such as childcare—and shows that in some cases it can be years, rendering services inaccessible for many. Clients had to actively maintain their spots on wait lists, with many people getting kicked off because they could not manage that process. Compliance costs also incorporate financial expenses, such as funds needed to order documents, pay a fee to participate in a program, or hire an advocate. Such costs can also be enormously consequential in terms of how compliance affects benefit access.
In a notable example, Tennessee dropped approximately 10 percent of Medicaid-enrolled children from the program in an onerous recertification process. Renewal occurred only by mail. Mail-in forms often do not reach target populations, mostly because poorer families move frequently. Families that failed to complete the forty-nine-page form, returned it late, or made errors, lost their health insurance. Almost 250,000 children lost coverage, the majority because the forms were not returned, incomplete, or late (Arbogast, Chorniy, and Currie 2022; see also Heinrich et al. 2021). Tennessee was one of many states that introduced new burdens to their Medicaid programs between 2016 and 2019. The result was to reverse a long-term trend of increasing health insurance coverage for children. Such burdens reduced public insurance coverage of children by an estimated 5.4 percent within a year of adoption (Arbogast, Chorniy, and Currie 2022).
Although it is challenging to access public programs such as SNAP or Medicaid (Barnes 2021; Homonoff and Somerville 2021; Michener 2018, 2019; Unrath 2021), it is often just as difficult use them. Recent work by Carolyn Barnes (2021) contributes to the conceptual framework by identifying an additional cost: redemption costs. In short, it is not just the costs of accessing and maintaining access to basic rights and benefits, it can also be the process of actually using benefits. The WIC program, for example, has an incredibly complex set of requirements, that are frequently changing, regarding what foods can be bought with these benefits. Barnes’s work finds that the process is so complicated, some beneficiaries effectively abandon the program.
By contrast, alleviating the burden of compliance costs can improve the experiences of program beneficiaries. When compliance costs lessened as the result of remote appointments in response to the COVID-19 pandemic, WIC participants reported having an easier time accessing and maintaining their benefits (Barnes and Petry 2021).
PSYCHOLOGICAL COSTS
Perhaps the least understood category is psychological costs. These include stress, frustration, anxiety, loss of autonomy, or sense of stigma that arise from trying to access and maintain as well as use benefits or services. The concept of stigma is long standing (Moffitt 1983) and the idea of frustrations from paperwork is both intuitive and observable at a physiological level (Hattke, Hensel, and Kalucza 2020). Research on stigma is the most well developed among the psychological costs, with evidence pointing to significant variance in its impact. Research on food stamps has shown that participants can feel stigmatized, such as being shamed at the grocery store (Rogers-Dillon 1995). Research on the EITC, however, finds little evidence of associated stigma, perhaps reflecting its link to paid employment and its framing as a tax benefit rather than a welfare program (Bhargava and Manoli 2015; Halpern-Meekin et al. 2015).
Other kinds of psychological costs go beyond stigma. The stakes are high for those failing to navigate these burdens. Worrying that you cannot feed your children or access critical medical care is obviously stressful (Lowrey 2021). Evidence is robust that high levels of cumulative stress, including that associated with adversity, can be damaging (Herd and Moynihan 2020). Interactions with social welfare systems can also be humiliating or degrading. More nuance is needed to understand other psychological costs. In many respects, qualitative studies that describe people’s experiences are better attuned to the nuances of psychological costs and how they shape people’s interpretation of the state (see, for example, DeLuca, Katz, and Oppenheimer 2023, this volume, issue 5; Barnes, Halpern-Meekin, and Hoiting 2023, this volume, issue 5; Michener 2018).
Our understanding of psychological costs is sharpened by contributions in this double issue. First, the nature of psychological costs in coercive or rights-depriving venues is likely to extend beyond stress, frustration to incorporate fear, or the experience of deprivation (Brown 2023, this volume, issue 5; Edwards et al. 2023, this volume, issue 5; Moynihan, Herd, and Gerinza 2022). Second, although we focus primarily on psychological costs for the participants and beneficiaries, imposing or negotiating burdens can also be experienced as stressful or demotivating for public servants at the frontlines, or for third-party actors trying to help. For example, Lilly Yu (2023, this volume, issue 4) reports that immigration lawyers experienced the deliberate imposition of immigration burdens by the Trump administration to be demotivating, making it more difficult to attract and retain employees. A senior attorney said, “All my attorneys tell me ‘I’m feeling overwhelmed, I’m feeling stressed.’ We have, network-wide, lost a lot of attorneys because of exhaustion and stress. One came back after a year absence, part-time. But, um, a lot of time we—I remember one young attorney was like, ‘I started doing this under Obama and I just can’t do this anymore.’ And she just quit, you know, she just can’t take it. So we see that a lot.”
THE EFFECTS OF ADMINISTRATIVE BURDENS
In this section we summarize the ways in which administrative burdens affect access, outcomes, and inequality. We describe the evidence along five main themes: access, social control, inequality, incidence, and the accumulation of burdens.
BURDENS HAVE LARGE EFFECTS ON ACCESS TO RIGHTS AND PUBLIC SERVICES
The ubiquitous and mundane nature of administrative burdens may lead us to systematically underestimate the scale of their effects on access to basic rights and services. To illustrate this point, we offer two examples. First, burdens have been central in restricting access to the ballot, the most fundamental political right in a democracy (Herd and Moynihan 2018). As the franchise was nominally extended to Black voters, burdens became the way governments constrained their political power. The record of voter registration in Louisiana is an example. Before the 1898 Louisiana constitution, Black and White voters had similar registration rates. Unable to formally bar newly empowered Black voters, the White power structures relied on a mix of literacy tests, poll taxes, and property requirements, even as they exempted more Whites from such clauses using grandfather clauses and discretionary application of the rules by election officials. When one method of exclusion was removed, another took its place, until the Voting Rights Act of 1965 (Keele, Cubbison, and White 2021). Even with federal guarantees, the effects of these practices were so enduring that it was not until the start of the twenty-first century that Black Louisianans regained the voter registration parity they had held a century earlier (see figure 1).
A second example is provided by Parolin, Cross, and O’ Brien (2023, this volume, issue 5), who estimate, using Current Population Survey data from 1990 to 2019, how much administrative burdens limit access to cash and near-cash transfers that people are eligible for after a family member loses a job. Their approach represents a thought experiment about what the reach of the social safety net might look like if burdens were low enough that all families eligible to receive a given benefit should do so. The results point to a significant and large gap between earnings replacement rates—the share of lost earnings made up for with social safety net benefits—under current practice and the counterfactual. Figure 2 presents the actual and counterfactual replacement rates for all—Whites, Blacks, and Hispanics. Although replacement rates are declining across this period for all subgroups, the scale of the administrative burden effect remains consistently large, indicating that families experiencing job loss could receive replacement rates up to 10 percentage points higher in a condition of zero burdens relative to current practice. If the scale of such effects is surprising, it may reflect our tendency to underestimate the effects of burdens.
These burdens can be so large that reducing them can have just as large an impact as expanding eligibility for a program. Emily Rauscher and Ailish Burns (2023, this volume, issue 4) demonstrate how cumulative reductions in burdens in Medicaid had as large of an impact on infant health as the Affordable Care Act (ACA) Medicaid expansion did.
BURDENS FACILITATE SOCIAL CONTROL
The primary focus of administrative burden research thus far has been on limiting access in rights-granting venues, such as political rights or access to benefits. However, as contributors to this double issue point out, burdens also matter as a method of social control in rights-depriving venues, that is, citizen-state interactions in which people have involuntary contact with coercive state institutions (Edwards et al. 2023, this volume, issue 5; see also Brown 2023, this volume, issue 5). An obvious example is the criminal legal system. For example, Black Americans express a stronger sense of fear about interacting with the police (Pickett, Graham, and Cullen 2021), even as the police represent the more direct and present face of government for many Blacks (Soss and Weaver 2017). Frank Edwards and his colleagues (2023, this volume, issue 5) point to child protective services as another such venue, where administrative burdens are a key component of contemporary punitive and racialized poverty governance. Another example is immigration, where administrative burdens have been weaponized with the goal of creating fear among both undocumented immigrants, and those seeking legal refugee status (Moynihan, Herd, and Gerinza 2022; Yu 2023, this volume, issue 4).
The role of burdens in rights-restricting domains is less well understood. It is reasonable to assume that given the stakes involved—loss of liberty, familial rights, legal status—the consequences of burdens in such venues may be greater, reflected in more intense psychological costs such as fear or despair. Burdens can both lock people out of desired benefits and lock them into unwanted long-term involvement with punitive state organizations.
One way burdens facilitate social control function is via entry into and exit from different administrative categories. In some cases, such as establishing disability, burden limits access to desired categories that provide additional income supports, help, or relaxation of standard administrative demands such as work requirements (Sommers et al. 2020). In others, however, burdens make it difficult to escape from undesired categories, such as deficient parent, itself a psychologically degrading identity the parent has to both accept and document that they have graduated from in order to exit the system (Edwards et al. 2023, this volume, issue 4).
BURDENS REINFORCE INEQUALITY
Administrative burdens tend to exacerbate patterns of inequality. This is a consistent theme across the articles in this double issue. Parolin, Cross, and O’Brien (2023, this volume, issue 5) find that administrative burdens contribute to higher poverty rates and lower income rates for Black and Hispanic families experiencing job loss than for White families. These racialized burdens are also evident in higher education (Ray, Herd, and Moynihan 2023). Among those with high student loan debt, those living in lower-income and higher-percentage Black neighborhoods are less likely to take advantage of administratively complex but highly beneficial repayment plans provided by the federal government (Goldstein et al. 2023, this volume, issue 4). Applications for disaster support after Hurricane Katrina result in higher rate of denials for documentation reasons when they came from zip codes characterized by higher poverty rates and more people of color (Raker and Woods 2023, this volume, issue 5). Burdens that reduce public health insurance for children have a larger impact on Hispanic families, those with a noncitizen parent, and those without a college-educated parent (Arbogast, Chorniy, and Currie 2022). If burdens hurt marginalized groups, this also implies that efforts to reduce burdens reduce inequalities. Stephanie Pierce and Stephanie Moulton (2023, this volume, issue 5) point out how relaxing documentation requirements tended to help avoid foreclosure for those with the most difficult-to-document situations, who were more likely to be Black and women. Reductions of Medicaid burdens also tended to benefit marginalized groups (Rauscher and Burns 2023, this volume, issue 4).
Burdens reinforce inequality via three main mechanisms. First, the reach and effectiveness of programs intended to provide a measure of equality is undermined by administrative burdens. To the degree that programs designed to provide baseline support for people living in poverty are inaccessible, they are unable to achieve their policy goals.
Second, some groups are more targeted by burdens than others, even within the same benefit system. Our ideas about who is—and is not—deserving of government benefits are reflected in how we deliver those benefits and services. Our tax system, which has become an important mechanism for distribution and redistribution, benefits the White and well-off, by making programs such as subsidized employer provided health insurance and home ownership simple and accessible. Benefits for the working poor, such as the EITC, have higher layers of burdens (and higher rates of auditing). Even in the near universal, albeit temporary, Child Tax Credit (CTC) expansion (available in 2021), burdens were largely faced by the poorest beneficiaries. In particular, those who had filed a tax return for 2019 or 2020 received the monthly CTC payments automatically and, for many, via direct deposit; those who had not filed taxes, that is, those with the lowest income levels and thus not required to file, had to apply for the benefit.
Groups with less access to political power are more vulnerable to the imposition of hassles. It is not just that women, poor, Black, and disabled Americans differentially bear the brunt of them, burdens are in fact rooted in racism, sexism, and ableism (Ray, Herd, and Moynihan 2023). The relationship between political power, perceived deservingness and the targeting of burdens reflect the point that burdens do not just reinforce patterns of inequality; they are also the product of inequality. Those deemed undeserving are also exposed to burdens in rights-depriving venues like child welfare bureaucracies (Brown 2023, this volume, issue 5; Edwards et al. 2023, this volume, issue 5). The burdens that tribal governments and their community manage in child welfare cases reflects the gap between the promise of tribal sovereignty and actual political power (Brown 2023, this volume, issue 5).
The third way that burdens reinforce inequality is unequal distribution of resources needed to overcome them. We consider this point in greater detail in the next section.
PEOPLE WITH FEWER RESOURCES ARE LESS EQUIPPED TO MANAGE BURDENS
Our perspective on the relationship between burdens and inequality challenges existing theories that have justified burdens as a way to ration public services programs. Referred to as ordeal mechanisms (Nichols and Zeckhauser 1982; Zeckhauser 2021), this model follows some straightforward assumptions: not everyone who qualifies for public benefits actually needs them; people value the time and effort required to manage burdens, and thus will rationally opt out of these hassles if they are high relative to the potential benefits. The ordeals perspective offers policymakers a reassurance that burdens fulfill a useful social function by optimally targeting scarce resources to those most in need.
A framework that presents burdens as an ordeal mechanism creates some blind spots. One is that targeting may be counterproductive. For example, hassles reduce take-up of health insurance and are most likely to discourage the type of people—younger and healthier—the insurance pool needs to balance the numbers of older and sicker participants (McIntyre, Shepard, and Wagner 2021). In such cases, automatic enrollment not only increases participation for some groups but also generates a collective good. The comparison holds in the quite different policy domain of voting. Evidence is ample that hassles such as voter registration requirements differentially exclude more marginal voters: younger, poorer, minority, or immigrants (Grumbach and Hill 2021; Rigby and Springer 2011; Braconnier, Dormagen, and Pons 2017; Michener 2016). Indeed, the general gap between rich and poor when it comes to voting is a registration gap, not a voting gap: both poor and rich people who are registered vote at the same level (Herd and Moynihan 2018).
When hassles are not well targeted, they often end up affecting those who most need help. Why might this be the case? If we accept that some burdens are not simply tedious, but also complex and cognitively demanding, those with more resources become more likely to succeed in overcoming them. Here resources can include education, administrative literacy, money, social networks, flexible work schedules, reliable internet and phone connections, cognitive skills, health, and time. As a result, an administrative burden Catch-22 emerges: those needing the most help are less well positioned to overcome the barriers on which that help is conditioned (Christensen et al. 2020). Rather than targeting services only to those who needed them the most, burdens can end up becoming unnavigable for those without the resources to deal with them. Pierce and Moulton (2023, this volume, issue 5) find as much when it comes to giving mortgage relief to homeowners in trouble: reductions in burdens made targeting more efficient by helping those who were needy but struggled to satisfy documentation requirements.
The degree to which burdens result in better targeting or exclude the neediest is an empirical question. Evidence from behavioral science seems more consistent with the Catch-22 perspective: individuals experiencing time or financial scarcity will be less apt to manage the hassles involved in administrative processes (Mullainathan and Shafir 2013). Empirical studies beyond those in this double issue also raise concerns about the exclusionary effects of burdens. Families facing poverty deal with more complex life challenges, more administrative demands across multiple programs, with fewer tangible resources or sources of help (Heinrich et al. 2021). A study of Social Security field office closures found that they tended to reduce access to benefits for those with lower income, less education, and more severe disabilities (Deshpande and Li 2019). Amy Finkelstein and Matthew Notowidigdo (2019) point out that frictions in SNAP enrollment tend to be more consequential for lower-income and sicker individuals, who need the benefit more. Mental health problems, such as attention-deficit disorders, are associated with higher reported burden and lower take-up (Bell et al. 2022).
These burdens have implications beyond program access. Negative interactions with government in the form of burden reduce trust in government (Ali and Altaf 2021). For example, the learning costs that result from moving across state lines—because Medicaid eligibility and the rules around documenting that eligibility vary across states and even counties—leads to frustration and distrust, with profound consequences for the health of our democracy (Michener 2018). The deliberate use of ordeals conveys a formal recognition of inequality. In the eyes of the state, some people’s time is of less value than that of others.
THE EFFECTS OF BURDENS ACCUMULATE OVER TIME
Although the categories of learning, compliance, and psychological costs help unpack the frictions that people face in particular encounters or programs, this diagnostic process should not obscure the broader picture of burdens as a cumulative experience. In short, some people are systematically more likely to be exposed to burdens. It is important to remain attentive to the accumulation of burdensome experiences across multiple programs or domains, both rights-granting, such as applying for benefits, and rights-depriving, such as when social services take a child away from a parent (Edwards et al. 2023, this volume, issue 5; Sackett and Lareau 2023, this volume, issue 4). In the former, someone is a claimant, whereas in the latter someone is the subject of the state, but across these domains the costs accumulate. Families seeking help from the safety net are negotiating with multiple institutions. Poorer families seeking help encounter burdens in WIC, childcare supports, public housing, SNAP, and Medicaid. Blair Sackett and Annette Lareau (2023, this volume, issue 4) illustrate this point in their study of refugees negotiating multiple venues, which are both vital and unfamiliar. This patchwork system of in-kind programs, rooted in patriarchal and punishing practices directed at women, people living in poverty, and Black individuals, leads to a much larger set of costs than would be present with simple cash assistance.
Even after someone has overcome enrollment burdens to access a benefit, they continue to face redemption costs in using those benefits (Barnes 2021), especially for voucher-based benefits such as housing supports (DeLuca, Katz, and Oppenheimer 2023, this volume, issue 5) and WIC (Barnes 2021). On the one hand, learning costs should mechanically decline with familiarity with an institution or program. A number of articles in this double issue point to the way that the novelty of a new context raises learning costs. Claudia Aiken, Ingrid Ellen, and Vincent Reina (2023, this volume, issue 5) examine how a new rental assistance program struggled to succeed. Sackett and Lareau (2023, this volume, issue 4) examine the challenge facing immigrants negotiating unfamiliar spaces. Yu (2023, this volume, issue 4) documents how immigration lawyers struggled to keep up with a rush of policy changes. However, familiarity may also breed contempt. Negative experiences in using benefits may cause people to exit programs (Barnes, Halpern-Meekin, and Hoiting 2023, this volume, issue 5). The exit, or “churn” of eligible applicants at points of renewal is another juncture where those who have shown a capacity to overcome a set of burdens in the past are no longer willing or able to do so.
For many, the experience of the state is the experience of burdens. People who rely on social welfare programs, from food assistance to the EITC, simply spend more of their lives navigating complicated bureaucracies to meet their basic needs than do those with more resources (Land 2018). Thus the accumulation of burdens will reinforce inequality to the degree that people systematically experience the same sort of burdens in the administrative venues they are assigned. Marginalized groups may look at the history of their experience with the state, and view their interactions through that lens. Using the American Time Use Survey, Stephen Holt and Katie Vinopal (2023) find that low-income people are 3 percentage points more likely to spend part of their day waiting for services, and the duration of their waits are, on average, twelve minutes longer. These differences are not based only on income: high-income Black people experience the same wait times as low-income groups.
The effects of the accumulation of burdens should be considered, but burdens can be exceptionally consequential in crucial state interactions. As noted, burdens are especially salient for high-stakes rights-restricting venues, such as child protective services, or for crucial rights, such as voting, or at crucial moments, such as trying to recover from a disaster (such as Raker and Woods 2023, this volume, issue 5). The Joe Biden administration has identified surviving a disaster as one of a series of “moments that matter” when public services should be designed to be accessible. Other such moments include the birth of a child, retirement, transitioning from military to civilian life, financial shocks, and children with disabilities transitioning to adulthood.2 A focus on identifying and providing support to such moments of vulnerability is a useful way for governments to think about how to prioritize burden-reduction efforts.
SOURCES OF ADMINISTRATIVE BURDENS
In this section, we summarize the evidence on the sources of administrative burdens including the role of policy design, bureaucratic dysfunction, federalism, and the private provision of social welfare benefits.
BURDENS ARE POLICYMAKING BY OTHER MEANS
If administrative burdens are such a big problem, why tolerate them? Why is it not a core function of governments to relentlessly identify and reduce burdens to the greatest extent possible? These deceptively simple questions also ask why burdens emerge and persist. No single reason explains the phenomenon.
At the broadest level, burdens persist either because they serve an end for some political actors or because the problems they create are kept hidden. Burdens, of course, can reflect normal bureaucratic dysfunction, when administrative actors fail to understand, are indifferent to, or feel they cannot change how state actions and inactions affect how the public experiences the state. Despite individual behavioral reasons that burdens are consequential, such as individuals being present-biased and unwilling to deal with immediate hassles (Frederick, Loewenstein, and O’Donoghue 2002), we focus on issues under the control of the state and consider individual differences only to the extent that they worsen patterns of inequality. Whether burdens are deliberate or accidental, they are largely a function of state choices, including choices not made, to attend to them.
What sort of functional purposes can burdens serve? Processes that give rise to burdens are sometimes necessary for basic administrative tasks, such as distinguishing between those who are eligible and ineligible for a program. Burdens help (or are claimed to help) achieve values that policymakers say they care about, such as reducing fraud, targeting scarce resources to the neediest, or other policy goals, such as encouraging labor-force participation. Burdens can also help achieve outcomes policymakers desire but do not explicitly acknowledge, such as restricting access to the ballot or curtailing abortion services. They also undermine the effectiveness of programs for those who are opposed to those programs or the constituency they serve. Such opposition can coincide with, or be driven by, indifference, paternalism or antagonism toward marginalized groups. For example, Victor Ray, Pamela Herd, and Donald Moynihan (2023, 139) argue that burdens that make it hard to vote emerged “when more explicit forms of racial bias in policies or administrative practices become illegal, politically untenable or culturally unacceptable.”
In other words, burdens often function as a form of policymaking by other means. The utility of burdens and their consequences for inequality are tied to their opacity. The more complex and submerged are the details of program implementation, the less they are subject to close inspection by media or other policymakers.
POLICY DESIGN MATTERS, BUT UNIVERSALISM IS NOT A SUFFICIENT FIX
Existing theories and research, particularly by social scientists throughout the 1980s and 1990s, largely considered burdens through the lens of policy design or whether benefits should be targeted or universal.
Sociologists and political scientists thought about burdens in the context of universalism versus targeting (Esping-Andersen 1993). In short, universal policies are assumed to be simple, accessible, and popular, whereas means-tested policies are laden with burdens, inaccessible, and unpopular (Skocpol 1991). This work also documents the ways poor and racialized minority women are far more likely to be segregated into targeted and burdensome programs, making this divide a central way that social welfare policies reinforce race and gender inequality (Mink 1998). Programs created by the 1935 Social Security Act make the point. Social insurance programs such as Social Security became generous and accessible, especially for White men, and unemployed White women, whereas means-tested programs such as Aid to Dependent Children, to which Black women had relatively more access, became miserly, inaccessible, and ultimately written out of the Social Security Act altogether (Mink 1998; Quadagno 1994). The same contrast is also drawn between Medicare and Medicaid, both of which were created in 1965 as amendments to the Social Security Act. In part, targeting centered narratives around deservingness and focusing means-tested policies often on populations deemed “undeserving” (Schneider and Ingram 1993). Indeed, survey experiments with politicians show that they become more supportive of imposing burdens when deservingness framings are invoked (Baekgaard, Moynihan, and Thomsen 2021).
However, even though targeting versus universalism continues to hold predictive power regarding how policy design influences burdens and their impacts, a closer look complicates this narrative. In terms of how burdens impact inequality, it is not only that women and Black people end up more subject to hassles because they are more likely to be segregated in targeted programs: even within targeted programs they face more burdens. For example, states with higher proportions of Black workers are more likely to have tougher welfare-sanctioning policies (Hahn et al. 2017) and less spending on cash assistance out of the TANF block grant (Hardy, Samudra, and Davis 2019). It is also the case that, within social insurance programs, Black Americans are more concentrated in burdensome parts of those programs. For example, Black Americans are disproportionately likely to receive Social Disability Insurance rather than Social Security Survivor benefits (Harrington Meyer et al. 2006). Whereas it takes just a few minutes to apply for survivor benefits, disability benefits are laden with burdens. In 2020, it took an average of five to six months for an initial application to be processed (SSA 2021). If an individual had to appeal a denial, they would wait another four months on average for it to be reconsidered (SSA 2021). If it then goes to an administrative law judge, the wait becomes even longer. Given that one-third of disability recipients had to appeal to receive their benefit, the average costs are extraordinarily high (CBPP 2021).
Policy trends in the aftermath of 1990s welfare reform have seen burdens generally cut in targeted programs (Greenstein 2022). Mean-tested policies that have grown in size—most obviously EITC, SNAP, and the CTC—also have less burdens and higher take-up than in the past and the programs they now dwarf, such as TANF. At the same time, burdens have become more pronounced in some universal programs. For example, Medicare has become far more complicated to navigate over the last thirty years with the introduction of private health insurance plans for Medicare beneficiaries (Herd and Moynihan 2018; Herd 2021). When beneficiaries enroll in Medicare, they now face a dizzying array of health insurance options. Depending on what they choose, they may actually end up with three separate plans: Medicare Part B, a private Medigap plan, and a private prescription drug plan. The evidence is that most people pick poorly and expose themselves to potentially large financial and health costs (Herd 2021). Further, to ensure they have the best plan to meet their needs, given that their health changes and the plans change, beneficiaries are required to make choices every year. In contrast, Medicaid has expanded and in some ways become less burdensome over time (Herd and Moynihan 2018), given the removal of asset tests, reductions in paperwork, longer periods between recertification, and a host of other changes driven by the ACA and state discretion (Rauscher and Burns 2023, this volume, issue 4).
BUREAUCRACIES ARE NOT NATURALLY INCLINED TO DETECT AND MINIMIZE BURDEN
Bureaucratic has become a synonym for delay, pointless paperwork, and frustration. Such critiques are often lazy stereotypes that do not recognize the complexity of public services. After all, private organizations also create burdens. Think of how easy private companies make it to sign up for services such as gym membership, a magazine subscription, or a cable package relative to the challenge of canceling those services. Such burdens are often driven by profit incentives. For public bureaucracies that do not have a profit incentive, the dynamics are somewhat different, but still compatible with the creation of burdens (Peeters 2020). The root of these dynamics can be linked not only to politics and organizational culture, but also to the capacity public organizations have to effectively deliver services and benefits.
Public organizations are responsive to political messaging and directives about maximizing political values that result in burdens (Moynihan, Herd, and Harvey 2015). Even absent political pressure, though, they too often lack direct mechanisms of feedback that measure burdens, or incentives to reduce them. Robert Merton (1963) long ago observed that bureaucracies tend toward goal displacement: focusing on rule creation and rule following at the expense of their core purpose. Such tendencies are reinforced by incentives such as blame avoidance and negativity bias, where bureaucrats face greater punishment for failure than benefits for positive outcomes (James et al. 2020). In other words, bureaucrats may be sheltered from knowledge about the negative externalities that rules that generate new burdens create but be acutely conscious that the absence of such rules makes them vulnerable to criticism. For example, Aiken, Ellen, and Reina (2023, this volume, issue 5) point to administrators mistakenly believing their programs were less burdensome than they truly were, while being closely attentive to the risk of fraud and the potential for federal audits in ways that made burdens more likely to emerge.
Institutional errors—which can be a function of a mistake by a state actor or the client—can have calamitous effects (see Bouek 2023, this volume, issue 5; Raker and Woods 2023, this volume, issue 5; Goldstein et al. 2023, this volume, issue 4; Sackett and Lareau 2023, this volume, issue 4). For example, such mistakes for refugees can cascade, undermining their efforts to establish some stability (Sackett and Lareau 2023, this volume, issue 4). Groups lacking political power have less capacity to draw attention to administrative problems, or win resources to provide solutions (see also Moynihan, Herd, and Gerinza 2022).
Policymakers are also more comfortable with burdens if they are not on the receiving end. Bureaucrats and elected officials in particular tend to be more educated and wealthier than the public generally, making their personal experience less representative of those they serve. Relatively few policymakers have had to worry about negotiating work requirements for SNAP, for example. This lack of experience matters to how policymakers see burdens. Policymakers tend to be more opposed to burdens in means-tested programs when they have personally depended on similar programs in the past (Baekgaard, Moynihan, and Thomsen 2021).
Studies of citizen-state interactions show how street-level bureaucrats can alter each of these types of costs. They can choose to engage in active outreach or withhold information in ways that adjust learning costs. They can impose excessive demands for documentation and paperwork, relax requirements, or even help individuals to overcome compliance costs. They may offer a welcoming setting, conveying to individuals that they are accessing rights rather than supplicants to the state. Some of these choices may reflect personal biases against marginalized groups or ideological beliefs (Bell et al. 2021). Audit studies offer evidence of such biases among street-level bureaucrats, though not consistently and seemingly less systematically than among other political elites or private actors (for a summary, see Olsen, Kyhse-Andersen, and Moynihan 2022). It would be a mistake, however, to assume such micro-level interactions as purely a reflection of individual biases, given that they may also reflect embedded meso-level administrative practices, which street-level actors are only vaguely aware of, or are resigned to accepting (Ray, Herd, and Moynihan 2023; Compton et al. 2022).
Administrative capacity matters to the ability of bureaucratic actors both to manage burdens themselves and to shift burdens from the individual and onto the state (Herd et al. 2013; Heinrich 2016). For example, Aiken, Ellen, and Reina (2023, this volume, issue 5) detail that a lack of infrastructure and data capacity was a key explanation for the burdensome processes that people encountered when trying to seek housing assistance during the pandemic. Examining a foreclosure program in Ohio, Pierce and Moulton (2023, this volume, issue 5) find that an effort to reduce compliance costs by relaxing documentation requirements helped applicants, but also created congestion because the state lacked the administrative capacity to manage the increased demand that followed. Lack of capacity among state actors may be compensated for by nonstate actors, but they also face capacity constraints. Yu (2023, this volume, issue 4) details how third-party brokers, in the case of immigration attorneys, were overwhelmed during the Trump presidency as the Department of Homeland Security dramatically increased burdens across the immigration process. These immigration lawyers simply did not have the resources to adequately help people navigate these burdens, resulting in stratified access to representation.
THE FEDERATED AND FRAGMENTED NATURE OF U.S. POLICY IMPLEMENTATION ENHANCES BURDENS
A primary lesson emerging from new scholarship on administrative burden is the identification of federalism as a key source of burden. Part of the issue is the basic reality that more players means more veto points and more opportunities to add burdens. Federalism also means locating service delivery in venues that are subject to less direct attention from the public. For example, federal policies that impose work requirements or new hassles in legislation or executive order are more apt to draw attention and pushback than equivalent changes at the state level.
Programs that involve federal, state, and local control tend to be more complicated and burdensome, and consequently more heavily reinforce race, gender, and class inequality, than those solely controlled by the federal government (Michener 2018). In short, federalism plays a large role in shaping the level of burden in U.S. social welfare programs (Michener 2018; Mettler 2011). For example, although both Social Security and unemployment insurance evolved out of the 1935 Social Security Act and are social insurance programs, control of UI is shared between federal and state governments. The Social Security retirement program is the least burdensome U.S. social program, especially when taking its scope and impact into account. Take-up is nearly 100 percent and users do not need to keep track of, or document, their lifetime earnings. Benefits can be claimed in a matter of minutes, either online or via one of thousands of field offices around the country.
Unemployment insurance is a very different story. Its dysfunction during the pandemic was not just the result of an unusual surge in applications. Even in “normal” times, only about three-quarters of those eligible for unemployment insurance actually receive benefits, with many jobless people deemed ineligible. The fraction of unemployed people receiving benefits ranges from about 10 percent in North Carolina to 57 percent in New Jersey (U.S. Department of Labor 2022). Notably, differences across racial groups are large. Using data through 2015, Elira Kuka and Bryan Stuart (2022) find that only 42 percent of eligible Black individuals receive UI relative to 55 percent of eligible White individuals and 20 percent of the racial gap is accounted for by Black individual’s greater residence in the South. More generally, states with higher proportions of Black workers also tend to have less generous unemployment insurance benefits, with learning costs contributing to low take-up of benefits across all racial groups (Gould-Werth and Shaefer 2012). Benefit eligibility varies across states in terms of job classifications, wage requirements, and reasons for leaving a job, meaning that many unemployed workers are unaware of their eligibility status (Wandner and Stettner 2000).
Challenges accessing UI during the pandemic reflect prior policy choices made, especially after the Great Recession, to narrow eligibility and broaden burdens (Badger and Parlapiano 2020). The cumulative effect was to reduce the overall fraction of unemployed workers receiving benefits from around 31 percent between 2004 and 2007 to 23 percent between 2012 and 2016 (Vroman 2018). Compliance costs played a key role in this decrease. States such as Florida (Fineout and Caputo 2020) and North Carolina (Lee 2020) confounded the application process with burdens to reduce unemployment insurance spending. Across states, new programs to facilitate employment, such as job counseling and required documentation of job-seeking activities, led to people losing unemployment insurance for failing to meet administrative requirements, rather than to increased employment (Vroman 2018).
Many burdens are structurally induced as subnational governments navigate requirements from the federal level. For example, concern about federal audits triggered more burdens in rental assistance programs (Aiken, Ellen and Reina 2023, this volume, issue 5). Tribal governments with limited resources struggle with the administrative requirements placed on them by state governments in the domain of child welfare, and in some cases states have deliberately limited tribal capacity (Brown 2023, this volume, issue 5).
Federalism is a contributing factor to another source of burdens, which we label here as administrative fragmentation: the multiplicity of administrative actors that an individual must interact with to complete a task. It seems axiomatic that administrative costs increase when a person has to negotiate with multiple organizations. Most obviously, learning costs increase because someone must be aware of more than one relevant organization and understand how to engage with multiple sets of rules. Compliance costs also increase as the person commutes between organizations, and has to provide documentation multiple times to different actors. The frustrations of being shuffled back and forth also mount.
Fragmentation can occur both within and across policy domains. The U.S. health-care system is a very fragmented domain, with some exceptions, such as the Veterans Health Administration, which operates more like a national health system for its users. Adam Goldstein and his colleagues (2023, this volume, issue 4) point to the deleterious effects of fragmentation in student loan programs. Income-driven repayment programs benefit most borrowers, but borrowers must master more than one such program with varying provisions and eligibility. The Department of Education relies on private loan service providers, and when loans are transferred from one servicer to another it can be calamitous for borrowers when basic information is not transferred. Borrowers often have to play the role of their own administrative representative, collecting administrative data, or coordinating organizational relationships between, for example, their bank and lender. Within the broader safety net, different programs may have different rules, definitions, and measurement about things like dependents or assets that make little sense to their users.
Federalism contributes to administrative fragmentation, but it is not the only source. As discussed in the next section, a preference for marketized provision of public services, and a reliance on supportive nonprofits begets fragmentation. A person walking between organizations can get lost. Sackett and Lareau (2023, this volume, issue 4) illustrate how refugees often find themselves tied up in institutional knots in part because they cannot resolve competing or contradictory demands from the multiple organizations they must negotiate with. In other words, it is easier for people to fall through the cracks when the cracks are gaps between organizations.
THIRD PARTIES CAN BUFFER OR AMPLIFY BURDENS
Another critical emerging area of administrative burden research, with important contributions from the articles in this double issue, involves the role of third parties. Private, nongovernmental third-party actors can play multiple roles when it comes to burdens that can reduce or amplify the effects of burdens and associated inequalities: as service providers, as advocates that lobby to increase or reduce burdens, by providing help to users manage burdens, or as the subject of burdens themselves. In discussing immigration, Yu (2023, this volume, issue 4) shows how immigration lawyers can simultaneously help reduce burdens for clients, be the subject of burdens imposed by the state, and reinforce patterns of disparate access by charging a fee for their help. Hilary Hoynes, Nicole Maestas, and Alexander Strand (2022) find that the use of legal representatives significantly speeds up the access to Social Security disability benefits without changing who ultimately receives the income support program. Aiken, Ellen, and Reina (2023, this volume, issue 5), document how burdens increase when emergency rental assistance programs require the participation of both the landlord and tenant. This is what is called a “double take-up challenge.”
Third-party actors are only becoming more important over time. Since the 1990s, the growth in the use of private entities—both nonprofits and for-profits—in delivering access to social welfare benefits has clear implications for burdens. The growth of the private sector in delivering social welfare benefits provides another potential source of burdens and subsequent inequalities for some of the same reasons as federalism: more veto players, more room for discretion to impose burdens. Other challenges are specific to private actors. A desire to marketize public services increases learning and compliance costs. Medicare is a clear example (Herd 2021). As of 2020, nearly 40 percent of beneficiaries were exclusively in private health insurance, via their participation in Medicare Advantage, with estimates that this could reach 47 percent by 2029 (Freed et al. 2022; Terry and Muhlestein 2021).
Figure 3 provides an overview of the maze of options people encounter when they enroll in Medicare, which they must deal with on an annual basis, when they either remain in their existing plans or select new coverage. Given the absence of robust regulations regarding how plan changes are communicated to Medicare beneficiaries, users often act in ways that increase their costs and reduce their benefits. A study of Part D prescription drug plans found that only 5 percent of beneficiaries chose the cheapest plan to meet their needs (Zhou and Zhang 2012). Unsurprisingly, these burdens fall hardest on those who are disabled or have more significant health problems—including cognitive impairment.
The addition of a profit motive will in some cases align with imposing more burdens that are at odds with the best interest of the client. The privatization of poverty-based programs has been a hallmark of welfare reform movement since the 1990s. When the Aid to Families with Dependent Children was removed from the Social Security Act, it was replaced with TANF. TANF was a block grant program that gave states enormous flexibility in how it distributed those funds. States such as Florida put significant funds toward work readiness programs, requirements for people receiving any kind of cash assistance. They also chose to administer those benefits through for-profit organizations, organizations that were more likely to sanction beneficiaries for failing to obey administrative requirements (Soss, Fording, and Schram 2011). This could cost them not only their cash assistance, but also food stamp benefits and Medicaid. Strikingly, studies have found that 30 percent of sanctions were applied erroneously to people who were obeying programmatic rules (Pavetti et al. 2003).
In some cases, the perceived hassles make third parties withdraw essential support that clients need to access benefits. With the example of student loans, Goldstein and his colleagues (2023, this volume, issue 4) show that private organizations might formally have incentives to provide help to individuals and serve their best interests—by directing them to more affordable income-driven repayments—but often fail to do so because loan-service staff do not have the time to deal with the additional hassles of helping borrowers into the program. In a similar manner, the unwillingness of private actors to accept vouchers as forms of payment because of the perceived additional administrative hassle involved can make those vouchers less attractive for users also (Aiken, Ellen, and Reina 2023, this volume, issue 5; Barnes, Halpern-Meekin, and Hoiting 2023, this volume, issue 5; DeLuca, Katz, and Oppenheimer 2023, this volume, issue 5). Immigrants depend on doctors and police to provide essential documentation for visa processes, which the third party experiences as a burden (Moynihan, Herd, and Gerinza 2022). In such cases, the mismatch between the rewards for clients and third parties makes the client dependent on an actor who has little incentive to provide help.
Stefanie DeLuca, Lawrence Katz, and Sarah Oppenheimer (2023, this volume, issue 5) illustrate how third parties can help individuals negotiate the fragmented domain of federal housing assistance. The Creating Moves to Opportunity program was able to significantly increase the success of families using their vouchers to move to high-opportunity neighborhoods. The intensity of the help offered by housing navigators, and the multiple points where help is needed, illustrates the necessity of such support and how onerous the process looks without it.
TOWARD SIMPLE, ACCESSIBLE, AND RESPECTFUL GOVERNMENT
To reduce administrative burdens requires building alternative frameworks, a new toolbox for change and revised institutional norms. What are the normative underpinnings of such an ethic of change? We assume that the provision of public services should be designed to be simple, accessible and respectful. The introduction of burdens that impose costs on the public should be recognized and imposed only when it demonstrates a benefit that exceeds those costs. Burdens should not be used for unsanctioned goals that policymakers are unwilling to explicitly defend, such as undermining the reach of a program, or to target some groups more than others, or to use ordeals to filter out eligible recipients.
Our normative claims may, on their surface, not appear terribly controversial. They are, however, routinely violated in practice. Embedding these norms in government and making them part of practice will require new approaches to measuring citizen-state interactions. This is no easy thing. The articles in this double issue illustrate this point by showing the wide variety of burdens that could be measured, and how they may matter in unexpected ways. As governments seeks to better measure administrative burdens, they will face such practical problems (Executive Office of the President 2022). It is better to at least approximate such measurements, even if imperfect, than ignore them completely.
ADMINISTRATIVE DATA AND TECHNOLOGY CAN HELP, WITH CAVEATS
The pandemic generated a need for massive social support and laid bare the importance of administrative burdens in either facilitating or blocking access to those supports. The collapse of UI benefits pointed to the fragility of systems seemingly designed to be inaccessible. The slow response and relative failure of targeted supports such as rental relief or student lunch supports underlined the challenges of designing complex new systems with many veto players. Other aspects of the state response were more effective. Expansions to SNAP reached needy families quickly and automatically (Bitler, Hoynes, and Schanzenbach 2020), and the waiving of in-person interviews (a documented administrative burden [see Homonoff and Somerville 2021]) led to fewer program terminations (Barnes and Petry 2021). Stimulus payments based on tax data quickly reached those eligible if they were recent tax filers. The expanded CTC saw a dramatic and rapid decline in poverty (Creamer et al. 2022). If we think of the pandemic as an experiment, the lesson seems to be that the use of the tax system can offer a way to significantly reduce burdens.
The long-term shift toward benefits distributed through the tax system reinforces this impression. The growth, relatively high take-up, and political durability of the EITC seems to be a testament to the post-welfare-reform shift to using the tax system. Spending on the EITC far exceeds what was historically spent (on an annual basis) on traditional welfare. Theories as to its relatively low burden, high take-up, and generous benefits range from its link to employment to its concealment within the tax system, allowing it to slide under the political radar (Howard 1997). Moreover, evidence indicates that the EITC promotes social inclusion and political participation (Halpern-Meekin et al. 2015; Shanks-Booth and Mettler 2019).
The use of the tax system, however, is a double-edged sword. Its concealment undermines its impact by significantly increasing learning costs and submerging the role of the government in providing benefits, thus potentially undermining support for the program (Shanks-Booth and Mettler 2019). For the EITC, private tax preparers help reduce this learning cost, often for a fee (Hoynes 2019). Additionally, the very high rate of audits of EITC recipients creates another ex post burden, reducing the benefits of the program (Kiel 2019).
Additionally, for those with low earnings, and disconnected from the tax system, this offers little help. Nearly two-thirds of eligible EITC beneficiaries not receiving a benefit have not filed a tax return, unaware of the benefits they are eligible to receive (Census Bureau 2013). The pandemic stimulus was sent automatically to those who had recently filed taxes, but that left out an estimated eleven million eligible families who would have to navigate the new Internal Revenue Service process for applying for the stimulus. Similarly, the expanded CTC was distributed automatically to recent filers, and those who do not file tax returns faced additional administrative burdens, making them much less likely to receive the payments. Because being a nonfiler is associated with the deepest poverty, this effectively means that generous poverty-reducing programs systematically exclude the poorest eligible individuals because they operate through the tax system. The press to maintain work requirements in a permanent expansion of the CTC would also effectively exclude many such families.
Distribution of social welfare benefits through the tax system also has implications for inequalities in terms of who bears the brunt of burdens. White and well-off Americans largely receive health insurance subsidies via the tax system with employer-based health insurance where administrative burdens are borne by government or private bureaucracies instead of beneficiaries (Ray, Herd, and Moynihan 2023). Indeed, many Americans receiving subsidized employer-based health insurance are completely unaware they are receiving the benefit, the burdens are so low (Mettler 2011). By contrast, racially marginalized groups are more likely to rely on health insurance subsidies through the Medicaid program or the Affordable Care Act health insurance exchanges, which involve substantially higher burdens (Herd and Moynihan 2018; Michener 2018).
If the tax system offers a lesson, it is that the state can dramatically reduce burdens and deliver support quickly for most people by better using administrative data and information technology that shift the burdens away from the individual and onto the state (Herd et al. 2013). It requires not just new investments in state capacity to make such a shift possible, but also a new way of thinking about citizen-state interactions. Rather than waiting for the individual to complete an onerous process, the state reaches out to help those eligible. This is akin to changing the default in our current model of thinking about access to public benefits.
Technology offers multiple promises. It can be used to target outreach, simplify enrollment processes, and streamline the delivery of services. For example, an intervention with income-driven student loan repayment saw large increases in take-up when the default was changed from requiring individuals to go to a website to complete an application, to sending a prefilled application to beneficiaries. This required a combination of administrative data and technology, and increased take-up from 24 percent to over 60 percent (Mueller and Yannelis 2022). Technology can also be used to reduce psychological costs associated with benefits. Barnes, Halpern-Meekin, and Hoiting (2023, this volume, issue 5), examine the effect of the use of electronic debit cards in WIC, a tool previously adopted in SNAP, as a way to reduce people’s sense of stigma.
The lure of technology should not lead us to overlook that it can also be a method of exclusion (Eubanks 2018). Algorithmic bias can be used in ways that target burdens toward marginalized populations. Reliance on online-only application processes is a clear way to exclude populations with low technological literacy (Sommers et al. 2019). In short, technology is a tool that can be used for good or bad.
NUDGES ARE NOT ENOUGH TO REDUCE BURDENS; SOMETIMES HELP IS NEEDED
In some cases, informational nudges can reduce learning costs. For example, Heinrich Hock and his colleagues (2021) find that informational nudges with tailored messages increased take-up in Social Security Disability Insurance. The tailoring of such messages becomes more effective when it can draw on administrative data. Third parties that directly reach out to populations can also play such a role. Such targeted forms of nudges appear to be more successful than broad-based advertising campaigns.
Redesign of application processes can reduce both learning costs by better conveying information, and compliance costs by reducing documentation demands. For example, Code for America, a civic tech company, has reduced frictions in California SNAP application processes by reducing unnecessary steps as well as developing a mobile friendly platform for applications. In one experiment, it simplified the terminology used to describe self-employed (many who are self-employed do not select into this category even though it is beneficial for them to do so) and provided a self-attestation form to reduce documentation requirements. The result was to increase the claiming of larger benefits associated with self-employment status by about one-third (Moynihan et al. 2022).
Some processes are so complex that nudges are not enough: help is needed. DeLuca and her colleagues (2023, this volume, issue 5) document the role of Navigators in aiding people as they negotiate the housing vouchers, echoing previous findings on the role of health-care navigators or aid in FAFSA applications (Bettinger et al. 2012). Hoynes, Maestas, and Strand (2022) show that using attorneys at the beginning of a Social Security Disability Income application substantially reduces the wait time for benefit receipt. Pierce and Moulton (2023, this volume, issue 5) show that reduction of burdens in a foreclosure prevention program increased take-up, and especially benefited female Black and older applicants, and those facing more complex documentation requirements.
Help is a salient solution when compliance costs are high or for hard-to-reach populations who sit on the margins of society, beyond the reach of standard administrative tools. Simply sending more information in such cases can be of limited value if people cannot act on that information. For example, Elizabeth Linos and her colleagues (2022) find that multiple messages failed to increase take-up for the EITC, suggesting that those not already accessing the benefit needed more direct help. In their experiment to increase SNAP take-up among eligible older adults, Finkelstein and Notowidigdo (2019) find that adding a phone line for an enrollment specialist doubled the increase in SNAP take-up relative to a simple information treatment.
One challenge is that an intervention may help many but still increase inequality within groups. Finkelstein and Notowidigdo’s (2019) SNAP study demonstrated how these inequalities play out: those most likely to take advantage of the information or offers of help tended to be those with relatively higher income and better health than the average potential claimant. This underlines not only that frictions tend to deter the most needy, but also that interventions that seek to reduce burdens may come with their own frictions that advantage better-off eligible nonparticipants.
PRACTICAL BURDEN-REDUCTION POLICY FRAMEWORKS ARE EMERGING
Although technology and administrative data provide new tools for policymakers and third parties to reduce burdens, and individual nudges or other types of interventions add to the general toolbox of solutions, the broader challenge is developing and embedding within institutions an ethic to identify and reduce administrative burdens where possible.
We see glimmers of hope on this front. The Biden administration incorporated the framework of administrative burdens into government-wide guidance to agencies, first under the auspices of an inauguration day executive order on social equity (U.S. Office of Management and Budget 2021) and later via another on trust in government—“Transforming Federal Customer Experience and Service Delivery to Rebuild Trust in Government.” The latter declares that “Agencies must work with the Congress; the private sector and nonprofit organizations; State, local, Tribal, and territorial governments; and other partners to design experiences with the Federal Government that effectively reduce administrative burdens, simplify both public-facing and internal processes to improve efficiency, and empower the Federal workforce to solve problems” (Executive Office of the President 2021).
Although previous administrations have emphasized customer service, the Biden administration has explicitly tied customer experience not just to the experiences of people who interact with government, but also to social equity, promising to “deliver services more equitably and effectively, especially for those who have been historically underserved.”
In practical terms, the executive order institutionalizes a set of routines under which the Office of Management and Budget oversees efforts to improve people’s experience of government, reporting on their progress and setting new goals on an annual basis. Administrative burden initiatives may be championed most warmly by behavioral science nudge units in government (see, for example, UN Innovation Network 2021). This is understandable, and in no small part reflects the work of Cass Sunstein and Richard Thaler, the original Nudge authors, in elevating attention to what they label as “sludge” (Sunstein 2021). It also comes with some risks, as behavioral scientists in such units will naturally be inclined to look for behavioral solutions. A more hopeful possibility is that attention to administrative burdens should broaden the perspective of such units. The Biden executive order is explicit in proposing that addressing burdens means thinking not just about nudges and process redesign but also looking at underlying structural causes of burdens: “efforts to improve customer experience should include systematically identifying and resolving the root causes of customer experience challenges, regardless of whether the source of such challenges is statutory, regulatory, budgetary, technological, or process-based” (Executive Office of the President 2021).
The executive order also triggered new agency guidance when it comes to the implementation of the Paperwork Reduction Act of 1995, which has generally not lived up to its original promise of minimizing the federal paperwork burden on individuals. The revised guidance is notable not just in urging agencies to better identify administrative burdens, and reduce them in the name of greater equity, but also expands the types of costs that government agencies should track (Executive Office of the President 2022). Paperwork Reduction Act estimates have historically focused on time spent completing forms, but the revised guidance emphasizes other types of compliance costs, such as travel time, and recognizing the role of learning and psychological costs. Agencies are told that the Office of Management and Budget will assess how well they
simplify the request for information, while ensuring the continued utility of the information they do collect;
enhance communication, navigation, and outreach tools and processes to reduce learning costs to the public;
improve information collection and submission processes to mitigate challenges that underserved and marginalized communities may disproportionately experience; and
use leading design practices to assess, evaluate, and then improve forms and information collection experiences.
Other governments could follow. Although certain conditions of the U.S. political context make the presence of burdens more likely—intense polarization, federalism, a patchwork safety net where basic public services are provided through multiple actors—the topic has been fruitfully explored in many contexts across the globe. The United Nations (UN Innovation Network 2021) recently focused on reducing burdens as a priority. In the U.S. context, state governments could fruitfully use the Biden executive order as a model to direct attention to reducing administrative burdens. One opportunity is to partner with the civic tech community as a potential partner with both an expertise in technology and an interest in improving public services (Kennan, Soka and Sullivan 2022; Lowrey 2022).
CONCLUSION: SOCIAL SCIENCE CAN CONTINUE TO PLAY AN ESSENTIAL ROLE IN REDUCING BURDENS
Taking administrative burdens seriously means revisiting how policy is designed and how the administrative state functions. It invites the marriage of multiple skills, such as behavioral science, human-centered design, and deep policy knowledge. One example is the redesign of practices recruiting low-income students with high performance to college (Dynarski et al. 2021). An experiment at the University of Michigan provided personalized communication, which reduced both learning and psychological costs (feelings of uncertainty and inclusion) while activating support from the broader social network of families and college advisors and minimizing application hurdles. This multifaceted approach nearly doubled both applications and enrollments in the target population even without offering any new financial resources.
At a time when U.S. administrative capacity is characterized as being in decline (Klein 2022), the administrative burden agenda offers insight into how investments in capacity could improve the quality of government. Such investments can ensure that the state, rather than individuals, bear the brunt of burdens in public programs and in the long run can reduce administrative costs. For example, the administrative costs for the Social Security program are less than 1 percent of its budget, in large part because of a system that ultimately automated procedures to determine both eligibility and benefit size. This is in stark contrast to programs like SNAP, where administrative costs can run ten times as high (Geller et al. 2018). Variation is also significant across states in terms of administrative costs, which are driven, in part, by levels of administrative burden that make it more expensive to administer each case. County level of administration of SNAP, an example of federalism at play, is tightly linked to more burdens and lower take-up rates, as well as substantially higher administrative costs—24 percent higher—than state-administered programs (Geller et al. 2018).
The Biden executive order has also modeled ways social science research can help, encouraging a wide range of research techniques including “experiential data (including, as appropriate, through randomized controlled trials or other rigorous program evaluation); ethnographic research; feedback from public engagement; human-centered design methodologies such as journey mapping; operational and administrative data analysis; direct observations; examination, from a customer perspective, of how to navigate the agency’s service offerings, apply for a benefit, or comply with a requirement of the agency; observations of customer interaction with the agency’s website or application processes and tools; or observations of customer support service delivery such as activities at call centers” (Executive Office of the President 2021). For social scientists of all stripes then, this is a unique opportunity to advance research to reduce frictions and to help build the general toolbox to reduce burdens.
This double issue showcases how social science can play a role. Although across the articles we see the utility of early theory development in this domain—for example, the ability to illuminate the nature of burdens via learning, compliance, and psychological costs—they also point to the need for a variety of empirical and analytical approaches. The studies balance both qualitative and quantitative approaches. The qualitative analyses leverage in-depth interviews but also rely on observation and consumer complaints as novel sources of data. For quantitative analysis, administrative data play a central role in both diagnosing and solving problems. The lack of prior insights on burdens in many policy domains illustrates the great value of descriptive work that documents the nature and scale of problems previously overlooked.
Much of the original framing of administrative burdens examined how individuals were affected by state action. Articles in this double issue emphasize the need to broaden beyond that perspective. For example, Sackett and Lareau consider the experience of families as a unit of analysis, examining how those experiences transfer across organizations and programs (see also Heinrich et al. 2021). Other work argues for examining general policy regimes as a unit of analysis, rather than focusing only on individual interactions, because doing so provides a fuller sense of the cumulative nature of burdens that individuals face (Moynihan, Herd, and Gerinza 2022; Pierce and Moulton 2023, this volume, issue 5). In a similar spirit, Rauscher and Burns (2023, this volume, issue 5) make the case for examining a combination of burden-reduction reforms in a policy domain rather than changes to just one part of the process, showing that combinations of reforms are more effective than any particular reform at improving infant health and prenatal care. The consistent logic is that understanding the accumulation of burdens sometimes needs a broader lens than examining discrete citizen-state interactions. Other work offers different units of analysis, which can be conceptual—for example, Sackett and Lareau’s idea of institutional knots as a complex blockage of institutional procedures that individuals seem unable to free themselves from—and descriptive, such as Bouek’s (2023, this volume, issue 5) attention to wait lists as a promising venue to understand how state actors manage burdens.
The translation of ideas into government practice is a fraught endeavor. It remains to be seen what will come of the Biden executive order. Whether agencies will act to reduce administrative burdens, how permanent any changes will become, how deeply they will embed themselves within agencies, and the degree to which they become mechanisms to address equity in government are all open questions. Institutionalization begets standardization, and some government employees could easily come to see burden-reduction efforts as one more administrative burden. Government-wide administrative reforms by their nature tend not to be uniformly successful, but to depend on local conditions such as organizational leadership, culture, and the nature of the task. Again, capacity matters: reforms need not just buy-in among lower-level workers, but also investment from a wide range of stakeholders, and new skillsets when it comes to measuring burdens, mapping user journeys, applying principles of human-centered design, and making better use of technology and data. The framework for change, however, is in place. Social scientists have an opportunity to use their research skills as a partner. For researchers interested in this topic both an intellectual framework and a real demand from policymakers exist in a substantive way that did not a decade ago. Researchers also need to maintain a capacity to look beyond specific burden-reduction efforts, both to continue to draw attention to deeper structural causes of problems, and to engage in blue-sky thinking about how to redesign the administrative state to make it work for the people it is intended to serve.
FOOTNOTES
↵1. Author order is alphabetical.
↵2. “A Human-Centered Approach to Government,” Performance.gov, n.d., https://www.performance.gov/cx/projects/ (accessed January 5, 2023).
- © 2023 Russell Sage Foundation. Herd, Pamela, Hilary Hoynes, Jamila Michener, and Donald Moynihan. 2023. “Introduction: Administrative Burden as a Mechanism of Inequality in Policy Implementation.” RSF: The Russell Sage Foundation Journal of the Social Sciences 9(5): 1–30. DOI: 10.7758/RSF.2023.9.5.01. Direct correspondence to: Pamela Herd, at ph627{at}georgetown.edu, Georgetown University, McCourt School of Public Policy, 37th Street NW O Street NW, Old North #100, Washington, D.C. 20057, United States.
Open Access Policy: RSF: The Russell Sage Foundation Journal of the Social Sciences is an open access journal. This article is published under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.