Timing And Turnout In Ohio

Walter Mebane

July 21, 2005

Walter Mebane is professor of government at Cornell University.  He was a member of the team that produced the DNC Ohio 2004 study.  Before that, he worked as part of the National Research Commission on Elections and Voting created by the Social Science Research Council to monitor, analyze and report on problems and successes in the 2004 election.

On June 29, 2005, John Tanner of the U.S. Department of Justice (DOJ), wrote to Nick Soulas, Assistant Prosecuting Attorney, Franklin County, Ohio, to report the results of "an investigation into the November 2, 2004 general election in Franklin County, prompted by allegations that Franklin County systematically assigned fewer voting machines in polling places serving predominantly black communities as compared to its assignment of machines in predominantly white communities."  While far from a glowing evaluation, the conclusion Tanner stated must have pleased Ohio officials: "Franklin County assigned voting machines in a non-discriminatory manner.  Accordingly, there was no violation of Section 2 of the Voting Rights Act, 42 U.S.C. 1973, and we are thus closing our file."  Headlines two days later broadcast that gist. "Department of Justice finds no racial disparities in Ohio voting," went an AP story.  "Justice Clears Ohio in Voting Booth Bias," said another.

The truth is not so simple. The DOJ report concedes there were more registered voters per voting machine in precincts that had many African-American voters than in precincts that did not.  In fact, using data from Franklin County, I calculate that the quartile of precincts with the largest proportion of African-American voters had 23.7 percent more registered voters per voting machine than the quartile of precincts with the smallest proportion.  One might think that fact would end the argument, but the DOJ notes that different results appear when one looks at the number of ballots cast.  The DOJ reports states that "the allocation of voting machines actually favored black voters because more white voters were voting on each voting machine than black voters."  Indeed, in the quartile of precincts with the smallest proportion of African-American voters, there were 3.7 percent more ballots cast per machine than in the quartile with the largest. This is the basis for the DOJ's conclusion.

The Turnout Question

It is easy to see the big hole in the DOJ analysis.  The Democratic National Committee's study of the 2004 election in Ohio (Democracy at Risk: The 2004 Election in Ohio) finds that throughout the state, failures in election administration caused racial disparities in voters' election day experiences.  African-American voters reported waiting an average of 52 minutes before voting, while white voters reported waiting an average of 18 minutes. Three percent of voters who went to the polls left their polling places and did not return due to the long lines.  Throughout Ohio, reductions in voter turnout are associated with inadequate provision of voting machines.  The DNC study finds that turnout is typically two to three percent lower in precincts that were allocated fewer machines.

The DOJ report completely ignores the effect the machine allocations had on turnout.  Data from Franklin County show that the typical reduction in turnout associated with the machine allocations is more than four times larger in the quartile of precincts with the highest proportion of African-American voters than it is in the quartile with the smallest.

The DOJ report mentions voter turnout only in relation to general facts about racial differences in turnout that purportedly explain why voting machines may have been allocated as they were.  The DOJ report states, "In considering voter turnout, the Board tended to allocate fewer machines to the 54 predominantly black precincts per registered voter because of the long history of lower black turnout."  To take turnout differences into account, Franklin County officials use a concept of "active voters" to make the machine allocations.  Data provided by Franklin County include counts of the number of active voters in each precinct.  The DOJ report says nothing about active voters.  In fact, the quartile of precincts with the highest proportions of African-American voters had 13.6 percent more active voters per voting machine than the lowest quartile of precincts.  Measured by the active voters standard, machine allocations in the county still appear to be seriously discriminatory.

The Timing Question

But we are not yet at the simplest truth. The data from Franklin County include counts of active voters at two points in time.  The figures I just cited use the active voter counts dated November 4, 2004—two days after the election.  Another set of counts is dated April 27, 2004.  Using the April active voter counts, there are more active voters per voting machine in the quartile of precincts that have the lowest proportions of African Americans (189 voters per machine) than in the quartile of precincts with the highest proportions (177 voters per machine).  The DOJ might have chosen to say that Franklin County officials used the April counts of active voters to allocate voting machines.  The DOJ notes that "the Board [of Elections] had to make allocation decisions well before the election."

Of course, to say that Franklin County election officials relied on the April information would have been to pinpoint grossly negligent administrative behavior.  As every observer of Ohio during 2004 knows, there was substantial voter mobilization between spring and fall.  Between April 27 and November 4,  the number of active voters in the county increased by 27.4 percent.  No one could have been surprised by that, and certainly not election officials who were struggling to keep up with a flood of new registrants.  Between April 1 and November 4, voter registration increased by 15 percent in Franklin County. The surge of new registrants should have been a clear warning that any plans made based on the spring information would fail.

Alas, it wasn't.  Racially disparate effects on voter turnout were the readily forseeable consequence.

Negligent Or Malicious?

What about the DOJ report result that there were more ballots cast per machine in the precincts having the smallest proportion of African Americans?  All that tells us is that, in these precincts, each voter typically took less time to vote than voters did elsewhere.  In the precincts with the lowest proportion of African Americans, typically 13 ballots were cast per voting machine per hour that the polls were open. In the precincts with the highest proportion of African Americans,  typically 11 ballots were cast per voting machine per hour. The number of ballots cast per machine is not a measure of the fairness of the voting machine allocation.

Should the DOJ have exonerated Franklin County?  The argument they used to do so ignored an important fact: The allocation of voting machines reduced voter turnout more among African American voters than among white voters.  Probably the simple truth is that Franklin officials made their voting machine allocation decisions using the spring data on active voters, doing little or nothing to respond to the new voter registrations.  Was that negligence malicious?  Was it criminal?  I am not a lawyer, but a clever attorney might argue that the Franklin County machine allocations had a rational and nondiscriminatory basis in the April data.  That might legally exculpate them. But the DOJ did not take this approach.  The least one can say is there was shamefully negligent administrative failure that deprived thousands of voters of their right to vote.