Mike Parsons has posted this draft on SSRN (Vanderblit L. Rev En Banc). Here is the abstract:
How much does an election cost? For a democracy as old as ours, the answer is surprisingly unclear.
As election officials prepared to hold the 2020 election in the midst of a pandemic, many tried to ballpark the change in resources necessary to run a safe and effective process. The Brennan Center recommended $2 billion based on rough estimates, only to revise that estimate a month later to $4 billion to cover the range of elections that states and localities would hold over the year. Congress eventually provided $400 million 3—a figure so inadequate that Mark Zuckerberg and Priscilla Chan’s charitable contributions alone matched it. Additional sums came from private philanthropy, corporate in-kind donations, and (of course) state and local funding—not to mention a wide range of unreimbursed expenses likely incurred by election workers themselves.
Thanks to the heroic work of administrators, poll workers, and civic groups, the 2020 elections were largely a success. But how much was ultimately spent, how much should have been spent, and how much is typically spent remains a mystery. For example, what would it have cost to ensure that no one waited longer than 30 minutes to cast a vote in 2020? What would be the price tag for 2024? There is no central source of authoritative information to help answer these questions. “States know how much they spend on roads, health care, education and other big-ticket items, but no one knows how much they spend on elections.”
With Democracy on a Shoestring, Professors Joshua Sellers and Roger Michalski step into this opening and make vital contributions. To start, the authors give us something new: hard data. Through immense and detailed research, the authors have compiled a dataset that provides real-world insights into spending patterns in California, Arizona, Texas, and Florida. Using predictive machine learning, the authors then build on this dataset to offer detailed average spending estimates across multiple governmental units.
From these estimates, Sellers and Michalski glean important insights. First, while there is great variation both across and within states, the average spend is small. Second, and more surprisingly, the variation in spending “is seemingly unconnected to poverty, race, and other traditional explanations of electoral disadvantage.” Finally, the authors leverage these insights to sketch out potential policy options, doctrinal implications, and future research paths.
Their intervention could not be more urgent. Scholarly debates, legislative policy discussions, and judicial remedies often unfold in the abstract, with insufficient attention paid to hard costs and administrative concerns. This would be enough of a problem if all the major players in the electoral system were operating in good faith. But they increasingly are not. When every small hiccup provides fodder for conspiracy theories designed to draw the outcome into doubt, the stakes of adequate funding become higher than ever. And when across-the-board cutbacks risk potentially increasing the costs of electoral access for some voters more than others, equitable funding becomes more important than ever.
Among the many contributions of Democracy on a Shoestring, then, is to spur more concrete thinking about the costs and consequences of our country’s devolved and varied spending patterns and decisions—a topic that generated substantial interest after the 2000 election but has since waned. Sellers and Michalski have given us a wealth of empirical, doctrinal, conceptual, and practical information to kickstart these conversations.
In this response I offer two areas that warrant further emphasis and examination. In Part I, I highlight how many of the potential policy solutions identified by Sellers and Michalski could be far more powerful if implemented at the federal level rather than the state level. This includes the benefits of increased funding, uniform data collection, and soft consolidation of purchasing and expertise to leverage economies of scale. While the authors understandably focus on state-level action given the historically “hyper-decentralized” nature of election administration, recent elections reveal new reasons to believe the politics underlying this traditional arrangement may (and should) be shifting.
In Part II, I examine the authors’ surprising finding that variations in election spending are unconnected to poverty, race, and other traditional explanations of electoral disadvantage. This finding suggests that disparate electoral opportunities are not simply a matter of unequal funding. Instead, we must look elsewhere to identify the source of these disadvantages. While Sellers and Michalski seem skeptical that electoral disadvantage may thus be attributable to variations in the relative needs of communities, I believe that this kind of “structural suppression” could be more substantial than the authors credit, and that future research should focus on rigorously examining and quantifying this phenomenon.