The COVID-19 pandemic has exposed many fragilities in the United States’ healthcare, economic, and social infrastructure. Food bank lines have extended for miles. Low wage earners, women, and workers of color—demographic groups most common in the ‘essential workers’ category—have experienced employment disruption due to quarantine orders. Hospital capacity has been pushed to its outermost limits and, almost simultaneously, pandemic shutdowns have purportedly caused overdose deaths to swell. This precarious state of affairs is not helped by lackluster administrative data. Fractured health data sources and voluntary data submission have been cited as reasons for what has been, quite simply, a sub-par pandemic response. It might be argued that the magnitude of the COVID crisis has simply overtaxed our health-data infrastructure. However, this is not the only area where the United States’ administrative data collection is sparse.

The FBI’s Uniform Crime Reports show that between 2004 and 2015, there were over 73,000 reported hate crimes in the United States and, in particular, there was a spike in hate-based incidents during the 2016 election cycle. However, The Bureau of Justice Statistics reported that the true figures are more than 50 times greater, suggesting that Americans are subjected to over 250,000 hate-based victimizations every year. This is largely because the Hate Crimes Statistics Act of 1990 does not require states and localities to turn this data over to the federal government. This has led to decreases in reporting and, in reporting districts, unreliable data. For instance, most jurisdictions report zero hate crimes—an impossible statistic.

The Social Security Administration (SSA) shares a similar records issue. The Government Accountability Office explains that even though the Death Masterfile (DMF)—the federal government’s database used to regulate benefits and detect fraud—is the authority on the deceased used essentially by all financial institutions, the SSA does not verify certain death reports or even record others. The agency is primarily focused on ensuring that it does not make benefit payments to deceased Social Security beneficiaries, which means they only verify death reports received for program beneficiaries. Even then, verification is only for those reports received from sources considered less accurate. There is also a legal interpretation issue at fault here. Specifically, Section 205r of the Social Security Act has been read in a way that constrains state data disclosure to banks and many federal agencies. The result? People like Barbara Murphy have their lives upended due to bureaucratic errors. Their bank accounts are frozen, credit cards canceled—as much happened to Murphy—and in severe cases, they are arrested for stealing their own identities. While this is statistically uncommon, the approximately 12,000 people who experience cyberdeath every year would implore the SSA to address the issue. Coupled with administrative data woes in education, welfare, and other areas, this information indicates a couple of things. That is, not only is the COVID-19 data issue non-unique, but that it is a symptom of a larger issue. The United States’ administrative data infrastructure—an infrastructure all the more necessary with new federal programs and initiatives—is hobbled, diluted, and political. This also begs the question: what is to be done?

Currently, multiple legal reforms have been proposed in these specific policy areas. Sen. Blumenthal’s NO HATE Act, for example, would help states and localities shift to the National-Incident Based Reporting System with a series of grant programs to state/local agencies. However, the law’s text suggests the disbursement of these grants are subject to the Attorney General’s discretion, meaning that the politicization issues with the country’s data management are still cause for concern. Similar legislation for the DMF has been proposed in the form of the “Stopping Improper Payments to Deceased People Act.” Grants with reporting requirements were included in both bills has also been employed as part of the federal government’s COVID-19 relief. Altogether, the legislative approach to the United States’ voluntary data disclosure issues has been the ‘grant programs with strings attached’ approach set forth by Congress in 2006 with subsequent amendments.

However, it is still curious that the administrative data issue is widespread, not concentrated in certain policy areas. This phenomenon seems to suggest that there is a sort of common feature causing the problem—a feature beyond the policy specific ones mentioned before. In other words, we may have to admit that fractured data bureaucracies are a natural consequence of the country’s federalism—or an interpretation thereof. Beginning in the 1960’s the United States’ mode of governance transitioned to what some scholars have dubbed coercive federalism: a type of federalism distinguished by, among other things, more policy conditions on federal aid and unfunded mandates. For example, the current administration has seemingly granted coronavirus aid requests in a way that benefits political allies more than Democratic districts. This suggests states are incentivized to not cooperate in data sharing because the information states provide may be used against them punitively. California seems to have experienced this first hand when the Environmental Protection Agency used the state’s homelessness statistics as a sword of Damocles to encourage compliance with deregulatory agendas. Some argue Wyoming experienced the same in the late 80’s when they raised their drinking age to preserve federal highway funding (neighboring South Dakota actually tested the rule in the US Supreme Court’s 1987 South Dakota v. Dole case.) While this case is not necessarily data driven, it illustrates how the state/federal relationship might be viewed as coercive.

The data-sharing issue has solutions at two levels. First, on a purely administrative level, the data sought by the federal government should be operationalized in a uniform way as much as possible. Different federal agencies and programs have widely disparate submission requirements, deadlines, and the like. Obviously, certain agencies will require different or specialized data to accomplish their goals (e.g. education data versus Social Security data). But by encouraging cooperation and modernization of federal submission infrastructure—similar to what the NO HATE Act does for hate crime figures—state governments would face fewer barriers in reporting.  This would require comprehensive legislation to give states the funding needed to update systems and for the federal government to modernize its own.

Next, in light of the coercive federalism issue, local and state data should be protected from political use. That is, the data local and state governments share should be shielded from use in considering how the state fares on unrelated federal initiatives. This would help prevent placing states and localities in an impossible situation where they either fail to report and receive no aid or report and are punished in other areas. Additionally, the Government Accountability Office (GAO) has suggested that increased compliance with E.O. 13132 could increase collaboration and communication between the federal government and state/local officials. This executive order requires the designation of a federalism official to consult with state and local governments over policy implementation when significant federalism interests are implicated. Currently, implementation of this order—which is the purview of the Office of Management and Budget—has been sparse. The GAO reported the OMB was unable to identify oversight steps in regard to the order, signaling that there is much work to be done on this front.

Disclaimer: The views expressed above are the author’s own and do not reflect those of SALPAL or Georgetown University.