The following post has been cross-posted from the Election Law Journal with the generous permission of author Doug Chapin:
The growing enthusiasm across the country for non-precinct place voting (NPPV) presents the election administration field with a series of challenges and opportunities with respect to the design and implementation of jurisdiction-specific programs to put NPPV into practice. Much of the impact of NPPV has been temporal—i.e., tied to the expansion of the notion of Election Day. Traditionally, Election Day marked the only opportunity for the vast majority of voters to cast their ballots; today, Election Day is merely the last day a voter can cast a ballot. Much of the popular scrutiny of NPPV to date, then, has focused on this temporal expansion, along with its attendant effects on candidate and voter behavior. Equally important, though, is NPPV’s spatial expansion of election administration. NPPV has inexorably eroded the traditional equivalence between electoral geography—that unique combination of candidate and non-candidate contests that comprise a voter’s ballot style—and the physical location where a voter actually casts that ballot. NPPV’s temporal and spatial effects have combined to create a modal expansion for voters and election officials alike. Because voters now have more choices about when and where to vote, election administration has had to evolve to become an increasingly complex system to cope with ballots cast at different times and at different places, but also in different forms.
This three-dimensional expansion has created a series of policy challenges for the field.
Overlay with voter registration
Despite NPPV’s overwhelming change in how, when and where voters cast ballots, the underlying requirement of voter eligibility—specifically, the voter registration process which exists in every state but North Dakota—remains. States with Election Day registration are already familiar with the process of citizens registering and voting simultaneously, but law and policy have had to adjust in states where the registration deadline falls before Election Day.
Different approaches have emerged to deal with this challenge. North Carolina, for example, has developed ‘‘one-stop voting centers’’ where voters can register or update their registration up to the Sunday before Election Day and then cast a ballot on the spot (but, interestingly, not on Election Day). This is not always intentional; a one-week statutory overlap between early voting and the registration period set off a fierce debate in Ohio in 2008 after then-Secretary of State Jennifer Brunner (D) issued a directive creating a so-called ‘‘golden week’’ where voters could register and vote on the same day. Elsewhere, the growing use of electronic pollbooks—whether or not linked in real-time to a voter registration database—has allowed voters to cast ballots outside the limits of the traditional polling place without sacrificing the eligibility check implicit in the registration process.
Synchronization with the political map
While NPPV has, in some sense, loosened geographic and temporal restrictions on voters casting ballots, such restrictions (including the bold grey area that is domicile) are still important to the determination of whether voters are eligible, and what contests they are eligible to decide. As polling places drift further away from ‘‘home precincts’’ under NPPV, election offices face a two-fold challenge: making sure, first, that each voter is eligible for the ballot he casts and, second, that each voter receives the ballot he is entitled to cast based on that eligibility. In some states, a third challenge has emerged: assuring that the early voting returns can be reallocated back to the geographic precinct for the purpose of political canvassing and redistricting (keep in mind that ballot styles span many precincts, and votes tallied at a county office or satellite location are not necessarily ‘‘coded’’ by precinct of origin).
The response to this challenge has been largely technological. Initially, the advent of direct recording electronic (DRE) machines was seen as a promising means of ensuring that voters receive the correct ballot regardless of voting time and location. As doubts about DREs have grown and more jurisdictions migrate toward opticalscan, we have seen the development of the concept of ‘‘ballot on demand’’, which allows an election worker to produce a ballot as voters appear at NPPV stations.
Ballot on demand technology is still being tested and developed, however. In the meantime, election officials continue to search for ways to efficiently and effectively make voter-specific paper ballots available at NPPV stations. The alternative is a difficult decision on the size of print runs for ballots: too few, and a jurisdiction runs the risk of falling short on Election Day as was the case in Bridgeport, Connecticut in 2010; too many, and the result is a waste associated with huge numbers of unused ballots—a common problem, especially in low-turnout off-year or special elections.
Impact on returns – unofficial and official
It is well-known—indeed, has been wellremarked-upon—what effect NPPV has had on the ballot casting process. There are now so many different options for casting a ballot that the traditional precinct-based Election Day ballot is usually a choice rather than a necessity for most voters. And yet, while election offices open new and different modes for voters to cast ballots, the laws and procedures for counting ballots are largely unchanged. This creates numerous challenges.
First, the growing percentage of NPPV ballots cast outside the traditional polling place has completely upended the typical Election Night unofficial reporting experience. In 2010, the Associated Press announced that its unofficial tallies would be reported as a percentage of the ‘‘expected vote’’ instead of ‘‘precincts reporting.’’ Thus, as more and more races come down to the wire—a wire comprised often of NPPV ballots—it is increasingly dangerous for anyone, especially candidates, to make an assumption that Election Night totals will hold.
Second, whether or not one or more contests are close, the advent of NPPV means that before a jurisdiction can begin counting ballots and preparing to certify returns, it must first collect and sort those ballots into a form that allows for counting. In places like California, where vote-by-mail ballots (N.B. that they are no longer called ‘‘absentees’’) can be delivered to any polling place in the county before the close of polls, this sorting process is not trivial and can slow the pace of the count even if it does not delay the official certification of returns. When questions arise about ballots themselves—as they did famously in Minnesota’s razor-thin 2008 U.S. Senate race—scrutiny during counting (often accompanied by litigation) can delay the results past the official deadline. How to reconcile this growing need for deliberation with the ever-accelerating public and media demand for information about election results will be a constant concern for election offices for the foreseeable future.
Administration and efficiency concerns
Even assuming that election offices overcome all of these other challenges, they will be left with the question of how to make NPPV work—and how to pay for it. To date, every state that employs NPPV is layering it onto a pre-existing election framework (with the exception of Oregon and Washington, which are now all vote-by-mail). As the proportion of voters using NPPV grows—with the dollars available to cover election costs staying flat if not decreasing—jurisdictions must find a way to align demand for voting across all different modes with the resources available to support them.
We are already seeing initial efforts in this direction, both in terms of jurisdictions closing or consolidating Election Day polling places and in their rethinking how and when to offer NPPV. Georgia recently reduced its early voting period by a week, after its research indicated that most voters did not use early voting during the first week it was offered. (Anyone who has heard ELJ Co-Editor Paul Gronke speak in the last several years will be familiar with this research.)
The next step, however, will be for the election community to engage in more detailed calibration in response to NPPV’s ‘‘supply and demand.’’ As NPPV expands, researchers are going to learn more and more about who uses NPPV and why.
Similarly, as more and more jurisdictions begin to collect data on what it costs to administer elections, they will get clarity on what NPPV costs—both alone and in relation to traditional polling places.
Someday, these two strands of data will converge and election officials will be able to allocate resources to NPPV and traditional polling places much as an investor does to stocks, bonds and cash—maximizing return at the most affordable cost. Such analysis will never replace the tough policy decisions—unlike funds in a portfolio, voters are not fungible—but it will almost certainly result in a better-managed election system.
Address correspondence to:
Director of Election Initiatives
PEW Center on the States