Break the Tie

Neither the horse race matchups between President Bush and Sen. John Kerry nor the president's job-approval ratings have moved much in the past three months. Indeed, the Associated Press/Ipsos national surveys, which pegged Bush's overall ratings at 48 percent approval 50 percent disapproval in April, held steady in May and June.

And in trial-heat polling, both Bush and Kerry have averaged about 45 percent support, with most polls falling within 3 points of that average and with independent Ralph Nader drawing 3 to 6 percent when he's included.

Not surprisingly, most observers have spotlighted the fact that the presidential race has been essentially tied for months. What they have not stressed enough is that undecided voters almost always end up breaking away from well-known, well-defined incumbents. So an incumbent who is basically tied with a challenger at 45 percent, give or take 3 points, is actually in much graver danger of losing the election than a superficial look at the numbers suggests.

What's more, recent polls showing Nader attracting 3 to 6 percent of the electorate's support are reminiscent of surveys taken before the 2000 election. That year, he received only 2.7 percent of the actual vote. Given that much of the nation views Nader as having cost Democrat Al Gore the election, and given that Nader isn't the Green Party's nominee this time, most observers expect Nader's vote total to shrink -- which would likely benefit Kerry.

In 2000, exit polls indicated that 47 percent of Nader's supporters would have voted for Gore, if Nader had not been on the ballot. Twenty-one percent would have backed Bush. And the remainder would have stayed home or voted for someone else.

Bush's strategists acknowledge that their boss's overall job-approval ratings are lower than those of Presidents Nixon, Reagan, and Clinton were at this point in their successful quests for re-election, but they point out that Bush's numbers are better than those of Presidents Ford, Carter, and Bush I, who all lost. The strategists also note that the current president's job-approval ratings are closer to those of his winning predecessors than to those of his losing ones.

Yet it is also true that a presidential election is, first and foremost, a referendum on the incumbent, not a choice between candidates. So, the incumbent's approval rating is critical.

This fact was made abundantly clear in a recent PowerPoint presentation by Doug Sosnik, who served as a top political adviser in the Clinton White House. Looking back at the re-election campaigns of Carter (1980), Reagan (1984), George H.W. Bush (1992), and Clinton (1996), Sosnik found that in the six months leading up to Election Day, each incumbent's approval ratings were "clear and unambiguous."

Sosnik says, "At this point in the previous four incumbent presidential election cycles, the American public had made its mind up about whether to re-elect the incumbent. While the horse race numbers may fluctuate, the underlying dynamics were established by now and were reflected in the final vote on Election Day."

What this tells us is that even though the 2004 presidential race appears today to be even, George W. Bush can't afford to go into November 2 tied with Kerry. The incumbent needs to be ahead, because undecideds are very likely to end up in Kerry's column and any shrinkage of the Nader vote would likely help Kerry.

In the end, the incumbent's approval rating needs to be more like 49, 50, or 51 percent, unless Nader really can draw 3 to 6 percent of the actual vote. When all the 2000 votes were counted, Gore received 48.38 percent, Bush 47.87 percent, Nader 2.7 percent, and Pat Buchanan 0.42 percent (a host of other candidates collectively got about half of a percentage point).

This time, for all the talk about same-sex marriage, prescription drug benefits, judicial nominations, and tobacco buyouts, we can expect a big-issue election in which only two issues matter: the economy and Iraq. The economy is turning around. Job creation is strong. And the Bush campaign has to hope -- make that pray -- that, as in the past, the public acknowledges a turnaround two to five months after economists spot it. More than likely, such an acknowledgment will be given to Bush in most states, although expecting voters in hard-hit Michigan, Ohio, and western Pennsylvania to reward him for the improvement is a very iffy proposition. In most of the country, though, the president should get some lift.

That leaves Iraq. Americans are now waiting on the edge of their couches to see whether the handoff (or "fumble," as MoveOn.org's advertising puts it) of sovereignty will stabilize Iraq, reduce U.S. casualties, and give hope that the United States can soon reduce its presence there. Recent polling suggests that Americans have grown decidedly more pessimistic about Iraq. And most of the foreign-policy, military, and intelligence experts I have spoken with think that, at most, the situation in Iraq has a 30 percent chance of improving before November 2.

So the $64,000 question is whether Bush will get enough of a lift in the right states from an improving economy to offset the very real possibility that the situation in Iraq will worsen before Election Day. To clinch a second term,

Bush needs his economic boost to translate into job-approval ratings that are at least a few points higher than they have been since April. But even a gain that small might not be easy to achieve.