Bush Would Have Won! Or Would He?

Monday was supposed to be a really big day for the Miami Herald. After two months of squinting and peering at half-perforated and slightly dented ballots, a band of ballot-counters hired by the newspaper to comb through Dade County's so-called "undervotes" were ready to hand over a verdict.

This was heady stuff; the newspaper set up a stand-alone section of its web site and prepared for a frenzy of activity. Would the recount results fuel Democrats' conviction that the election was stolen by a partisan U.S. Supreme Court? Or would the new numbers establish unequivocally that George W. Bush won the presidency?

Perhaps, the optimists thought, the partisan bickering over legitimacy would end once the "real results" were out in the open.

Then Monday dawned and the paper hit the stands. The news was bad for the optimists, and, to a lesser extent, for Al Gore. After two months of work, the data were conflicting at best.

First, the bad news for Gore: In a review of 10,644 undervotes, the paper reported, the vice president gained only 49 votes (1,555 to Bush's 1,506). Those votes, combined with the recount numbers Gore requested from three other counties — Volusia, Palm Beach and Broward — and subtracted from Bush's 730-vote lead, would still have left Gore 140 votes behind Bush in the overall Florida count.

Of course, Bush partisans crowed over the numbers, which came, after all, from what was supposed to be a Gore stronghold. But there were also new points of contention, all of which will likely be advanced enthusiastically in the coming weeks by Democratic loyalists. While the recount, conducted by accountants BDO Siedman for the Herald, used the most generous definition of "vote" to tally numbers, the new numbers, they reminded anyone who'd listen, includes only undervotes — those ballots whose chads were not fully detached.

One thing the new numbers do confirm is what is now almost universally acknowledged: that in calling for recounts in specific counties, the Gore camp made a strategic blunder. Apart from fueling public suspicion that he was cherry-picking counties apparently most beneficial to his cause — a tactic that probably fueled already vociferous opposition — it also appears that he had little to gain in terms of votes by narrowing the recounts to so few counties.

The rest, however, is not so clear-cut. The recount, we are reminded, did not include the 1,840 ballots where voters cleanly punched holes assigned to no one, including 736 punched in the hole directly beneath Bush's name and 1,017 beneath Gore's. This phantom margin, apparently created by confused or distracted or perhaps just not very bright voters, would be, of course, enough to secure a Gore victory. If, that is, we were in the habit of counting votes next to or beneath or slightly to the left of where voters are supposed to register their vote — which we are not.

And it is perhaps in those floating, unassigned hole punches that any real agreement over this election is likely to result. It seems inconceivable that anyone, upon learning that fully 2,000 votes were firmly cast (in one county!) for no candidate at all, could argue against revising our balloting systems. Those votes, which will taunt Al Gore and 2,000 apparently disoriented Dade County residents, are a clarion call for ballot reform.

After all, we practically beg people to vote. Why, then, do we design ballots that manage to confuse enough voters that the outcome of a presidential election could be in doubt?

As for a more complete view of the Florida vote, that will have to wait until two separate counts of the whole state are completed. The Herald and its parent, the Knight Ridder chain of newspapers, has counted all but two counties. (Officials in upstate Duval and Holmes counties have postponed the recount, fearing further disruption if the ballots were subpoenaed in lawsuits.) Meanwhile, a consortium of news organizations, including the Associated Press, the New York Times and CNN, has hired the National Opinion Research Center, a non-profit firm out of the University of Chicago, to examine nearly 200,000 ballots that did not register any vote at all, including ballots where no vote was clearly marked, and where more than one vote was clearly marked. The NORC report is due out in April.