Collecting information is an inescapable necessity of governance. The Constitution requires the federal government to collect information from every person in the country every 10 years: the census.^[1] Paying taxes and fees takes more time than all other federal information collections combined, but collecting revenue is necessary to fund government programs.^[2] Important statistical data are gathered through information collections of many kinds. Noncitizens must provide personal information before they can be granted visas. Environmental, health, and financial regulations almost always require information collections—for example, emissions readings, accident information, or financial statements—to be effectively implemented. And almost all spending programs require information collections to function. For example, healthcare programs like Medicare and Medicaid must collect information to determine eligibility for services. Similarly, grant programs—such as those contained in the CHIPS and Science Act—often mandate that applicants provide highly detailed information justifying why their grant application should be accepted over others.

This comes at a cost, however. Information collections impose administrative burdens on the public—the time, money, and psychological costs involved in responding to them—that are substantial. Most obviously, information collections can impose excessive and differentiated burdens on businesses and other entities that generally do not benefit from the collection. For example, small businesses will tend to lack specialized compliance teams; as a result, imposing the same paperwork burdens on them as on large businesses may reduce competition and favor large incumbent firms. Administrative burdens also weigh on the beneficiaries of programs. These burdens often cause far fewer people to receive benefits than are eligible.^[3] For example, data suggest that as few as 18 percent of families eligible for Temporary Assistance for Needy Families get the benefits that they are entitled to, and only 61 percent of those eligible for Supplemental Security Income (need-based aid for the elderly, blind, and disabled) receive the benefits that they are entitled to; other prominent programs fall somewhere in the middle of these two programs.^[4] Perversely, administrative burdens often reduce uptake most for the neediest.^[5] 

How the Paperwork Reduction Act came to be

In light of the benefits and burdens of information collections, statutes have imposed a variety of procedural requirements on agencies. These procedures impose burdens on agencies, much like information collections impose burdens on the public. Accordingly, those burdens should be weighed against the benefits of these agency procedures: that they tend to reduce unnecessary burdens on the public, or enhance the value of the information collected. The most important set of procedural requirements for information collections are codified in the Paperwork Reduction Act (PRA), first enacted in 1980. However, it is important to remember that the PRA process exists within a broader, common set of procedural requirements governing agency action, the most important of which were codified in the Administrative Procedure Act (APA) in 1946.

Much of what became the PRA arose as a response to concerns about coordination and improvement of federal information collections in the 1930s. These concerns “increased markedly with the enactment of the early New Deal programs because ‘they involved not only [general-use] statistics but [also] the administrative figures required to establish and enforce . . . regulations for individual enterprises.’” ^[6] Recognizing that contradictory or poor-quality information could impede implementation of New Deal programs, President Roosevelt established the Central Statistical Board (CSB) in 1933 to coordinate and review efforts to collect information from the private sector.^[7] Congress approved of these efforts, and strengthened the approach in 1935 by enacting the Central Statistical Act, which expanded the CSB’s jurisdiction to cover statistical needs more broadly.^[8] 

By 1938, President Roosevelt wrote to the CSB that he remained “concerned over the large number of statistical reports which Federal agencies are requiring from business and industry,” and in particular was concerned about excessively detailed or duplicative reports.^[9] President Roosevelt wanted to enhance the “efficiency and economy” of information collections, “both to the Government and to private industry.”^[10] In the course of studying the issue, the CSB solicited public feedback, and volumes of angry complaints poured to them “from business, particularly small business.”^[11] Following the CSB’s work, President Roosevelt transferred the powers and functions of the CSB to the Bureau of the Budget (BoB) in 1939; the Bureau was the predecessor of the Office of Management and Budget (OMB), “which itself was transferred from the Department of the Treasury to the newly-created EOP,” the Executive Office of the President.^[12] This shift strengthened presidential oversight and control, consistent with the view of progressives that uncoordinated agency action—in this case, information collection—will often be myopic or even captured.^[13]

More legislation followed. The PRA’s direct predecessor, the Federal Reports Act of 1942, was passed to improve coordination and quality in information collections.^[14] The Act “empowered the Director of the Bureau of the Budget to determine if multiple agency information collections were duplicative or unnecessary and, in such cases, to bar agencies from collecting that information or require a consolidated information collection.”^[15] It also “required that agencies not conduct or sponsor an information collection with identical items from ten or more people unless the agency submitted the plans or forms to be used in the collection and any relevant regulations to the Director for approval.”^[16] And to guide this effort, the Act “directed that information collections ‘be obtained with a minimum burden upon . . . persons required to furnish such information, and at a minimum cost to the Government’ while also ‘in a manner to maximize the usefulness of the information.’”^[17] From the beginning, “[t]he limited nature of the Act” and its ability to improve information collections “was generally recognized.”^[18] Continued problems with agency-specific approaches led to the enactment of “the Budget and Accounting Procedure Act of 1950, which granted the Bureau the power to promulgate regulations ‘for the improved gathering, compiling, analyzing, publishing, and disseminating of statistical information’ by agencies.”^[19]

The Bureau of the Budget “achieved some remarkable results under its coordinating function by developing uniform standards” for federal statistics,^[20] such as the North American Industry Classification System (NAICS) codes.^[21] However, “[o]ver the years, congressional committees periodically held hearings and issued reports complaining that either the Federal Reports Act was inadequate” to its other task—reducing unnecessary or duplicative information collections—“or that BoB (and later OMB) was not adequately enforcing the Act.”^[22] Complaints also centered on the Federal Report Act’s exclusion of tax forms, and OMB’s inability to identify and correct agencies that flouted the Act’s requirements.^[23] For example, “for more than a decade the Department of Health, Education, and Welfare (HEW) had required states to prepare several detailed annual plans for welfare programs without ever having submitted the forms for clearance.”^[24] Another common view was that the OMB “lacked initiative in pursuing the directives of the Act, and refused to adequately staff or equip the office responsible for carrying out the Act.”^[25] Yet these complaints “contrasted sharply with [other] complaints that OMB” had been over-aggressive in enforcing the Act, “preventing the regulatory agencies from fulfilling their statutory functions.”^[26] Reformers hostile to the administrative state, such as Ralph Nader, attacked OMB’s role in seeking changes to information collections on the basis of concerns about their methodology or cost that had been raised by industry groups.^[27]

These Naderite critics won out as the failure of the Federal Trade Commission (FTC) and Federal Power Commission to “complete[] investigations into the oil and gas industries” in the wake of the 1973 Arab oil embargo was blamed on OMB review of their information collections.^[28] The result was a new statute “exempting independent regulatory agencies from OMB oversight” of their information collections, adding to existing exemptions for tax information collections and others.^[29] It transferred the authority to review these agencies’ reporting requirements for duplication and unnecessary burden to the congressional investigative arm then known as the General Accounting Office, but not the authority to assess whether the information itself was needed.^[30] This happened not only over OMB’s objections, but also even though “the FTC said the Federal Reports Act change was not significant” in slowing its investigations.^[31] 

In 1974, as “constituent complaints about paperwork burdens” continued and Congress grappled with “perceived difficulties with GAO’s role as overseer of the regulatory agencies,” Congress responded by creating the Commission on Federal Paperwork.^[32] In 1977, the Commission published its final report, describing “several failures of the existing system of control,” including that “the controls in existence had been designed to control statistical paperwork, rather than program management and operation which constituted the vast bulk of federal paperwork” and that “existing controls were spread among a number of different agencies.”^[33]

In 1979, President Carter responded by issuing Executive Order 12174, “Federal Paperwork Reduction.”^[34] Much of this executive order was included in a parallel bill, which was enacted as the Paperwork Reduction Act of 1980.^[35] The PRA “moved swiftly through committee to the floor” and “passed overwhelmingly,” as it was backed by President Carter, Congress, business groups, state agencies, and some public interest groups, and faced “weak to nonexistent opposition” that came almost exclusively from the heads of some (but not all^[36]) federal agencies.^[37] The PRA was amended in 1986, in part to require Senate confirmation for the Administrator of the Office of Information and Regulatory Affairs (OIRA).^[38] In 1995, the PRA was rewritten in full, changing several key provisions. Most dramatically, it added “a requirement that agencies seek public comment concerning proposed collections of information through [an additional] sixty-day notice to the public prior to submission to OIRA for clearance” and the existing thirty-day public comment period.^[39] 

How the Paperwork Reduction Act works

The PRA aimed to strengthen OMB’s role as the primary overseer of government information collection, building on its experience under the Federal Reports Act and its regulatory oversight responsibilities.^[40] To address concerns about OMB’s inattention and limited staffing in its role overseeing information collections, it created the Office of Information and Regulatory Affairs (OIRA) within OMB, and charged it with OMB’s information collection responsibilities.^[41] The PRA clarifies that OIRA’s review of information collections has two purposes: to “minimize the Federal information collection burden, with particular emphasis on those individuals and entities most adversely affected” and “maximize the practical utility of and public benefit from information collected by or for the Federal Government.”^[42]

Like the Federal Reports Act that preceded it, the PRA applies when agencies seek to collect information from 10 or more “persons” (individuals, firms, states, or other entities).^[43] The biggest change between the two statutes was the specificity with which the PRA details the clearance process. Following the 1995 amendments, it works like this:

• First, an agency publishes a notice of a proposed information collection in the Federal Register and solicits public comment for 60 days.^[44] 
• After reviewing the public comments, the agency submits the information collection (potentially revised in light of public comments) to OMB for review, publishes another notice in the Federal Register, and allows for at least 30 days of public comment on the latest version of the information collection.^[45] How OMB conducts its review varies with the particular information collection.^[46]
• OMB must approve or deny the collection within 60 days of submission (i.e., within 30 days of the 30-day public comment period ending),^[47] and the information collection can only be approved for up to three years.^[48] If the information collection is still needed beyond the period it is approved for, the agency must go through the entire process again, and complete the process by that date.
• A procedure exists to skip these public comment periods and related requirements, if certain criteria are met (colloquially referred to as an “emergency” information collection); OMB’s approval of such an emergency information collection is limited to 180 days.^[49] 
• When information collection is part of a proposed rule, the PRA public comment process is streamlined. A single 60-day public comment period, conducted alongside the rule’s comment period under the Administrative Procedure Act (APA), suffices under the PRA.^[50] 

The PRA also structures agencies in a variety of other ways. For example, each agency is required to have a Chief Information Officer to oversee the Act’s functions.^[51] 

The PRA tasks OIRA with a variety of other duties. For example, OIRA is tasked with also directing and overseeing “privacy, confidentiality, security, disclosure, and sharing of information.”^[52] In 2000, the Information Quality Act amended the PRA to task OMB with developing standards for federal statistical information.^[53] This piece only focuses on the PRA’s provisions relating to information collections generally.

OMB guidance on the Paperwork Reduction Act

To address concerns with PRA implementation, OMB has issued nine memorandums and one frequently-asked-questions document. Each clarified agency actions that are not subject to the PRA or other flexibilities.

• The first was issued in April 2010 and provided a primer on how PRA processes worked at the time of its issuance.^[54] Of note was its inclusion of OIRA’s longstanding view that for “non-substantive changes” to information collections an “agency is not required to seek public comment”; only OMB confirmation that the change is indeed non-substantive is required.^[55]
• The second was also issued in April 2010 and clarified that many website collections—social media and web tools for general solicitations of public input, ratings and rankings of posts or comments on websites, public meetings, “wikis” or other collaborative drafting platforms, information collected only for the purpose of creating a user account for an agency website or completing a financial transaction, and contests—are generally exempt from the PRA.^[56] 
• The third was issued in May 2010 and elaborated on the decades-old practice of approving “generic” information collections, which was also discussed in the first memo.^[57] A generic information collection request allows an agency to conduct many similar information collections as part of a proposed plan with just one approval, that is, without going through new public comment periods each time.^[58] For example, an agency might have plans “to gather views from the public through a series of customer satisfaction surveys” and be able to conduct all surveys with just one approval.^[59] OIRA reviews generic clearances within 10 business days.^[60]
• The fourth was issued in December 2010 and clarified that many kinds of scientific research—such as clinical information from individuals being treated or examined in connection with research on the prevention or treatment of a disorder—are either exempt from the PRA or qualify for streamlined review (such as the use of generic clearances for pre-testing and development work for surveys, or emergency clearances in response to a public health epidemic).^[61]
• The fifth was issued in June 2011 and built on the prior memorandum by creating a new “fast track” process for generic information collection requests.^[62] The fast track process was primarily meant for usability testing of agency “services, products, or communication materials.”^[63] The special feature of this process was cutting OIRA review down to five business days (with automatic approval if no action is taken within that period).^[64]
• While not a memorandum, OIRA provided guidance about a new option for use of common forms in November 2012.^[65] It allows multiple agencies to use a single form, meaning that—after one agency goes through the process—any other agency can use the common form without going through public comment.
• The sixth was issued in September 2014 and built on the second memorandum.^[66] It specifically clarified that web-based interactive technologies—such as data search tools or calculators—are also generally not subject to the PRA.
• The seventh was issued in September 2015; while primarily devoted to the use of insights from behavioral science, it discussed how agencies applying those insights would benefit from OIRA’s longstanding approach to “de minimis” changes to information collections.^[67] It noted that de minimis changes—ones “that affect the look and feel of a collection, but do not change the nature or type of information (e.g., data elements) collected,” and “do not increase the burden of a collection, though they might reduce its burden”—do not require OMB approval under the PRA at all.^[68] 
• The eighth was issued in July 2016; it discussed and organized the flexibilities discussed in the prior seven memos.^[69] It also discussed the process for emergency reviews.^[70] Most importantly, it communicated OIRA’s longstanding position that information obtained through direct observation by an agency, or through non-standardized conversations in connection with such direct observations, is not generally subject to the PRA.^[71]
• The ninth was issued in November of 2024; it built on the eighth’s discussion of direct observation and non-standardized conversations to clarify—in the clearest terms possible—that most usability testing is not subject to the PRA.^[72] It also discussed how agencies can quickly implement improvements to their forms and websites based on the results of their usability testing.

This last piece of guidance went through usability testing itself, to ensure that it would unequivocally clear up a longstanding confusion that usability testing would always require first going through the full PRA process.^[73] But the need for its issuance speaks to one of the key problems of the PRA.

Agencies adopt overly conservative PRA interpretations and baroque clearance procedures

The OMB guidance discussed in the previous section kept being issued for a reason: It was fighting the “cascade of rigidity” that afflicts the implementation of so many laws.^[74] This cascade occurs as some or all of those implementing the law come to believe that they must hew to conservative interpretations to avoid censure for overstepping their authority.^[75] Depending on the context, such “conservative” interpretations may be overly narrow (e.g., of the PRA’s exceptions) or excessively broad (e.g., of the PRAs’ definition of “information”). This rigidity is reflected in many agencies’ slow implementation of the OMB memoranda. For example, as noted above, OMB’s 2016 memorandum stated its long-held view that usability testing is generally exempt from the PRA.^[76] Yet some agency PRA officials continued to claim that usability testing was not exempt from the PRA for years afterwards.^[77] That led OMB to issue its 2024 memorandum, which employed usability testing itself to ensure that it was as clear as possible.^[78]

Part of the problem was likely that OMB’s guidance was only that: guidance. It provided agencies notice of OMB’s view of what the PRA required. But that guidance was not binding on courts if an information collection was challenged. While information collections are rarely challenged in court,^[79] the prospect of a lawsuit may nonetheless affect the agency general counsels who are generally in charge of ensuring compliance with the PRA (rather than programmatic divisions).^[80] As a matter of training and culture,^[81] agency counsels will tend to adopt conservative interpretations of statutes—such as by not relying on guidance interpreting a statute—to reduce the risk that an action exceeds the agencies’ authority.^[82] They will also often advise an agency to take steps that can avoid a potential accusation that the agency acted unlawfully. Over time, the prudential decision by an agency general counsel to go through unnecessary public comment procedures can harden into a rigid understanding that all such information collections must go through the PRA’s public comment procedures.

This “cascade of rigidity” is reinforced by the “bureaucratic anxiety cycle” whereby concern about critics leads “bureaucrats to seek refuge behind … layers of procedure” in order to show that they did things “by the book.”^[83] Incentives contribute to the problem, beyond training and culture: Someone will almost certainly not be reprimanded by a superior for incorrectly determining that something exempt from the PRA is subject to its most onerous requirements; conversely, the same person may get in serious trouble for incorrectly opining that something subject to the requirements of the PRA was exempt. This, in turn, can lead to unnecessary layers of review in agencies before an information collection is even sent to OIRA; these layers of review often delay information collections even more than the public comment process. Agencies could remove excessive layers of clearance procedures without any change to the PRA, but the minimum delay imposed by the PRA’s comment periods reduces their incentive to do so. Agencies would have a greater incentive to fix their internal clearance procedures if doing so would hasten the average time for an information collection to go from developed to approved from nine months^[84] to one or two months, rather than the four or five months needed to go through both public comment periods.

Most information collections receive little useful public input and most reviews of information collections have little benefit

For decades, the delays associated with the public comment periods imposed by the PRA have been criticized on the grounds that very few information collections attract public comment. Looking at a random sample of approved information collections from May 2004, GAO noted that “[a]n estimated 7 percent” of “the 60-day notices in the Federal Register” had “received one or more comments.”^[85] Stuart Shapiro found in 2012 that a similar “8.7% of active collections received comments the last time that they were submitted for review.”^[86] Shapiro also conducted “[i]nterviews with agency officials indicat[ing] that meaningful public participation in the comment process was concentrated in a small number of collections.”^[87] To be sure, comments on the collections that receive them may be particularly valuable; for example, social scientists providing comments on statistical collections. But for the vast majority of cases regarding websites and forms, direct usability testing and user feedback is a far more effective method of gathering public input than hoping for responses from the very few who will ever respond to a Federal Register notice.^[88]

In essence, the problem is two different broken policy-implementation feedback loops.^[89] At a more granular level, the PRA includes two public comment periods to ensure that information collections can benefit from public input before burdening individuals. But with no input received in the overwhelming majority of cases, the public comment periods generally do not inform agency policy. And at a higher level, the PRA itself has not been updated to reflect this fact, despite decades of experience. As is often the case for Congress and the executive branch, “the separation of powers … creates structural distance between policy and its implementation.”^[90]

In addition, OIRA simply reviews too many information collections. Shapiro—looking to both public statistics and information from interview subjects—roughly estimated that only a few hundred information collections are improved by OMB each year, out of a total of a few thousand.^[91] In part, this is evidenced by the large number of routine renewals of information collections that OIRA approved with no change.^[92] This too reflects a broken policy-implementation feedback loop between Congress and the executive branch.

The costs of PRA public input processes and reviews are high

Both OIRA and agency personnel must spend time complying with the PRA. In 2012, Shapiro estimated that the opportunity costs of that time, along with the cost of Federal Register fees, were around $7 million to $10 million dollars per year.^[93]

Substantially more important are the costs of delay imposed by the PRA. Recall that an agency must go through both a 60-day and 30-day public comment period before issuing an information collection. If an agency receives comments during the 60-day comment period, it must take time weighing those comments and determining how to respond. Further, the Federal Register process—from submission to publication—generally takes three business days (but can take longer), adding over a week of additional delay per information collection.^[94] And as discussed previously, an enormous amount of extra time is consumed by unnecessarily conservative—and, in many cases, baroque—internal agency clearance processes each time a new step must be completed.^[95] Finally, OIRA review itself can be lengthy, particularly at times when its relatively small staff is overwhelmed with a large number of low-value routine reviews.

All of this means that agencies may not seek the information at all, if the agency determines that the benefit of the information collected is not worth the time that it would take to complete the approvals. That means that agencies often act with less information, and accordingly make worse decisions. This is another case of excessive “procedure … overburdening the bureaucracy, reducing its ability to deliver meaningful outcomes.”^[96] The costs of this are likely immense, but difficult to quantify.

The fact that a large number of reviews have little benefit should not obscure the importance of OIRA review to the remainder. Indeed, OIRA fills a key niche in the structure of the executive branch. First, the mere existence of the PRA and OIRA review can provide an opportunity for pushback within agencies against information collections that, while useful to the government, do not have benefits that justify their burdens on the public.^[97] That is one reason why the first Trump administration’s “Improving Customer Experience with Federal Services” initiative and the Biden administration’s burden reduction initiative both successfully harnessed OIRA’s PRA review process to achieve their goals.^[98]

More directly, OIRA’s ability to work across information collections and regulations often helps it identify important opportunities for reducing unnecessary burdens. OIRA has additional advantages stemming from its ability to convene relevant expertise from other parts of OMB, such as the Office of Federal Procurement Policy (for procurement issues) or budget examiners (as agencies frequently benefit from OMB support and guidance regarding potential budgetary authority to fix a problem). Often, OIRA’s value is in driving big changes that require multiple agencies—or multiple components of one agency—to not just work together, but also redistribute responsibility. Removing unnecessary burdens quite frequently will mean more work for one agency, even as it generates bigger time and paperwork savings for the public and other agencies. For example, if multiple information collections from different agencies are consolidated, the collecting agency will bear costs associated with collecting and sharing some information that only the other agencies have need for. Without OIRA to convene (and elevate and resolve the issue, with White House support, if necessary), it is far more difficult to get agencies and components to agree to such fixes.^[99] But this is all very abstract. To see how OIRA review under the PRA can be critical for reducing unnecessary burden, consider some specific cases.

In 2024, hundreds of civil servants were nominated for the Samuel J. Heyman Service to America Medals.^[100] Only 24 of those nominated were finalists for the medals; one of them was an OIRA desk officer, Kyle Gardiner.^[101] Gardiner was a finalist due primarily to his work with the Social Security Administration (SSA) to bring its disability recertification process online.^[102] Not only did he help push SSA’s systems into the digital age—a huge achievement in its own right—Gardiner also worked with SSA to ensure that applicants would not have to re-fill each aspect of the form during a recertification, instead having the form pre-populate with the applicant’s prior data.^[103] In total, his work with SSA saved enormous amounts of applicants’ and agency staff’s time.^[104] 

Gardiner is no exception. OIRA desk officers—when supported by political leadership—routinely reduce unnecessary administrative burdens by leveraging the PRA. Ricky Revesz, who previously served as the Administrator of OIRA, has recounted many such examples.^[105] Sticking to SSA, when OIRA reviewed an information collection on overpayment waivers and repayment rate changes—processes used to help beneficiaries who accidentally received overpayments that in theory should be returned—OIRA conditioned approval of the collection: SSA had to commit to work with OIRA on addressing how unnecessary overpayments were caused by its current rules and information collections, and how the waiver process could be improved.^[106] An overpayment is no small thing—an elderly or disabled individual will often lack the means to repay an overpayment that they already have spent.^[107] Overpayments have declined as a result of this work.^[108]

As examples beyond SSA illustrate, OIRA often prevents duplicative information collections. “For example, two components of the Department of Homeland Security—Customs and Border Protection (CBP) and the Coast Guard—require commercial vessels and their operators to provide several data submissions when arriving in the United States and when departing it.”^[109] Because it reviews both CBP and Coast Guard information collections, OIRA noticed the duplicative collections, and has worked with CBP to eliminate the redundancy.^[110] Similarly, the Centers for Disease Control and Prevention (CDC) recently updated the process (including information collections) by which individuals can import dogs into the United States.^[111] But CBP also controls much regarding imports, so OIRA ensured that only one agency would have to receive and verify each form.^[112] And ultimately, OIRA ensured that a new rulemaking was issued to create one unified approach to the importation of dogs.^[113]

OIRA can also simply identify unnecessary regulatory requirements through its review of information collections. For example, when OIRA reviewed a Federal Motor Carrier Safety Administration (FMCSA) information collection related to Driver Vehicle Inspection Reports (DVIRs),^[114] it noticed that almost all (95%) of DVIRs were filed after a driver found no defects to report.^[115] OIRA flagged this for the regulatory side of the agency, and worked with agency regulators to ensure that a rule was promulgated to waive the unnecessary requirement to file DVIRs when no defect was found; this both eliminated an enormous amount of administrative burden and allowed FCSA to better focus on the important reports that noted defects.^[116] 

OIRA review also is important in areas where individual agencies lack incentive to act or relevant expertise.^[117] For example, OIRA has pressed for more consistent translation of information collections that will likely reach individuals who do not speak English.^[118] Translations may not be prioritized by agencies for a variety of reasons, including that applicants may need to simultaneously fill out many agencies’ collections while addressing a given need (e.g., dealing with many agencies after a natural disaster), and therefore translation of only one agency’s materials does little to reduce burden on its own. By steering multiple agencies to move simultaneously, OIRA can resolve such collective action problems. OIRA review of information collections also ends up playing a critical role in allowing its statistical policy functions to be effective. When OMB updates a statistical policy directive, it is critical that all agencies comply as quickly as feasible in order to have data consistency across the federal government.^[119] OIRA uses terms of clearance to ensure that agencies—who, again, do not internalize the costs of inconsistent data formats across agencies (and often have agency-specific reasons to prefer their own data formats)—comply with updated statistical policy directives.^[120] 

All of this indicates that OMB review can add substantial value in an important set of cases. But still, it should not be forgotten that this set of cases is a rather small subset of all information collection reviews.

The PRA process elaborates separate procedural requirements within the more general requirements of the APA. This observation can point the way to several obvious reforms. It is possible to keep much of the benefits of the PRA process while eliminating most of the costs, simply by making statutory changes that align the PRA’s requirements to more closely align with the structure of the APA, which governs most other agency action.^[121]

Statutory amendments to reduce unnecessary procedure

Reform 1: Allow OMB to approve information collections for up to five years, rather than the current statutory maximum of three years.

Moving from a three-year approval period to a five-year approval period could reduce the number of information collection approvals by up to 40% in one fell swoop.^[122] This recommendation goes a bit further than the Administrative Conference of the United States (ACUS) 2012 recommendation to amend “the PRA to permit OMB to define a subset of collections that could be approved for up to five years in order to enable OMB to shift its focus to those information collections that require the most scrutiny.”^[123] Rather than trying to define a subset of collections that can be approved for up to five years in statute, it is better to allow those implementing the law to try different approaches, allowing for OMB to close the policy-implementation feedback loop itself. For example, OMB may initially start with a case-by-case approach (e.g., defaulting to five-year approvals with exceptions, as in the status quo for three-year approvals). After some time learning what categories of collections benefit from another approach, it may use its rulemaking authority (see Reform 4) to create clear rules for those categories of collections. Flexibility allows for experimentation and learning over time, rather than a rigid system that requires a new statute to fix.

Reform 2: If an information collection will have continuing use, agencies should have the option to forgo public comment if they make regular updates to the information collection on the basis of ongoing public feedback.

A statute could remove the statutory maximum altogether. But it is doubtful that this would have much in the way of additional benefits over a five-year maximum, and it could well have serious downsides. The problems that an agency is addressing by collecting information, as well as technology and other developments in the world, evolve over time; this makes it important to reconsider ongoing information collections regularly. As a result, any agency should periodically reconsider its ongoing information collections—certainly not less often than, say, once a decade. And as noted previously, OIRA review provides value in the form of solving for “collective action problems” and other forms of agency externalities, such as cross-government consistency in statistical data or coordination with other agency information collections.^[124] But OIRA has little to no ability to force an agency to update an information collection before it expires. So even with no cap, a prudent OIRA would almost never approve collections for extremely long (or indefinite) periods, because this would inhibit its ability to address these agency externalities. Five years—which matches the ACUS proposal—may not be optimal in all cases, but any exceptions are better dealt with through regulation (see Reform 5) than statutory reform. Given that unnecessary re-approvals waste both agency and OIRA time, there is more than adequate incentive to address any such problems through regulation.

Reform 3: For all other information collections, the PRA should require only a single 30-day public comment period.

With the advent of digital forms and websites, whole categories of information collections can receive better and more useful public input on a continuous basis,^[125] rather than in response to Federal Register notices that are posted every few years and overwhelmingly ignored.^[126] And OIRA already has a process that could easily be expanded to cover such changes: the non-substantive change process.^[127] Accordingly, the PRA should be amended to clarify that agencies can skip public comment altogether if they provide OMB with an information collection that specifies that they will regularly (the statute could specify more precisely, e.g., no less frequently than every 90 days) review—and if appropriate, update—their information collection on the basis of public feedback that the agency can receive and review in real time. These would have to follow best practices for collecting feedback; guidance on how agencies can effectively gather this type of feedback has already been developed by OMB.^[128] To ensure that agencies follow through, OIRA should be given the authority to reduce the period that such a collection is approved for in the event that an agency reneges on the public feedback process that it committed to.

Some information collections do not have continuing use, or it may not make sense to pursue them through the approach outlined in Reform 2 for a variety of other reasons. For such collections, the PRA’s requirement of two separate comment periods—one lasting 60 days and the other lasting 30 days—is difficult to defend. As noted previously, only a tiny proportion of information collections receive any public comments at all.^[129] In particular, the PRA’s process is strikingly inconsistent with how even the most important rulemakings are governed by the APA. For such rulemakings, only one public comment period is required, generally lasting a minimum of 30 days.^[130] And any information collection that accompanies a rulemaking also is not required to go through a second public comment period, creating an odd inconsistency in treatment. Reforms should rectify this mistake from the 1995 recodification of the PRA, instead aligning it with the APA.^[131] This change would, on its own, eliminate more than two-thirds of the time that it currently takes to get an information collection approved.^[132] Agencies would provide information collection material to the public before this comment period begins, as they currently do before the second, 30-day comment period. OMB review would now overlap with the start of the sole public comment period (just as it currently overlaps with the second, 30-day comment period); this would allow the agency and OMB to benefit from public comments, when they occur prior to OMB concluding review (no more than 30 days after the end of the public comment period, as in the status quo).

Reform 4: OMB should be empowered to waive public comment on an information collection for the same reasons that public comment can be waived for a rulemaking.

As discussed previously, the PRA has an “emergency” approval that allows an information collection to be exempt from public comment requirements for up to 180 days (to still be operative after that point, it must be approved again through the regular process). But the criteria for an emergency approval differs from the criteria for exempting a rule from public comment requirements under the APA (“impracticable, unnecessary, or contrary to the public interest”).^[133] This can have particularly absurd consequences when an information collection is necessary to implement a rule. The distinction between the two can lead to the rule, but not the information collection, being exempted from the public comment period requirement (or vice versa). But either one not being exempted causes enormous problems for the other. Further, the APA categories are more comprehensive and—because they are extensively litigated—can be applied to novel cases with less uncertainty. Accordingly, it makes sense for the PRA to adopt the APA’s standard for bypassing public comment. Further, the PRA should be aligned with the APA by not having a 180-day limit on approvals of such information collections. Instead, they should be able to be approved for up to the regular limit for approvals (see Reform 1).

Reform 5: Expand OMB’s authority to waive irrelevant requirements for specified categories of information collections.

Reform should not stop at simply codifying the APA exceptions to public comment requirements. An amended PRA should empower OMB to make rules specifying categories of information collections that can be exempt from OMB review at all (as de minimis changes already are) or exempt from public comment requirements (as non-substantive changes already are). In the past, OMB has tried to act through guidance, but rulemaking authority would allow OMB to go further. Unlike the status quo, a clear grant of affirmative authority to make such rules would give OMB the ability to waive requirements when the benefits simply do not justify the costs.

OIRA should begin with a few high-value categories that would be difficult to specify in statutory text, or inadvisable to freeze in place permanently, but greatly reduce unnecessary burdens. These include:

1. Adequate alternatives to public comment. When an agency can show that it will receive (or has received) public feedback that is as valuable as—or more valuable than—the feedback that it would likely receive through a Federal Register notice and 30-day public comment period, the public comment period should be waivable. This waiver generalizes the specific carveout in statute presented in Reform 2. And it is a standard that is not particularly onerous to meet, given the previously-discussed evidence that most information collections receive zero public comments. Agencies should be provided clarity on ways to meet the standard. For example, OMB has already developed guidance on how agencies can best collect public input to improve high-impact customer-facing services.^[134] An agency that conducts affirmative public outreach among those who will be impacted by an information collection will likely receive more useful feedback than it would receive in response to a Federal Register posting.^[135] But for the small fraction of collections likely to get high-quality public input, like certain statistical surveys, public comment (or a particularly rigorous alternative) may be appropriate. An advantage of this approach is that agencies would be able to receive up-front clarity that they can skip public comment by hewing to a well-established and superior public input process for certain categories of collections, giving them the certainty and speed necessary for effective program implementation.
   
2. Voluntary collections. Many PRA reformers have long advocated for excluding voluntary information collections (that is, information collections that there is no obligation to respond to) from the scope of the PRA.^[136] However, as Shapiro noted, there are two key difficulties with exempting all voluntary information collections from the PRA. First, voluntary collections are collecting statistical information in the majority of cases, the category of information collections for which OIRA review is most useful.^[137] Second, many information collections that are voluntary are perceived by the public as mandatory, and therefore pose most of the same concerns about burden that mandatory collections do.^[138] Rather than attempting to handle these difficult-to-define issues in statute, Congress should empower OIRA to create a category of information collections that are clearly voluntary and do not require review of statistical methodology, exempting them from the PRA’s requirements using a fast-track review process. For example, a waiver could allow federal agencies to more easily solicit voluntary information collections from their state counterparts regarding what technical assistance the states would most benefit from in certain programs. Agencies would be able to move forward with such collections in no more than five business days; if OIRA had concerns that the collection is not clearly voluntary or employs a statistical methodology that requires some review, the agency could simply proceed through a (now much speedier) normal review process, where those concerns can be addressed.
   
3. Studies already approved by recognized scientific bodies. In some cases, agencies are collecting information pursuant to recommendations (or directions) approved (or required) by scientific bodies. In those cases, the requirements of the PRA likely add little value.^[139] Accordingly, such information collections should be subject to fast-track approval without public comment (though agencies may seek public input voluntarily in certain cases). However, what should count as a recognized scientific body—the National Academies of Sciences, Engineering, and Medicine and university-affiliated research bodies, for example—should be established by regulation and updated over time, to adjust based on experience in implementation and changes in the world.
   

Reform 6: Affirm and codify OMB’s existing practice on PRA flexibilities, such as generic information collections.

As noted previously, the PRA has been plagued by folk wisdom at agencies that interprets the PRA more strictly than the text demands.^[140] Granting OMB explicit authority to waive PRA requirements for categories of information collections is critical to more effectively addressing this problem. In particular, it solves for agency counsels’ tendency to embrace rigid and conservative interpretations of the PRA, while also giving OMB the tools to address bureaucratic anxiety about potential non-compliance by creating rules that are legally binding. It also addresses the incentives of agency PRA officers, as the knowledge that OMB can waive requirements will create pressure from political leadership to move more quickly on policies that require information collections. This will also ensure that OMB will receive an appropriate amount of pressure to create such categories, while still leaving room for OMB to push back (with potential elevation and resolution by the President), in appropriate cases. Additionally, it creates incentives within agencies to simplify excessive layers of review of information collections, which are often the source of more delay than the public comment process itself.

Although an affirmative grant of authority to OMB is useful, a belt-and-suspenders approach is appropriate to preserve existing PRA flexibilities. While many PRA flexibilities (collected in the 2016 Shelanski memorandum^[141]) were implemented in the years immediately following the enactment of the PRA in 1980, their consistency with the PRA could be challenged. Particularly in a Loper Bright era,^[142] such challenges would be serious, and possibly devastating to the operation of good government. To start, amendments to the APA should preserve OMB’s statutory authority to promulgate any rules, regulations, or procedures necessary to exercise PRA authorities,^[143] not just to waive certain requirements (as discussed in Reform 5). But further, the statute should ideally cite current flexibilities as examples of valid uses of this statutory authority, to safeguard them against legal challenge.^[144]

The persons requirement

Some have alleged that the PRA’s carve-out for information collections involving fewer than 10 persons sweeps too narrowly. There are two primary reasons, however, that—if the other reforms detailed above are adopted—raising the information collection persons requirement would do little good, and much harm.

First, the previously-discussed reforms would eliminate PRA-imposed delays associated with all but the most significant information collections.^[145] The fear that a relatively unimportant collection from, say, 50 people would be unreasonably delayed or impeded is difficult to square with the reforms outlined above. The ordinary problem in cases where information collections are addressed to only a small number of people is that such collections should be exempt because they relate to usability testing, are minor in nature, should be approved quickly due to exigent circumstances, etc. The previously-discussed reforms—such as eliminating or consolidating public comment periods, creating authority to waive the PRA’s requirements for categories of information collections that should be exempt from those requirements, or aligning emergency approvals with the APA—adequately address these concerns.

Second, there are downsides to such a change, as the PRA’s definition of a “person” includes not just individuals but also businesses, tribes, and states.^[146] If the threshold were even just raised to 50 persons, it should be clear that demanding information from, for example, all 50 states, every major AI company in the United States, or the 50 largest universities in the United States could be an extraordinarily important and burdensome exercise. The problem cannot be solved simply by having different thresholds for individuals and other entities, either; the same hypothetical would work for governors, CEOs of major AI companies, or presidents of universities. There is good reason to have OMB ensure adequate review of such requests by relevant agencies and components of the Executive Office of the President, either to have OMB ensure that no unnecessary burdens are being imposed on the public or for consistency with the administration’s other policies. Conversely, some information collections involving more than 10 but fewer than 50 persons may be inputs to the most influential information collections that statistical agencies engage in, such as cognitive lab testing of economic data surveys. As discussed previously, OMB plays an important role in improving the quality of information collected and ensuring interagency consistency across statistical products in such cases.

Considerations beyond statutory amendments

There are additional considerations regarding PRA reform that, while important, would not best be addressed through amending the PRA.

1. Improving OIRA capacities

The original ambition of the PRA (and the Federal Reports Act before it) was centered on eliminating duplicative information collections across agencies. But to a great extent, achieving this currently requires long-tenured OIRA desk officers to accumulate personal knowledge of the information collected across agencies and sub-components of agencies. Technology adds little to this process, despite the fact that the back-end system that OIRA uses to track information collections—the RISC/OIRA Consolidated Information System (ROCIS)^[147]—has rich data extending back to the George H.W. Bush administration. ROCIS is antiquated, but it could be harnessed to substantially improve the efficacy of OIRA’s desk officers.

ACUS called for modernization of ROCIS many years ago.^[148] With advancements in artificial intelligence, a modernization is even more important, as artificial intelligence tools could help OIRA desk officers query the ROCIS database more effectively even than natural language search capabilities would allow (which, it should be noted, ROCIS also currently lacks). Any previous initiatives to improve ROCIS by integrating new features of this type that were underway may now face new challenges.^[149] Congressional committees of jurisdiction—such as the Senate Homeland Security and Government Affairs Subcommittee on Border Management, Federal Workforce and Regulatory Affairs and the House Judiciary Subcommittee on the Administrative State, Regulatory Reform, and Antitrust—should engage with OIRA to ensure that ongoing efforts are made to improve the usefulness of the ROCIS system for users inside and outside of government.

Questions of OIRA resources and capacities to manage workload may also be relevant.^[150] As noted previously, one problem in the status quo that these reforms are meant to address are delays imposed by OIRA reviews of information collections. Certainly, these suggested PRA reforms would radically reduce the number of information collections that OIRA would process each year, addressing such concerns. However, other recent initiatives may be substantially increasing OIRA’s workload.^[151] Appropriators and committees of jurisdiction should engage with OIRA to ensure that it has adequate staffing and other resources to fulfill its mission.

2. Improving agency and government-wide capacities

Staff in agencies whose primary mission does not focus on collecting information often lack basic training in user-focused design, usability testing, and other skills critical for improving many information collections. (By contrast, agencies developing information collections for statistical purposes—for example, at the Bureau of Labor Statistics or Bureau of Economic analysis—generally have staff with extensive training in effective statistical data collection.) Agencies should ensure that they provide staff with the training and other resources necessary to develop these skills, which will improve their information collections over time.

It may also be prudent to have a component of OMB’s management division or OIRA provide additional government-wide support for improved user-focused design and usability testing. Some agencies have such a large volume of regular information collections that they should have staff with those capacities within the agency. The combination of subject-matter expertise and more general design and testing skills allow those staff to be particularly effective. But for some agencies, which do not frequently collect large amounts of information from the public, access to a centralized body may be the only prudent way of benefiting from this kind of expertise. And such a centralized expert body—wherever housed—would also help OMB with its distinct tasks, like ensuring effective coordination across agencies.

3. Revitalizing the Council of Agency Paperwork Reduction Act Officials

The Council of Agency Paperwork Reduction Act Officials (confusingly known as CAPRA rather than CAPRAO) is a group of agency PRA clearance officers that meets monthly to discuss issues related to the PRA.^[152] Agencies obviously benefit from regular engagement on PRA issues through CAPRA.^[153] And OIRA has regularly presented to the group on PRA flexibilities, such as generic clearances.^[154] However, it is likely that CAPRA could be better harnessed to help push agency PRA officials to better understand existing PRA flexibility, and new flexibilities enacted into law. CAPRA could designate certain meetings to include both those responsible for PRA compliance and program officials, ensuring that accurate information is not distorted in transmission or siloed within only some agencies. OIRA’s presence at such meetings could be particularly helpful to dispel myths about limited PRA flexibilities. At the same time, some CAPRA meetings should likely take place without OIRA’s presence and focus on government-wide problems that agencies are running into with OIRA during the PRA process. That information, after being aggregated and anonymized, could help create a feedback loop driving improvements to OIRA’s process; it would also avoid any concerns among agencies that critiquing OIRA processes would undermine long-standing relationships among agency and OMB career staff.

4. Information collections implementing federal programs

Often, state, territorial, tribal, or local governments design and implement information collections that relate to federally-sponsored or jointly-administered programs.^[155] However, agencies have very limited ability to improve the information collections that are designed and implemented by these entities. There is a serious argument to be made that agencies should have the authority—although not the obligation, given the sheer number of such information collections—to review and replace state information collections relating to public benefits or services with less burdensome alternatives, when they are funded by the federal government. This could open up enormous opportunities for the federal government to work with state, territorial, tribal, and local governments to reduce the most important burdens imposed by their collections. Many entities may not even be aware of the flaws in their information collections, or have the capacity to address them.

However, moves to substantially alter the balance of federalism are likely to be far more controversial at this time than the other reforms outlined in this policy brief. Further, erecting guardrails to ensure that this authority is not misused to impose serious policy changes on these entities—rather than simply burden reduction—would be difficult. Statutes have moved authority to these non-federal entities, wisely or not, so that those entities could shape the implementation of each program. Accordingly, Congress should study this issue without codifying any changes to the PRA at this time.

That being said, it would be prudent for academic researchers to partner with individual agencies and states to learn if voluntary relationships of this type would be viable through small-scale experiments. For example, Governor Shapiro recently led an initiative to simplify the process of seeking a permit, license, or certificate in Pennsylvania^[156]; it would be natural to extend that effort to reducing unnecessary burdens in areas like benefits and grants. Voluntary agreements between federal agencies and states may enable a great deal of progress, even in the absence of federal statutory authority to compel changes.

5. Better capturing burdens

OMB has clarified that the accounting statement of an information collection should not merely include the hours it takes to fill out a form. An accurate accounting should include:^[157]

• “burdens such as time spent gathering records and documentation needed to prove eligibility, travel time associated with developing and submitting the collection, or even time waiting to speak with agency personnel”;
• “psychological costs that certain information collections impose on individuals, such as the cognitive load, discomfort, stress, or anxiety a respondent may experience as a result of attempting to comply with a specific aspect of an information collection”;
• “learning costs, which include the time and effort expended by a respondent to discover and determine the applicability of an information collection to their particular circumstances, as well as any research necessary for the respondent to understand how to comply with any program participation requirements beyond reading a form’s instructions”; and
• burdens that occurs “even when Federally sponsored information collections are implemented by state, territorial, tribal, or local governments.”

However, agencies still focus overwhelmingly on burden hours as the exclusive measure of burden, and have generally not attempted to systematically measure other sources of burden, especially psychological burdens. This is true even as research increasingly suggests that these other burdens can pose especially important costs to vulnerable individuals and small businesses when interacting with the government, limiting eligible individuals from accessing public services that they want and would benefit from.^[158]

It might be tempting to call for codification of the requirement that agencies more accurately capture the burdens of their information collections in accounting statements, with new teeth, beyond the PRA’s current requirement for “a specific, objectively supported estimate of the burden” of the collection.^[159] However, this aspiration is already clear to agencies.^[160] In this context, statutory codification is likely to do little good, and may do harm. Often, agencies will lack the information that they should ideally do a better job estimating; for example, the psychological harms veterans may experience while filling out certain forms regarding traumatic events to qualify for benefits programs. In these cases, a binding statutory requirement would only add to the existing problems of burden detailed previously: Agencies abandoning some valuable information collections, or delaying other information collections to complete the difficult analysis and often pulling agency staff away from higher-value tasks in the process.

In this context, it is likely best for agencies and OIRA to expand efforts and work together to slowly “level up” their analysis of the burdens of information collections, better estimating the full burden of each collection over time. The journey mapping and analysis needed to do so effectively can also be enormously helpful in identifying unnecessary burdens that can be reduced. As a result, agencies and OIRA already have incentives to produce these types of journey maps and to push for more complete analysis of the burdens of an information collection on the public. A model of effective collaboration between agencies and OIRA to account for psychological and learning costs can be found in the case studies of agency burden reduction efforts discussed in OIRA recent reports.^[161] That approach will yield better results than a sticks-without-carrots PRA amendment.