Unraveling The Presidential Horse Race

By far the most common criticism of news media coverage of American elections is that it's way too focused on the horse race - on predicting the winner. By far the most common question civilians ask political reporters, is, "Who do you think is going to win in November?"

So critics be damned, let's talk horse race before all eyes turn to those substance-fests known as the party conventions.

Besides gut instincts, there are several analytic perches used to forecast elections three months before they happen: national public opinion surveys, the state-by-state electoral map and academic prediction models.

The Polls

An ironclad rule of modern political marketing states that candidates get a temporary lift in the polls after their party conventions. It's called "the bounce."

George W. Bush has received a highly unusual pre-convention bounce.

Two national polls, one by Gallup for USA Today/CNN and the other by ABC News/The Washington Post, show that Bush's lead in the polls grew to 11 points on the eve of the Philadelphia convention. The most recent CBS News polls show Bush leading by between 4 to 6 percentage points.

Survey researchers caution that polls taken before the conventions don't mean much. Voters haven't focused yet. Okay, but 11 points for a candidate who has led in the polls almost all year isn't exactly public opinion chopped liver.

But, to put that in some perspective, recall that Bush the Elder led Bill Clinton by just two points going into the Democratic convention in July of 1992 and that after the convention, Clinton led by 10 points and never trailed again.

In 1988, Massachusetts Governor Michael Dukakis led then Vice President Bush before the Republican convention, which came second that year, by as much as 17 points. That lead vanished after the GOP convention and never returned.

So conventions matter, or at least seismic changes in polls tend to take place around convention time. It also seems that some time around the conventions, the polls settle down and reverses later in the campaign have become extremely rare. It seems we're seeing Bush's seismic bounce already. If Gore's doesn't come fast and big on the heels of the GOP convention, he'll be in dire straits.

The Map

Ten years ago, political scientists talked about a Republican "lock" on the electoral map. From 1968 through 1988, a winning block of states in the Far West, Rocky Mountains, Plains, Midwest, and Rust Belt went consistently Republican. And except for 1976, when Jimmy Carter won, the South also was predictably Republican.

Clinton changed all that.

Now the only hard pattern to be divined from the electoral map is that the Deep South and Plains remain Republican. Washington and Oregon seem to have become Democratic bastions along with Minnesota, Rhode Island, Massachusetts, Hawaii and Washington, D.C. Advantage Bush.

That advantage is apparent when you add up the pols taken in the individual states so far this year. In polls taken since March, Bush leads in 30 states that provide 342 electoral votes, according to the Hotline. Gore leads in six states that represent 120 electoral votes. The winner needs 270 electoral votes.

So at this point in the race, snapshots of the states buttress Bush's advantage in national polls and there is nothing in the electoral history of the individual states that warns us not to trust the current portrait.

The Eggheads

If the Gore camp wants some handicapping solace, they'd best look toward academia. Political scientists plug "macro" variables taken from economic statistics and public opinion research (minus the horse race polls) into elaborate mathematical models that actually have excellent track records in forecasting presidential elections.

And all the models say Gore will win, easily. "It's not even going to be close," University of Iowa prognosticator Michael Lewis-Beck told The Washington Post as early as May.

The basic recipe is pretty simple: the economy is booming, the country is at peace, and the voters have a favorable view of the incumbent president. Punch those variables into the computer models and Gore wins.

The ivory tower, however, isn't giving comforting Gore-ites much. As one of his top political numbers crunchers said bluntly, "There isn't a variable for an impeachment trial."

A few other factors deserve at least a mention in the racing form. Bush's core staff has remained unchanged throughout the campaign, while Gore is on his third team. A recent study by the Pew Research Center and Project for Excellence in Journalism finds that Bush gets treated marginally better by the press. The Republican primaries, thanks to John McCain, were exciting; the Democratic race was boring. Green party candidate Ralph Nader may nibble at Gore's support in a few swing states.

The betting counter is open.