LAWRENCE — The United States has been holding elections for more than 200 years. So, we should know how much they cost, right? A new book co-written by a University of Kansas researcher is one of the first in-depth studies of the cost of U.S. elections.
“A Republic If You Can Afford It: How Much Does it Cost to Administer Elections?” analyzes data from 48 U.S. states from 2008 to 2016 and found that some states spend an average of about $2 per voter, while others spend an average of almost $20. The book’s authors also found that in addition to varying costs, different regions of the country have seen changes in how much they spend and that trust in relation is associated with how much is spent.
Zachary Mohr, associate professor of public affairs & administration at KU, is one of the authors of “A Republic If You Can Afford It.” He studies costs and accounting in local government and recounted a story of a conversation he had with co-author Martha Kropf of the University of North Carolina at Charlotte.
“She said, ‘You know, Zach, the one thing we don’t know anything about is how much election administration costs.’ I said, ‘The data may not even be available, but one place we can probably find it is North Carolina,’” Mohr said. “We were doing elections for 200 years, and we were doing them at the local level, and there really wasn’t much problem until about 2000. Then there were hanging chads in Florida and other problems, and people started paying attention to election administration. The federal government then made money available to address and study the issue.”
Mohr, Kropf, Mary Jo McGowan, also of UNC-Charlotte, and independent researcher JoEllen Pope collected the data from every state except for Michigan and Wisconsin, which did not have data available, and began publishing studies in 2017, which eventually formed the basis of the “A Republic If You Can Afford It.”
The book, published by Cambridge University Press as part of its Cambridge Elements Series, is available for free online until Dec. 30.
Costs vary
The first key finding is that the cost of elections varies widely from state to state and even from one jurisdiction to another, such as county to county . The early part of the book breaks down the measure the authors used to analyze the costs. Much like accountants, the researchers spell out the differences between different cost objects and different measures of cost. For example, one cost object is the expense of holding an election, a short period of time from when candidates start filing ballots to after election votes are audited and certified. Another cost object is the cost of election administration throughout the year as the need to maintain equipment and voter registries is a year-round activity. Other issues included the difference between expenditures and expenses. The researchers settled on analyzing election administration expenditures throughout the entire year because that is the data that was most readily available.
The authors also detail why costs can vary so widely. One example is in highly populated jurisdictions, there are more taxpayers contributing to the government’s budget, which spreads out the cost more per person than what is observed in rural locations. Some areas see those economies of scale, but as the authors point out, other areas see diseconomies of scale as well. Priorities and political will of a location also factor into costs. The research found that in places where the governing party matches that of the larger populace, they tend to spend less on election administration than bodies who are of an “out party” who spend more, often in emphasizing voter turnout, Mohr said. Other factors, such as a populace’s average educational attainment, income and demographic makeup also factor into amounts spent.
Policymakers tend to take a one-size-fits-all approach to legislation, but the data shows that is not applicable in election administration.
“If you’re looking at the question of if the state or federal government should support local election administration, you have to realize that $1 per voter in one state goes much farther than it does in another. You see variations across states, but also within states,” Mohr said. “It’s interesting to us why some counties spend so much more than others. Most of the book is about why that difference exists.”
One of the reasons the cost varies so widely, the authors write, is that the system is complicated in most states, with various jurisdictions holding elections, often with differing guidelines.
Changes over time
In addition to covering previous research on the topic, “A Republic If You Can Afford It” examines changes in spending across regions of the United States in the data set from 2008-2016. The Great Recession of 2008 led to reduced spending on elections, which was expected, but in findings Mohr called “surprising and almost shocking,” how spending came back over the next eight years differed by region. In the South, spending increased, largely due to the fact that more people were moving to that region. In the West, spending dropped, due at least in part to states such as California and Oregon putting more emphasis on voting by mail. Spending in the Northeast dropped significantly but then came back to 2008 levels.
Spending and trust
The last key finding was that the more jurisdictions spend on elections, the more trust voters have in the system being free of fraud and representing free, open elections. Locations that spent much higher than others also saw correspondingly high trust, rising in correlation with the amount spent, the authors write.
The book points out that the figures it presents are not static and will not guarantee the costs of future elections. In fact, it can be considered the minimum, or a “lowest bar estimate” of the cost of American election administration.
“This is at least a step in the right direction, we feel, and can help guide the conversation knowing that there is not a one-size-fits-all approach,” Mohr said. “There are at least a dozen languages to translate ballots into in Los Angeles County that other counties don’t have to address, for one example.”
Public dataset
The book’s dataset is included and is publicly available for other researchers, policymakers and anyone interested in knowing more about the cost of election administration.
“This data is really needed in election science to control for models of voter turnout and other election administration outcomes,” Mohr said. “We want to make sure it is freely available for anyone who wants it.”