> Click to view the PDF version
|U.S. Department of Justice |
Civil Rights Division
|Voting Section - NWB||950 Pennsylvania Avenue, N.W.|
|Washington, DC 20530|
|June 28, 2005|
Nick A. Soulas, Jr.
Assistant Prosecuting Attorney
373 South High Street
Columbus, Ohio 43215
Dear Mr. Soulas:
The Voting Section of the Civil Rights Division recently conducted 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. As set forth below, the evidence our investigation has revealed establishes that 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.
We note at the outset that elections in Franklin County, as elsewhere in Ohio, are run by a six-member Board of Elections, three of whose members are representatives of each major political party. In Franklin County, the members of the two parties appear to work well together and share a common goal of running fair elections. The structural bi-partisanship of the Board extends to active cooperation in decisions on election issues, and that spirit of cooperation also has been extended to our investigation. We very much appreciate the readiness and openness with which the Board has greeted and filled our requests for records and other information.
It is clear that there were long lines at polling places across Franklin County, and it was not uncommon for voters to have to wait three or more hours to cast their ballots. This was especially true within the City of Columbus, where the ballot was exceptionally long. But the long lines were attributable not to the allocation of machines, but to the lack of sufficient machines to serve a dramatically enlarged electorate under any allocation. Voters came to the polls in record numbers last November. Within Franklin County, the number of registered voters rose from 681,949 for the 2000 general election to 845,720 for the 2004 general election, an increase of 24%. Voter turnout rose even more sharply, from 417,800 in 2000 to 535,575 in 2004, an increase of over 28%. In fact, Ohio as a whole had one of the largest increases in turnout in the nation from the 2000 election. To ensure that this type of situation does not occur again, the Board has determined on its own initiative to increase the number of voting machines from 2,904 to 5,000 for the 2006 election.
A major complicating factor in the appropriate allocation of voting machines was the artificially inflated voter registration rolls in the county. The 845,720 registered voters in 2004 actually exceeded the 2000 total voting age population of Franklin County (800,657) by 45,063 persons. This unsettling disparity resulted from the loss, during preparation for computer system changes in anticipation of the year 2000, of voter history data necessary for purging the voter rolls of ineligible voters as required by the National Voter Registration Act. The County chose to start fresh with new voter histories, with the result that there had been no voter purge since 1999. The County resumed regular purging of its voter list only after the 2004 election, and on June 20, 2005 removed approximately 114,000 ineligible individuals from its voter registration list. The 2005 purge brings the voter registration total well below the 2000 voting age population in the county.
In any event, the Board used the inflated voter rolls in the Fall of 2004 as one factor in its allocation of voting machines, and it also used past voter turnout as another factor. Of course, any allocation -- no matter how bipartisan -- was inevitably going to be inelegant and imprecise given that the Board had to make allocation decisions well before the election and was constrained by the number of voting machines available (2,904), the number of precincts (788), and the Ohio practice of having at least two machines in each precinct so that voting can continue if one machine breaks down. As Elections Director Damschroder acknowledged, the process involved "some math and some art."
With respect to the purported racial disparities in machine allocations, the reality is that while there tended to be more registered voters (based on the inflated rolls) per machine in predominantly black precincts than white precincts, the allocation of voting machines actually favored black voters because more white voters were voting on each voting machine than black voters. To understand how this came to be, it is important to note that voter turnout in Franklin County, as in much of the United States, varies significantly by race. Within the 392 precincts whose 2000 population was over 95 percent white in voting age population, voter turnout was 60.1% of total voter registration in 2000, and 68.1% in 2004. The 54 precincts over 66.6% black in voting age population had turnout of 51% in 2000 and 59.5% in 2004. Within the 12 most heavily white precincts, voter turnout was 78.9%. In the 12 most heavily black precincts, turnout was 61.8%.
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. But while there were fewer voting machines in black precincts than in white precincts based on registration, the disparity was reversed when compared to actual voter turnout in the 2004 election. In fact, voting machines in the white precincts were busier than the machines in the black precincts, and black persons who went to the polls were not at a disadvantage due to the number of machines. To the contrary, the allocation of voting machines actually favored black voters because more white voters were voting on each voting machine than black voters. (Specifically, white precincts averaged 172 voters per machine, while black precincts averaged 159 voters per machine.) The disadvantage for white voters (less than 10%) was relatively minor, and not enough to violate the Voting Rights Act. Nevertheless, we note that but for the decision of the Board to adjust for voter turnout ahead of time, the disparities would have been much greater.
Our investigation further revealed that the predominantly black polling places stayed open later than the predominantly white precincts in order to serve those voters who were in line at the normal 7:30 p.m. poll closing time. Polls in Ohio normally are open for 13 hours, from 6:30 a.m. until 7:30 p.m. Within the City of Columbus, where lines were longest, the 25 most heavily black precincts (averaging over 89% black voting age population) stayed open one hour and 15 minutes longer than the polls in the 25 most heavily white precincts (averaging over 99% white voting age population). To be sure, data are not available to determine the flow of voters through polling places, so it is not possible to establish from these figures whether voters in the black sites waited longer than voters in the white sites.
Allowing for problems or incidents in individual precincts (one white precinct was open for less than the 13 hours prescribed by state law), the principal cause of the difference appears to be the tendency in Franklin County for white voters to cast ballots in the morning (i.e., before work), and for black voters to cast ballots in the afternoon (i.e., after work). We have established this tendency through local contacts and through both political parties, and it accords with our considerable experience in other parts of the United States. Morning voters may wait in line several hours, as happened in white precincts, without keeping the polls open after 7:30 am; this is not the case, however, at sites where voters arrive after 5:30 p.m.
We have examined other possible factors related to the delay. The predominantly black precincts had larger numbers of newly registered, first-time voters -- 17.6% of voters in the 50 most heavily black precincts compared to 7% in the 50 most heavily white precincts. However, only 45.1% of newly registered voters in black precincts actually went to the polls, compared to 84.1% in the white precincts. There also were proportionately slightly more provisional ballots cast in black precincts (3.5% of all ballots) than in white precincts (2.2% of all ballots). The system for casting provisional ballots in Franklin County required the ballot to be cast on the machine, rather than on a separate paper ballot. As Mr. Damschroder pointed out, this was especially awkward and time-consuming in Columbus, where the ballot itself was cumbersome. Yet it does not appear that these factors were significant in terms of overall voting time.
The circumstances of the November 2 election clearly identify areas for improvement, and the Board has taken the principal step in its decision to increase the number of voting machines from 2,904 to 5,000. We commend the County for its outreach program to familiarize voters with voting machines, a program that will be expanded significantly as new voting devices are purchased. Such outreach will minimize delays in voting in future elections, as will the Board's agreement to attempt to identify precincts with voters who may require additional time in voting due to physical, educational or language barriers.
Again, thank you for the ready cooperation that both the Board and you personally extended to us during our investigation. We are equally ready to provide any assistance we can in identifying and addressing potential problems and to help you fulfill your goal of better serving the voters of Franklin County.