Blueprint to Fix Reporting of Federal Spending

Tracking the trillions our government spends each and every year has always been difficult, more difficult than it should be. But last year, the COVID-19 pandemic triggered a huge spending surge from the federal government. The extra spending brought with it a brighter spotlight of attention. And that spotlight has revealed stark deficiencies—failures in our reporting systems for federal awards that keep the public in the dark.

In the case of the COVID-19 relief spending, our lack of information is partly due to the fact that reporting and oversight provisions envisioned by Congress for pandemic spending were never fully realized. But in a much larger sense, the fault lies with major, preexisting deficiencies.

Linda Miller, who served a year as deputy executive director of the Pandemic Response Accountability Committee, recently shared her perspective on spending oversight problems in an op-ed. “Nearly all aspects of collecting and using data in the government are broken,” she concluded.

She isn’t wrong. Many of the questions being asked about the COVID-19 relief spending are questions that also regularly get asked about other federal spending. What did recipients use the money for? Were any communities or businesses missed? How many jobs did it save or impact? Where did it all go?

All reasonable questions. Unfortunately, all too often the answer is, “we don’t know.”

The good news is that there are clear fixes and improvements agencies can implement to ensure more meaningful accountability for all federal spending going forward. The first step, however, is to identify the problems plaguing our reporting system.

The federal government has major data quality problems with the information it currently collects—errors in the data, missing data, and data that doesn’t tell us anything. These are data points agencies have been collecting and reporting for years, and they still don’t work right.

But even if agencies fixed those problems, we still wouldn’t have the answers we want because the reporting system simply isn’t collecting all the information we need. Entire agencies and types of spending are missing from the current tracking system. For awards that are included in the system, key data points are missing that would allow us to understand how equitable government assistance programs are and whether any industry sectors have been overlooked. It is nearly impossible to adequately review whether emergency government spending—or any federal spending—accomplishes its intended goals.

So what do we need to do to address this? A large-scale revitalization of federal spending data should take three steps:

1. Fix current reporting
2. Fill reporting gaps
3. Track new data points

It has been more than a dozen years since USASpending.gov was launched as the primary disclosure portal for federal spending. In that time, the site has gone through numerous iterations, adding features and functions and expanding data. Congress has updated the requirements for the site several times with new laws like the Government Funding Transparency Act of 2008 and the Digital Accountability and Transparency Act of 2014. Yet, through all the changes, several fundamental elements in the reporting system for federal spending remain essentially broken. Below is a list of the most important fixes we need.

Fix #1: Require Meaningful Award Descriptions

Award descriptions are one of the most important fields in the spending data that agencies report. As the name implies, the award description should convey some summary of the intended use or the purpose for which funds are provided. This information should help the public answer one of the most basic questions: “What was the money used for?”

All too often, however, the information reported in the award description is of little or no use. Agencies sometimes fill the award description field with a program’s name, the authorizing legislation, or a simplified program mission statement over and over, providing no information about specific awards. Often, they use non-specific terms such as “fuel” or “meals” to complete the description field. Contract award descriptions sometimes consist of nothing more than strings of procurement codes that mean nothing to non-governmental users.

Why are we getting such insufficient responses to such an important field? It may stem from the lack of any guidelines or minimum requirements for what kind of information must be included. Currently, any information in the award description, no matter how vague or uninformative, counts as a completed entry.

POGO recommends establishing clear guidelines specifying what constitutes an acceptable award description and what types of entries are prohibited. Such guidelines should set standards for length and expected content. A word count minimum and maximum would prevent agencies from providing a couple of uninformative words or dropping in whole pages for users to hunt through for key data. For guidance on content, we suggest agencies be advised to ensure the award description provides a summary of the award or transaction that includes its purpose and scope.

To provide context for the funds being spent, the award description should also include any relevant quantities associated with the award, such as the amount of materials, length of time covered, or number of payments made. Including quantities associated with aggregated reports of assistance to individuals and households (for example, the number of payments, largest payment, smallest payment, and more) can provide important context while still protecting individuals’ privacy.

Fix #2: Address Inaccuracies in Subaward Reporting

Often the initial recipients of federal awards don’t hold on to the funding, or at least not all of it. Information on subawards—funds received from the federal government but then passed along through contracts, grants, direct assistance, or some other process—is critical to understanding where federal funds ultimately wind up and who spends the money. For example, state agencies are often responsible for distributing federal assistance to counties or towns. Local governments may hire contractors to help accomplish the goals of their project. Those contractors may, in turn, bring on subcontractors to perform portions of the work. Sometimes funds can move through several layers before the money is used for its intended purpose. For years, the federal government has been requiring information on only the first layer of subawards: Primary recipients are required to report only those subawards they themselves make with federal funds.

Even this limited subaward data, however, has remained almost wholly unreliable. Sometimes, subawards add up to far more than the prime awards given out. According to data on USASpending.gov, for example, in fiscal year 2020 the Department of Health and Human Services spent approximately $1.7 billion through the Community Services Block Grant program. However, subgrants reported for the same program total $2.5 billion. No one added money to the program in the middle of the distribution process. That $800 million difference is an error: the subaward data has duplicate entries, counting some subawards multiple times. In other cases, the subaward data is apparently under-reported or missing entirely. For example, in fiscal year 2020 the Department of Agriculture spent $83 billion through the Supplemental Nutrition Assistance Program (SNAP). However, USASpending.gov lists no subawards at all.

The SNAP data shows why accurate reporting is so important. We know these funds were awarded to states and counties, then further distributed to millions of households. But without reliable subaward data there is no way to track where the money went within the states. While individual recipients cannot be listed for privacy reasons, tracking the subawards—and posting transactions detailing the aggregate amounts spent by zip code and by month—would allow better evaluations of spending patterns and encourage research into communities that may be getting less than they should.

These are just singular examples, but the problems within the subaward reporting are systemic. The end result is that none of the subaward data can be trusted.

POGO recommends the entire subaward reporting system receive a comprehensive audit either from the Government Accountability Office or the inspector general of the General Services Administration (as the GSA manages the subaward reporting system) to identify the full extent of the errors and their causes. That audit should be followed by an immediate overhaul of the system to ensure complete and accurate reporting. Deadlines should be established for both the audit and the overhaul to ensure these issues are addressed quickly and this important fix is not delayed.

Fix # 3: Require Corporate Identification Codes in Public Reporting

Properly tracking federal spending requires a clear, consistent identification code for companies receiving these funds. Names listed for a company can vary depending on data entry, and companies in different states can have remarkably similar names. A corporate identifier is an essential component of the spending data.

Currently, we rely on the Data Universal Numbering System (DUNS), run by the private company Dun & Bradstreet. There have been several long-standing problems with DUNS codes, including their high cost to government agencies and the fact that companies can accrue multiple DUNS numbers, creating confusion and data fragmentation. But despite their clear drawbacks, the most significant problem with DUNS codes is that, for a large portion of federal spending, agencies are either failing to collect the codes or aren’t disclosing them to the public. The problem pre-dates the COVID-19 relief spending. Most disaster loans in the years preceding the pandemic have no DUNS number listed for the recipient company. But the pandemic spending surge exacerbated the problem as major COVID programs, such as the Paycheck Protection Program (PPP) and Disaster Assistance Loans, left DUNS numbers unreported for all recipients. These programs cover millions of awards.

Our inability to easily and reliably track companies in federal award data limits our ability to detect potential fraud and waste, especially across programs. Just last year, a POGO investigation revealed that a company called Atlantic Diving Supply (ADS) received a multimillion-dollar PPP loan. At the time it received this loan—funding intended to help companies struggling to meet payroll costs during the economic uncertainty of the pandemic—ADS was also receiving more than $1 billion in federal contracts, raising questions about the company’s need for assistance.

POGO recommends agencies be required to collect and report DUNS numbers for all companies receiving federal awards or subawards until such time as a government corporate identifier system is in place. Additionally, the administration should increase efforts to establish a government corporate identifier and incorporate it into the reporting system for federal awards as soon as possible.

Fix # 4: Codify and Expand the Use of Event Codes to Better Track Related Spending

The first step in evaluating any spending is being able to easily pull up related awards. Often, this means sorting awards by agency, program, location, or recipient. But some major events, most often disasters, require spending across multiple agencies and programs. Event codes are used to track such cross-cutting federal spending. The current system has two such codes. Neither works fully.

First, we have the National Interest Action (NIA) code. This four-character code has been used for years to track contract spending across multiple agencies in response to major disasters and some military operations. For example, unique NIA codes were assigned to track spending in response to Hurricane Harvey (H17H), spending for operations in Iraq and Syria (O14S), and spending for the coronavirus pandemic (P20C). But the NIA code system has serious limitations that keep it from being an effective tool for tracking event spending. First and foremost, the code is only used to track contract awards; it is not reported for assistance awards. Second, the authorization of a new NIA code or discontinuation of a code already in use depends on a fairly arbitrary process overseen by a few procurement agencies behind closed doors.

Our second event code, the Disaster Emergency Funding Code (DEFC), was recently introduced to the reporting system to allow users to better track COVID-19 relief spending. The code identifies the legislative source of funds for an award, specifically whether the funds were authorized by major COVID response bills. While the DEFC approach of tracking the source of funds is helpful, it—like the NIA approach—can provide an incomplete picture of spending in response to a disaster. The DEFC code misses agency spending in response to a crisis conducted with existing funds, which can be considerable.

POGO has found significant gaps in the application of the DEFC codes for COVID spending. None of the billions spent through the Coronavirus Food Assistance Program have been tagged with the DEFC codes indicating the program is spending COVID funds. Additionally, in comparing USASpending.gov data and COVID disaster assistance loan data released by the Small Business Administration, POGO has found almost 50,000 disaster loan transactions on USASpending.gov that are missing COVID-related codes in the DEFC field.

POGO recommends the administration fix both event codes going forward to ensure better tracking of federal spending in response to natural disasters, health crises, or other unique events. The NIA code should apply to both contracts and assistance, and there should be clear criteria for initiating and discontinuing a code. The use of the DEFC code should be audited by the Government Accountability Office to identify the gaps in its use, and updated guidance should be provided to agencies to ensure better application of the codes.

Fixing the broken reporting of specific data elements is an important step to getting the information we need to hold federal spending accountable. But it won’t address major loopholes built into the reporting system for federal spending: what the government has chosen to leave out. These gaps are not limited to individual data fields; they are larger issues. And while agencies already have this spending information, it is not included in the reporting system.

Gap #1: We Need Reporting On Tax Expenditures

Often, as a part of efforts to accomplish an economic goal like job creation or support for an industry sector, the government implements corporate tax breaks. For example, in response to the pandemic Congress authorized tax expenditures for employee retention, as well as for expanded sick and family leave. As a vehicle for assistance, tax expenditures are seen by some as a cleaner and more efficient approach, since they leave owed tax dollars with companies rather than collecting the funds and then having agencies redistribute them. But the lack of transparency for key tax expenditure programs is the single biggest accountability loophole in federal spending.

Currently, the federal government doesn’t report details on companies that claim these tax breaks. This means that the public has no idea who these tax expenditures benefit; whether they are equitably shared with small businesses, minority owned businesses, or women owned business; or whether the companies that benefit from the tax expenditures use the funds to maintain employment levels, to create new jobs, or to deliver the benefits to owners and executives in the forms of dividends and bonuses.

POGO recommends that agencies be required to track and publicly report individual claims made on major tax expenditure programs designed to achieve a broader public benefit. The reporting should include a similar level of detail to that required for federal spending through contract and assistance awards—taxpayers should be able to learn recipient names, parent companies, location, industry sector, and other critical information about those who receive these benefits.

Gap #2: We Need Reporting From All Agencies

Currently, several agencies are simply not required to report their awards through the system, creating a significant gap in public oversight around where taxpayer funds are spent.

The Federal Reserve is the best example of this loophole. In response to the pandemic, the Federal Reserve was tasked with overseeing several programs geared toward ensuring the stability of our financial markets. They made loans, purchased bonds, and extended credit. But none of these transactions were reported through USASpending.gov, because the Federal Reserve does not report any of its spending through that system. And while the Federal Reserve did post transaction-specific data for its COVID-19 programs—an unusual level of transparency for the agency—the data it provided did not include the same level of detail that other agencies regularly provide for their awards.

There are a few other agencies that do not report their contracts or assistance awards through USASpending.gov. The U.S. Postal Service has determined that it is exempt from the legislation that established and updated USASpending.gov requirements, so the public has never seen its contracts in that system. The Government Accountability Office, the investigative office of Congress, voluntarily reports its awards into USASpending.gov, but there is no requirement it do so because the reporting requirements are focused on the executive branch. This loophole for the legislative branch keeps several other entities connected to Congress—such as the Capital Police, the Library of Congress, and the Congressional Research Service—from reporting their awards into USASpending.gov. While these offices may not oversee emergency spending, full accountability for federal spending would require that we close these loopholes.

POGO recommends that Congress require the Federal Reserve, the U.S. Postal Service, and other missing agencies to report contract and assistance awards through USASpending.gov.

Gap # 3: We Need Access to Award Documents

Sometimes a string of data fields, even highly informative data fields, cannot provide the level of detail needed to fully understand the purpose and scope of a particular award. In those instances, the best piece of information would be the actual contract document or award record. These documents detail scope of work, services and products, roles of various entities, and more. But federal agencies don’t proactively disclose contracts and award records. Instead, anyone interested in reviewing them must file time-consuming Freedom of Information Act requests, delaying spending oversight.

Both Congress and federal agencies have explored the idea of proactively posting award records, particularly contracts, but the proposal has been derailed each time. This is often because of either concerns about cost or objections from contractors claiming the process would expose confidential business information. But at least 33 states have successfully posted contracts or detailed summaries of contracts for several years now. Indiana’s Contracts Portal is a model that not only allows users to see PDFs of contracts but also includes grant agreements, leases, and memoranda of understanding.

POGO recommends that the federal government establish requirements to proactively post contract and other award records or summaries of those records linked to the award data in USASpending.gov.

In addition to fixing broken reporting and closing reporting loopholes, an overhaul of our tracking and reporting system should also push agencies to start collecting new data points about federal awards. There are several useful data points for which the government simply isn’t collecting the information. Congress and the administration should work together to quickly address these reporting needs and require the following data points to be gathered for all awards. Without this data, we can’t answer standard, basic questions about government spending.

New Data Point #1: How Are Federal Awards Affecting Jobs?

A common inquiry around federal spending is its impact on employment or small businesses. In fact, some federal programs, like the recent Paycheck Protection Program, are targeted to support job creation and retention. Even so, it is almost impossible to answer whether these job support programs work, because we don’t currently track any employment numbers for recipients of federal awards. There is no way to determine if a recipient of a federal award has 1 employee or 100 or even 10,000.

With the exception of individuals and households, every recipient of a federal award or subaward should report the number of full-time equivalent people they employ. This data would provide a vital piece of information to evaluate the reach of federal spending in the economy.

We know gathering this information is feasible—the Paycheck Protection Program collected employee numbers as part of its application process and released the data separate from USASpending.gov. While the initial reporting included data quality problems, as there will be around almost any new data element, the process proved that employment data can be tracked and disclosed for millions of federal recipients.

POGO recommends the federal government establish employment reporting for all federal awards and include the data on USASpending.gov.

New Data Point #2: How Are Federal Funds Distributed Among Different Demographic Groups?

Recent questions around the equity of government relief programs have highlighted the fact that agencies do not collect any information on the ethnicity, gender, veteran status, or other demographic details about the owners of businesses receiving federal assistance awards.

For federal contracts, agencies track several data fields related to the owners of businesses. Reporting consists of simple yes or no fields that track whether a business is minority owned, woman owned, or veteran owned. There are also more specific fields to track the race or ethnicity of the owner.

Contract reporting also includes other fields that track important aspects of the type of business and communities served, including opportunities to report on Historically Underutilized Business Zone Firms, Labor Surplus Firms, 8(a) Program participants, Small Disadvantaged Businesses, and more.

POGO recommends the government establish assistance award reporting requirements for demographic data related to business owners and disadvantaged business data categories. These contract data elements can serve as a guide, and to the extent practicable, the demographic data collected for contracts and assistance awards should align. The process should also serve as an opportunity to evaluate whether the contract reporting in this area is sufficient or needs to be updated.

New Data Point #3: How Are Federal Awards Distributed Among Different Industries?

Despite genuine interest in knowing which industry sectors benefit from a particular federal assistance program and which are left out, agencies do not report industry sector data for companies receiving assistance. That data is collected for contract awards, but agencies do not collect industry sector data for loans, grants, direct payments, and other assistance awards.

There are obvious reasons why this information should be collected and made available. Some assistance programs are clearly focused on particular industry sectors, for example Head Start money going to education, mass transit grants going to transportation, or crop insurance payments going to agriculture. Industry sector data would help confirm that the assistance did go to businesses in those sectors, and it would help analysts identify which subsectors most benefited. Other programs, like disaster assistance loans, offer aid to a wide variety of businesses. For those programs, industry sector data is critical to determining whether assistance is disproportionately benefitting some sectors and if certain industries are missing out.

POGO recommends the government require agencies to collect and report the North American Industry Classification System code for each business that receives federal assistance awards.

New Data Point #4: How Efficient Are Federal Awards?

The basic question of “how much” is a common one that can be difficult to answer for federal awards. How much water was provided? How many meals? How many miles were paved? How many houses inspected, or patients tested? Under the current reporting system, the award description field is the only space in the spending data for information on quantity. But, as we noted in the section on broken data, it is rare that agencies include clear quantities in their descriptions.

A separate field to track the primary quantity or measurement for each award would give users an important tool for comparing the impact of similar awards. Given the wide scope of government awards, no one quantity would work for all of them. A system would have to be established and refined over time to identify the most common quantifiers for different agencies.

While the quantifier would change between widely different awards, for similar awards the quantity being measured would likely be consistent. Adding a quantity field to USASpending.gov would enable greater evaluation of the return or scope achieved for each award. Comparisons of similar awards could reveal cost savings or ways to increase the effectiveness of federal awards.

POGO recommends the government establish a new primary quantity reporting field that would require agencies to report the main measurement for each award transaction. The system would be challenging, but it could be phased in starting with the most common measurements and then updated periodically to add new options.

New Data Point #5: How Many People Benefit From Federal Awards?

Another common, and foundational, question about government awards that goes unanswered because the government doesn’t collect the data is, “how many people did this help?”

In their reporting of awards, agencies should offer an estimate of how many people benefit from the product or service being funded. If the award covers schoolbooks, how many students are getting them? If the award provides job training, how many people will receive it?

A metric on the number of people benefiting from a federal award would provide another important data point to evaluate the impact of federal spending and compare awards. The reality is that sometimes the important measure isn’t how much of a product the government provides but how many people benefit. Knowing that a half-mile bridge was repaired might not sound like much. But knowing that millions of people depend on the bridge to commute to and from work can change our perspective on the federal award that funded the repair. Similarly, tracking this data would also mean that big contracts and awards that benefit very few people, such as the famous bridge to nowhere, can be better identified and brought into question.

For some awards, the answer might be easy. Others, like defense contracts that help protect the entire country or scientific research grants that could help millions or could lead nowhere, would be more challenging to answer. Guidance would have to be carefully developed to give agencies the proper tools to define and consistently estimate the number of people benefiting from each award and perhaps to exempt certain types of awards.

POGO recommends the federal government establish a new spending data field to estimate the number of people benefiting from the award.

New Data Point #6: Who Is Lending the Funds for Federal Awards?

Loans are a common vehicle for financial assistance from the federal government. Sometimes loans are direct from federal agencies to businesses or households. But often loans are federally guaranteed, which means an agency approves loans from private lenders to qualified borrowers and agrees to cover the loan amount should the borrower be unable to repay the funds.

These private lenders deserve as much oversight and accountability in these programs as the borrowers. Currently, assistance reporting does not include any data on lenders. The absence of lender information prevents evaluation of how lenders are implementing government assistance programs. Are some lenders missing certain communities or business types? Are lenders in urban areas outpacing lenders in rural areas in providing access to government funds to their clients? When government assistance programs use guaranteed loans, data on lenders is vital to evaluating the effectiveness of those programs.

POGO recommends the government require lender information be reported for all loan transactions involving private lenders. Reported information should include the name of the lender, location of lender, parent company, NAICS code, number of employees, and any other data determined necessary to accurately track lender activities related to government loan programs.

The government spends trillions each year, and taxpayers should know that spending will be reported completely, accurately, and with enough detail that we can rigorously evaluate it. Our current system remains broken and incomplete.

Now is a critical time for Congress and the administration to act to repair these flaws. Recovery from the pandemic is ongoing. Some new programs started by the American Rescue Plan Act of 2021 don’t seem to have begun distributing their funds: this spending could still benefit from improved accountability. And beyond the pandemic, plenty of ongoing spending needs better accountability. The steps recommended by POGO will fix many of the biggest problems and establish more reliable transparency for federal awards. In the long run, this improved oversight will affect how Congress and federal agencies spend money, hopefully making federal awards more effective and equitable.

POGO urges the administration to work with agencies to immediately address as many of these issues as possible. Many of the fixes and improvements can be implemented without statutory requirements from Congress. Any data improvements implemented quickly could significantly improve the evaluation process and help identify problematic spending.

POGO also urges Congress to update the statutory reporting requirements around federal spending. While agencies can move quickly under their own authority, congressional action can ensure that all future administrations provide robust and effective reporting on federal spending. By establishing improved reporting requirements, Congress can ensure that the lessons learned from the difficulties in tracking COVID relief spending aren’t wasted.