Discover more from Can We Still Govern?
When policymakers add administrative burdens while expanding access
What to do about the good intentions behind program complexity
Note from Don: This week's newsletter is authored by Jordan Kyle, a researcher at the International Food Policy Research Institute, and former employee at the Office of Management and Budget's OIRA. She has thought a lot about the source of administrative burdens in government, which regular readers will know is a central topic of this newsletter. I’m excited to have her share her ideas here.
There is growing evidence that administrative burdens in social programs matter to access, and that we should do something about it. I’ve worked on such efforts in the federal government, helping to design revised guidance to the Paperwork Reduction Act to help to identify and reduce burdens. I’ve left the Office of Management and Budget now, but my time there and my academic research has given me new insights on the sources of administrative burdens, and some suggestions for how to reduce them.
Administrative burdens in social safety net programs are often attributed to two main causes. First, they can arise from “benign neglect”: unwitting small decisions made early in program design that can have outsized effects on the populations the programs are intended to serve. In these cases, burdens flow from bureaucratic processes that are not themselves intended to harm the public—indeed, they are often designed to improve public administration—but end up resulting in onerous processes for individuals. Second, burdens arise from “hidden politics,” when paperwork and wait times are used to purposively limit access to public services.
But there is an important third source of administrative burdens that’s not quite benign neglect or hidden politics: bureaucratic and legislative efforts to expand existing programs to cover new problems and new populations. Such efforts are benign but intentional, with the goal of broadening access rather than reducing it. We might think of this kind of institutional change as “layering;” essentially, adding new rules on top of existing ones.
This type of institutional evolution is common in a variety of domains: for example, the political scientist Paul Light has identified how efforts to improve government performance resemble “tides of reform” that accumulate over time. New reforms are started. Previous initiatives are left in place, with the result that public officials must manage multiple generations of well-intentioned administrative requirements that do not always form a coherent whole.
The motivation for policymakers to add is understandable: they respond to a particular need or political opportunity to make an incremental change. But the implication is that they rarely look at the whole picture. Such sedimentation of changes in welfare programs explains how programs that start off with a simple design can evolve into complex behemoths over time. The desire to respond to particular groups or circumstances with bespoke administrative processes tends to add burdens for all.
The evolution of SNAP: Burdensome but inclusive
The history of food stamps—now the Supplemental Nutrition Assistance Program (SNAP)—illustrates this conundrum.
The 1964 Food Stamp Act was relatively simple. The Act limits eligibility to those “households whose income is determined to be a substantial limiting factor in the attainment of a nutritionally adequate diet.” States are given the responsibility of establishing standards to determine eligibility in practice. Around half a million people participated in the program in its first year. The program expanded throughout the 1960s, but implementation questions started to crop up. The time between application and receipt of benefits in some states was slow, and states had not yet expanded program implementation to all jurisdictions.
A series of major legislative fixes to the program across the 1970s to address these challenges substantially increased the size and reach of the program, while simultaneously ballooning its complexity. The Agriculture and Consumer Protection Act of 1973, for example, required states to expand the SNAP program to all jurisdictions, to include individuals in rehabilitation centers recovering from substance use disorders among eligible populations, and to establish a version of the program that could be ramped up in the case of natural disasters. In the same breath, the definition of income became more complex to allow for these more specific scenarios and types of income.
Similarly, the Food and Agriculture Act of 1977 simultaneously expanded reach and increased burdens in SNAP. It established statutory income requirements, while also formally allowing for a number of allowable income deductions. Spelling out allowable deductions broadens access by acknowledging that households with different incomes and circumstances can face similar levels of food insecurity and poverty. The practice of allowing certain expenses to be deducted from household income before calculating benefits tries to ensure that food insecure households aren’t left out or receiving benefits insufficient to meet their needs even if they are above the general income eligibility threshold.
Such changes added nuance to the program, seeking to cater to people in different situations. But every deduction adds complexity, representing an additional question and documentation requirement on an application form.
The cumulative effect is significant. A single question about income becomes a multi-page questionnaire contemplating tens of different possible income scenarios. Yet, this period also saw significant expansion of the program, reaching over 22 million households per year by the end of the 1970s.
On the legislative front, the simultaneous expansion of welfare programs with more burdens is not surprising. Grand bargains between those interested in shoring up the social safety net and those concerned about program fraud is often necessary to pass legislation.
But the additional burdens are not just about protecting programs from fraud; they also arise precisely because new scenarios are contemplated, and rules need to change to allow for the inclusion of new populations.
Older Americans with high medical expenses are a good example. Based on income alone, many older Americans would receive paltry, if any, SNAP benefits. Yet, older Americans with chronic diseases or living with disabilities often face high fixed medical costs. After covering these costs, they may have as little or less income leftover for food as households with far lower income levels.
Helping this population means allowing them to deduct medical costs from their income to calculate eligibility. Again, this also means additional questions, documentation, and longer forms for a population that struggles with such complexity, thereby lowering program take-up. But program take-up and benefit allotments might be even lower absent these burdensome but ultimately inclusive provisions.
Death by a thousand bureaucratic fixes
Beyond the legislative sources of administrative burden, bureaucracies often intentionally take actions with the intention of “fixing” programs that can simultaneously expand access and increase burden. Given our current era’s legislative gridlock, agencies increasingly rely on administrative actions to alter program rules and to offer new interpretations of statutory obligations. Of course, these bureaucratic alterations can be part of the hidden politics of increasing paperwork burdens to raise the costs of program participation (see immigration policy).
But bureaucratic fixes can go in the other direction as well: increasing access and burden at the same time. Administrators might seek to help people who probably should be served by a program based on their food insecurity, lack of health care, or other qualifying factors, but for whom the existing income and deduction rules don’t work. Or, they might seek to carve out a vulnerable population from a specific requirement. In either case, determining whether or not an individual qualifies for the exception requires seeking more information.
Here’s a very specific example. Persons living with disabilities due to rare diseases are often invited to participate in essential medical research on their conditions. When they are compensated for their time and efforts for this participation, should this compensation really count as part of their income in the same way that it would if they were able to work outside the home for purposes of qualifying for disability benefits? Most people would say no. Congress agreed, and in 2010 amended the Social Security Act to allow income from participating in clinical trials to be excluded for purposes of calculating Supplemental Security Income. Social Security Administration officials then were tasked with clarifying through administrative actions exactly which types of trials and payments count, as well as documentation necessary to prove that the specific case meets the criteria.
Addressing this issue helps to ensure that benefits are not unduly reduced, yet it also means more questions on the application, more documents to keep track of, and more information required of applicants. The cumulative effects of decades of these types of specific fixes result in byzantine application processes collecting seemingly endless amounts of information.
It would be inaccurate to characterize such fixes as benign neglect. They are intentional actions aimed to adapt government processes to a myriad of individual circumstances and complexities, which in turn must be documented. Concerns with fraud can lead to unnecessarily conservative approaches to collecting information. But here bureaucrats are concerned instead about about due process and whether sometime who might be legally eligible for a benefit could be unduly denied due to lack of information: “If we just knew this one extra thing, we could prevent unfairly denying a benefit in x, y, z (specific, and often uncommon ) scenario...” Thus, overly conservative approaches towards either inclusion or exclusion error can spawn administrative burden.
Simplicity versus complexity
In many ways, the tensions between burden and access get to the heart of legal debates about the tradeoffs between bright line rules and flexible standards, between simplicity and complexity.
Bright line rules offer predictability, uniformity, low decision costs, and easier and more consistent implementation. Bright line rules, however, mean accepting that edge cases that don’t fit the rules well might result outcomes that would seem unfair if we knew the facts. Standards, on the other hand, define a set of relevant factors for consideration while leaving room for case-specific flexibility. Standards are messy to implement, however, and require a lot of case-specific information to support deliberation; in other words, more administrative burden.
Think about how this conundrum plays out in program design. New programs that are designed simply from the start are appealing. Under the hood of simple applications, however, are bright line rules that might not always seem fair. An application that relies on data matching to automatically link your employment information with program eligibility looks like beautiful and simple design. But, what if the agency implementing the program lacks employer data for sole proprietorships or small businesses? These applications may be excluded or otherwise treated differently from those from individuals employed by larger entities. Without enough staff, governments will struggle to pursue manual solutions for complex cases who may need the most help. The very things that support simple design like web accessibility and data matching are often least useful for more vulnerable populations.
These tradeoffs are easier when we contemplate user experience and design in the private sector. Indeed, it’s one of the many reasons why private sector design is consistently better than public sector design. Private sector companies can design for their target customer and not worry too much about anyone else. But the public sector needs to serve all, even those with complex scenarios and limited access to internet or mobile devices. Being too far at either end of the simplicity-complexity spectrum is likely never going to work for a public program trying to balance equity, fairness and procedural justice, program integrity, access, and high quality user experience.
Ugly applications might sometimes be the messy way that bureaucracies try to reach hard-to-reach populations or those with unusual circumstances. The question on the form—the source of burden—might also a way of saying, “we care about this issue and we didn’t forget about you.” Of course, additional burdens can equally be a way of obsessing over program fraud: we care more about the small chance that one person may benefit from a loophole than we do about increasing paperwork burdens for millions of the U.S.’s neediest citizens. It’s difficult to address burdens in large, long-lasting programs when they arise from so many different sources and with such distinct underlying logics and motivations.
Three ways to reduce burdens in large, long-standing programs
None of this is to say that it’s impossible to reduce burden in large, long-standing programs. Rather, would-be reformers must proceed with caution and awareness about why specific burdens have developed over time and how they may link back to access. There are as many ways to reduce burden as there are sources of burden, and this recent OMB memo offers a lot of very specific ideas. I want to throw out three additional pathways for burden reduction that I hope the current administration, with its emphasis on good customer experience and burden reduction, considers.
1. Creating two-track applications
First, policymakers should acknowledge that they might not be able to create one beautiful and simple administrative process that works for all. Instead, they may want to embrace both simplicity and complexity by having two application tracks, which I’ll call the “default” and the “context-specific” routes.
Although it’s true that case-by-case considerations may be important for specific populations, many individuals qualify for programs without any of the detailed information. These individuals should be able to choose a blissfully brief application experience that asks only the essential questions, and they shouldn’t even have to scroll through the more complicated scenarios. Some people may inadvertently lose benefits who may have otherwise qualified if they had taken the “context-specific” route. It is difficult to reduce burden without allowing for this possibility.
The risk of taking the wrong route can be significantly curtailed by providing excellent up-front explainers and navigational tools. For example, an application could advertise that households with individuals living with disabilities and with high medical expenses or shelter costs should likely take the “context-specific” route, but ultimately leave the choice to individuals. It is possible that the most important thing to that individual is to qualify quickly, and they just want to use the default application.
Choosing between default and context-specific routines is analogous to what Americans do every year with their federal income taxes. Many opt for the standard deduction and never go through the exercise of checking whether itemized deductions may have yielded preferable tax treatment. In some cases, it’s possible that they leave money on the table, but the time it would take them to figure it out is also valuable.
Designing a simpler default—and allowing people to use it—would vastly improve the experience for millions of individuals. Meanwhile, if the majority has a far simpler and less burdensome experience, this could free up resources from frontline service providers and call centers to help support the smaller number of individuals with specific cases.
2. Reducing the cognitive load
Some compliance burdens will be very difficult to address over short time horizons. But, the government could do a much better job of making dizzyingly complex processes easier to navigate. When tasks are multi-step and complex, allowing for non-linear processes can be very helpful. Maybe I have my income information on hand, but not my supporting documentation. Or, I can easily fill out information on my household composition, but I need to talk to my spouse to fully account for household shelter and utility expenses. Processes that lay out in advance all of the steps that will ultimately be required, allow for frequent interruptions and interim saving, and that clearly delineate mandatory from voluntary information can go a long way in reducing overall cognitive load.
Better explainers for each piece of information and why it is needed are also sorely missing from benefits applications. Explainers should not come in the form of 30-page manuals separate from the application which themselves become a source of administrative burden, but in clear, simple language side-by-side with the question on the form for which they are relevant. Hints and examples could be provided that guide individuals toward accurate responses and documentation. Is there a certain type of documentation that individuals frequently submit incorrectly? Or something that people think is needed, but is in fact not needed? These types of common errors could be pointed out, with supporting photographs and images. Side-by-side explainers could also contain reminders of when a question does not need to be answered by everyone.
In addition to reducing cognitive load of understanding what is being asked for, question prompts that offer neutral explanations on why a question is on the form in the first place also have the potential to reduce psychological costs. For example, if the section of an application asking about medical expenses began with a prompt like, “Some families experience high fixed medical costs which regularly reduce their net income. If this describes your family, then you may want to fill out the following questions so that the government can use this information to more accurately calculate your potential benefits.” A neutral question prompt like this would convey both that a specific section of the form is voluntary and how the information is used. But it also potentially sends a very different message, one of caring about families with high medical expenses, compared to a long list of questions about medical expenses without any explanation. User testing could go a long way to finding appealing and understandable explanations, and research could investigate whether they affect psychological costs of a given process.
3. Shout it from the rooftops
If program officials recently changed a rule or simplified a process, they should relentlessly communicate this to internal and external stakeholders. Public officials need to claim their wins, and tell their success stories, because no-one else will do it for them.
This is also important because so many different parties and entities are involved in policy implementation, including the clients receiving services. They all need to be on the same page when it comes to burden reduction efforts. The frontline service providers that clients interact with could be employed by federal, state, or local governments. Ensuring that they know about changes within the programs that they operate is not always easy. The U.S. welfare system also includes a myriad of third-party actors whose knowledge of program rules is equally essential, including civil society organizations, advocacy groups, legal service providers, and entities that may receive contracts or grants to run essential benefits programs.
If any one of these actors isn’t aware or doesn’t understand the rule change, the beneficiary won’t experience the burden reduction in practice. Recent research on Special Supplemental Nutrition Program for Women Infants and Children (WIC) found that many beneficiaries did not know about program flexibilities enacted during the pandemic. Even when they did, they sometimes faced third-party retailers who did not know about the flexibilities, leaving them unable to benefit from them. This is another source of burden: beliefs at the front lines that a certain amount of red tape is required that does not actually match formal requirements.
Especially in states that prefer a certain amount of hassles in the safety net, it is essential that beneficiaries have full knowledge of their rights and the tools to push back if they encounter poor information. This is a universal issue. In work on how to increase access to benefits programs in Indonesia, we found that creating common knowledge between local implementers of programs and program beneficiaries was more effective in increasing implementers' adherence to program rules than sharing private knowledge about program rules with beneficiaries. For individuals to claim their rights effectively, they needed to know that program implementers not just knew both about those rights but also were aware that all beneficiaries also knew about those rights. Common knowledge increased beneficiaries’ bargaining power and increased scope for collective action.
There is more that government could do to reduce burdens. Some of these are well-documented: lengthening recertification periods, appropriately staffing frontline service providers, more frequent contact attempts across multiple contact methods, and so on. But the sort of incremental fixes I outline here are perhaps less obvious, and can often be done without significant new resources, making them a good place to start.
Jordan Kyle is a Research Fellow at the International Food Policy Research Institute (IFPRI) in Washington DC, where her research focuses on how to improve the delivery of public services across different countries and contexts. Previously, she was a policy analyst at the Office of Information and Regulatory Affairs (OIRA) at the U.S. Office of Management and Budget (OMB). @jkyleindc
Can We Still Govern? is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.