ABSTRACT
Voter identification laws (“voter ID laws”), laws that require voters to present identification when voting, launched the modern Voting Wars. After the Supreme Court blessed Indiana’s voter ID law in Crawford v. Marion County, voter ID laws proliferated across the country. Their prevalence belies their notoriety. They remain one of the most hotly contested category of election laws and are often referred to as a voter suppression law, if not the modern voter suppression law.
While these laws first served as a rallying cry for the election law, they have become a sore spot for what is historically a collaborative and close community of social scientists, lawyers, and legal scholars. Many social scientists have concluded that voter ID laws have had negligible effects, if any, on voter turnout. That conclusion may seem surprising—even difficult to believe—given how many eligible voters lack IDs. It has raised uncomfortable questions about whether the progressive legal alarm over voter ID laws—including litigation challenging those laws—has been warranted.
By harmonizing the causal social science literature and descriptive evidence unearthed in the course of litigation, this Article is the first to offer an account of why empiricists have consistently failed to detect a turnout effect from voter ID laws. Upon investigation, what is surprising is not that a turnout effect has not been detected, but why an effect should have been expected in the first place. Evidence from litigation suggests that more than 99 percent of registered voters who habitually vote may have the requisite ID for voting, even though large numbers of eligible (but not registered) citizens lack IDs. It is therefore unsurprising that the best causal studies suggest that voter ID laws decreased turnout (that is, voting conditional on registration) by no more than 2 percent. Those studies should not have expected any other result: existing causal studies sought to detect an effect that descriptive evidence did not support. Thus, the discord in the literature results not from the sidelining of important causal findings, but rather from the lack of interaction between the causal academic literature and litigation-derived descriptive evidence.
Resolution of the debate on the turnout effects of voter ID laws has far reaching implications for the election law community. For legal scholars in particular, doing so highlights important responsibilities in maintaining an interdisciplinary relationship with social scientists. The traditional notion of the interdisciplinary relationship between empiricists and lawyers in the field of election law is one of answering questions and questioning answers, in which social scientists offer empirical answers to questions posed by lawyers, and lawyers in turn question the relevance, importance, and weight of the empirical answers provided by social scientists. Resolution of the debate over voter ID laws’ effects suggests that election law scholarship should also question questions: lawyers should not only question the empirical answers that social scientists offer, but also their hypotheses and methods in reaching those answers.
The voter ID debate supplies two additional examples of questions worth questioning. First, is the estimated effect big or small? Social scientific assumptions in interpreting empirical effect sizes do not hold for legal evaluation. While social scientists are interested in comparing the effect of election laws against all other drivers of turnout, legal interest is limited to how an election law compares to other laws. Second, is the law in question a voter suppression law? In assuming that laws that do not depress turnout are not voter suppressive, social scientists measure vote suppression, which is not the same as voter suppression. Understanding an election law’s suppressive effects solely through turnout evidence ignores burdens that voters take on to comply with onerous laws, as well as mounting barriers that further discourage disaffected individuals from voting.
Questioning questions also helps clarify doctrine. I consider how courts hearing challenges to voter ID laws have applied—and misapplied—turnout evidence in conducting the burden inquiry in the Anderson/Burdick standard governing federal constitutional protections for the right to vote. The Anderson/Burdick standard balances the burdens imposed by the challenged law on the right to vote against the state’s justification for the law. Causal evidence of turnout effects is a clearly efficient—but also radically incomplete—measure of burdens on the right to vote. Conceptual clarity of both what turnout estimates measure and what the doctrine asks ensures not only that all relevant evidence is presented and considered in voting-rights cases, but also that the social science literature is better positioned to produce doctrinally responsive research.
