Public Policy Forum Blog

Milwaukee and the Voter Fraud Debate

The long-debated issue of election reform resurfaced in Milwaukee recently, when two groups conducting major voter registration drives – the Association of Community Organizations for Reform Now and the Community Voters Project - reported that several of their employees were submitting voter registration cards under the identities of “dead, imprisoned, or imaginary people.” In both cases, the workers involved were fired, and the organizations voluntarily notified the Milwaukee Election Commission of the inaccuracies. Nonetheless, the incident raised worries that Wisconsin’s election system could be vulnerable to widespread fraud.

Such worries are hardly new in Wisconsin: the November 2004 presidential election was marred by a number of discrepancies. In Milwaukee, the number of ballots cast outnumbered the list of registered voters by 4,609 votes. Revelations of suspect voting practices triggered a firestorm of criticism and attracted national attention. Then, as now, some commentators, like Chris Lato and Brian Fraley, reacted by calling for the state to require registered voters to show photo identification at the polls.

The argument over voter ID laws in Wisconsin has unfolded along the same lines as similar debates in other states and at the national level. Advocates of more stringent ID requirements claim that they will safeguard against the fraudulent impersonation of registered voters. Opponents counter that requiring IDs effectively disenfranchises poor, elderly, and minority voters, who are less likely to have photo identification. A political current runs just beneath the surface of the discussion: minorities and working-class people, the groups voter ID opponents claim will suffer from voter ID requirements, tend to hold more liberal political views.

But a look at the evidence suggests that the hype on both sides of the voter-ID debate is largely unmerited. On one side of the debate, there is little empirical proof that requiring ID would suppress turnout in a discriminatory fashion. A paper by Timothy Vercellotti and David Anderson concludes that voters in states that require photo ID at the polls are 2.9% less likely to vote, but that photo ID laws had no statistically significant effect on African-American and Hispanic turnout. Other analysts argue that Vercellotti and Anderson’s methodology led them to overstate the negative effect of voter-ID laws.

On the other hand, there is little evidence to substantiate fears that election fraud will become widespread enough to tip the balance of an election. In 2004, even if each of the 4,609 questionable ballots in Milwaukee had been cast for John Kerry, they would not have decided the contest, which Kerry won by a margin of 11,384 votes. Moreover, the vast majority of the ineligible votes did not constitute criminal election fraud, which Wisconsin law defines as an intentional act. Many of the inappropriate votes appeared to be the result of errors by Election Commission staff. A lengthy investigation by the Milwaukee County District Attorney and the U.S. Attorney for Milwaukee found sufficient evidence to prosecute fourteen cases of fraud; five of those cases resulted in convictions. The U.S. Attorney, Steve Biskupic, concluded there was no “massive conspiracy” to change the results of the election in Milwaukee. That the ineligible votes cast in 2004 were mostly the result of unintentional error is problematic, but not necessarily an indicator of the potential for widespread criminal election fraud.

The voter ID debate is moot in Wisconsin for the time being, since Democratic control of the state Senate will prevent the proposal from passing until at least 2011. In the meantime, though, a much less controversial reform may well eliminate many of the errors that have led to calls for tighter regulations on voting. The state recently unveiled a long-overdue voter registration database, which local election clerks will use to check newly registered voters against lists of felons, deceased people, and licensed drivers. Many of the problems that emerged in 2004, and again in this month’s registration-drive flaps, could have been prevented if the statewide database had been fully functional. If the new system performs as well as promised, it could help curb the potential for election fraud and allow policymakers to more fully debate the need for a voter identification law.