The ballots are missing, so start finding answers   (FL)

Palm Beach Post Editorial  03 September 2008

The first rule of election recounts should be to start - and finish - with the same number of ballots. Palm Beach County has failed that basic test.

As a result, there's no way to know for sure who won the Group 23 judicial race between incumbent Richard Wennet and challenger William Abramson. As a result, there's no way to be confident about the November election. The variations between election day tallies, then a machine recount and then a hand recount are so wide as to reach a whole new level of concern for voters who thought they had seen everything in Palm Beach County.

Start with the missing votes - 3,478 of them, or 3.4 percent of the election day total. Then there's the 5 percent in the number of valid votes after the machine recount. How can candidates in other races accept a 1 percent or 2 percent loss - too wide a margin under state law to merit a recount - when the recount machinery rejects without explanation 5 percent of the votes?

Supervisor of Elections Arthur Anderson, who finished last in a three-way race for reelection, is breaking in a new voting system and learning, literally, as he goes. He'll be in charge in November, when voters pick his successor as well as the next president. Turnout could be five or six times greater than last week, which would stress a system that already seems too fragile.

Dr. Anderson's staff worked overnight Friday and Saturday to complete both the machine and hand recounts before a 5 p.m. Tuesday deadline for certifying the result to the state. But could workers have met the deadline if they were counting 600,000 ballots? How many would they lose? How can the public trust that all the votes will be counted correctly in the first place if the machines have such variability? How could Dr. Anderson and the county canvassing board certify vote totals they knew to be wrong?

Mr. Abramson led Judge Wennet by 17 votes out of 102,523 cast after voting ended eight days ago. The ballots were fed through counting machines on Friday night and Saturday - different machines than those in polling places - and the number of votes judged to be invalid grew by nearly 5,000. But many of those votes were not invalid at all, as election workers studying the ballots individually found in redeeming nearly 2,000 of them. In the end, the county certified a 60-vote victory for Judge Wennet.

State law requires hand recounts only in races where the difference is 0.25 percentage points or less after the machine recount, which is triggered when the margin is 0.5 percentage points or less. Had the difference been slightly greater in the Wennet-Abramson race, those redeemed votes never would have been redeemed. Somehow, the recount machines had a higher standard for what constituted a vote. Why?

Palm Beach County wasn't the only county to do a recount, and it wasn't the only one to lose votes. But the level of discrepancy stood out. In Broward County, recounting a similar number of votes, officials found 236 fewer valid votes in the machine recount, compared with 4,884 in Palm Beach. Broward couldn't account for about 100 votes, compared with 3,478 in Palm Beach.

There is no simple explanation for Palm Beach County's confusion. Dr. Anderson's spokeswoman warned not to expect answers before the end of the week. But those will be answers from the people who made the mistakes. While state law doesn't authorize intervention, Dr. Anderson has to seek help, starting with Secretary of State Kurt Browning, a former Pasco County elections supervisor.

For Mr. Abramson, the county's explanations will be too late. He'll surely sue. That's one way to get answers. The better way would be for Dr. Anderson to realize that the public can't wait for a lawsuit. The general election is 62 days away. He must provide answers, and quickly.