The debate over COVID-19 aid for state and local government revealed a broad range of opinions that were only partially informed by the facts. Lacking complete information about the history and trajectory of state and local government revenues and expenditures, federal policymakers had to make educated guesses about how to size and allocate the taxpayer-funded aid—or if the aid was needed at all. Policymakers and taxpayers could avoid a similar conundrum during a future crisis by investing in financial data standards, increased reporting frequency and fully transparent predictive models now.
The data void at the federal level regarding the fiscal position of state and local governments is not due to the absence of state and local government financial data. All states and most sizable local governments produce annual audited financial statements, but these usually appear 6-9 months after the end of the fiscal year. Most states and a few local governments produce monthly or quarterly financial reports—typically on a cash basis and limited to general fund activity—with a much shorter time lag.
Most annual audits comply with pronouncements from the Governmental Accounting Standards Board yet appear in widely varying formats. Interim reports are not governed by national standards and lack even a modicum of consistency.
The Securities and Exchange Commission (SEC) has raised the issue of stale municipal financial disclosure in recent years. For example, an SEC transparency subcommittee observed:
[T]he average municipal issuer provided its annual financial information within 12 months of the end of its fiscal year provided such annual information 188 days after the end of the applicable fiscal period. If a municipal issuer does not provide interim financial disclosures and it files its annual financial disclosures within the averages referenced above, the financials available to investors could be over 500 days old as the next submission date approaches.
During the coronavirus pandemic, the lack of timely and standardized state and local disclosure prevented policymakers from quickly determining with any degree of accuracy how much revenue was lost as the pandemic unfolded.
If every state, major city, and county produced monthly cash reports in a standardized, machine-readable format within two weeks of month-end, federal policymakers would have had a much better picture of how revenues and expenditures are evolving.
At the federal level, the Treasury Department shows the way by producing its Monthly Treasury Statement in Excel format each month. Excel files are easier to parse and thus consolidate than Adobe PDF files. Even better than Excel are non-proprietary formats such as CSV, JSON and XML. All of these options are supported by Treasury’s new FiscalData website, which includes an exportable dataset of Monthly Treasury Statements.
Among states, Texas provides its monthly state revenue collections in both Excel and CSV formats. Connecticut offers a monthly revenue data set exportable to Excel, CSV and XML formats. Local governments are generally less advanced, but New York City—as the nation’s biggest city—does provide extensive revenue and expenditure data in a Google Sheet updated quarterly.
If the nation’s 200 largest state and local governments provided data like this, the vast majority of state and local financial activity would be available in near real time. But collecting this data would still be non-trivial due to variances in the number, type and identification of the data elements provided. Ideally, of course, all state and local governments would adopt a single reporting taxonomy (i.e., data dictionary) that could be readily consolidated. Our Standard Government Reporting working group at XBRL US has produced a draft taxonomy for annual financial reporting. A small subset of items from this taxonomy could be applied to monthly reporting. Our taxonomies can be used with CSV, JSON and human-readable HTML reports.
In the absence of this standardization, data collection is typically handled by organizations that are willing to make the investment of time and energy to process all the varying government disclosures. This too often means that compilations are not available to the public for free or that the data is selectively reported in support of a certain political narrative (e.g., state and local governments are suffering and need a large federal aid package, or they are doing just fine and do not require a bailout).
The federal government—through the Census Bureau—undertakes a parallel data collection effort which is likely free of commercial or ideological motives. The Census Bureau conducts an Annual Survey of State and Local Government Finance and produces a Quarterly Summary of State and Local Tax Revenue. Unfortunately, these products appear after significant time lags, and because they are based on separate data collection instruments than those of state and local governments they may not fully agree with public financial reports issued by each state and local government. The Census Bureau’s efforts could be improved by the availability of public, standardized, machine-readable data
Commercial and ideological players also dominate the discussion of projected revenue losses and spending, using proprietary models to reach their findings. There’s nothing wrong with that, but as we learned during the global financial crisis of 2007-08, if government agencies rely on proprietary models operated by self-interested parties with financial stakes in the game, these models can help produce unacceptable results for taxpayers. Although the Congressional Budget Office is not always correct, it has achieved a level of openness and objectivity that instills confidence across most of the ideological spectrum. As federal involvement in state and local finance increases, citizens need institutions whose analyses and projections of state and local government finances would be met with similar levels of confidence. Perhaps CBO could expand into this area, or a non-ideological, non-profit could take on this challenge. In any case, taxpayers should have easy access to the data and modeling software code and assumptions would ideally be placed in the public domain so that third parties could fully review the analysis.
Economic historians will ultimately weigh in on the question of whether the American Rescue Plan’s $350 billion state and local aid component was reasonably sized and allocated. For those of us condemned to live in the present and without the benefit of hindsight, we should be taking every opportunity to improve our data and analytics to make the best possible decisions during the next financial emergency.
A version of this column previously appeared on Federalism US.