If Michael Bloomberg runs for president as an independent or somehow on the Unity '08 ticket, most of the political class has convinced itself that Bloomberg would steal more votes from Republicans than from Democrats.

They base this conclusion on one supposition and one fact. The supposition is that Bloomberg won't run if an independent-minded Democrat, or a Democrat who attracts independents like Barack Obama (or maybe Bill RIchardson) does, gets the nomination. Also -- he won't run if an independent-minded Republican -- John McCain, gets the GOP nod. It's unclear just how Bloomberg regards his mayoral predecessor, Rudy Giuliani, right now. So -- if the supposition holds, then Bloomberg would only run if the Democrats select a polarizing Democrat, the Republicans select a cultural conservative, and the level of discourse between the main candidates reaches a new low.

The fact is that in Southern states like Mississippi, South Carolina, Kentucky and Georgia, Democrats have a solid, unbreakable base of black voters that approaches 35% of the general election voter pool. They point out that Ross Perot, who won nearly 19% percent of the popular vote in 1992, stole more votes from George H.W. Bush than he did from Bill Clinton. That conclusion is supported by the political science literature.

And if the exit polls were correct, Perot managed to sap as many liberal independents from Democrats as he did conservative independents from Republicans. But Bill Clinton's base was solid enough in the South; more solid than the Republican base. Perot's other demographic strongholds included younger voters, unmarried Midwesterners, rural voters, evangelicals and protestants;

But there are at least three reasons to discount the argument from history.

One is very obvious: 1992 resembles 2008 in only one real way: change-agents have an edge. But Perot appealed to downscale Reagan Democrats, to economic populists, to opponents of NAFTA and free trade. He tapped into a vein of populist discontent with both parties, but particularly those branches than ran through Republican arteries once Bush abandoned his no-tax pledge. His technocratic appeal -- "I'll do what works; I'll balance the budget" -- stole a page from the GOP. The foreign policy debate was mostly about the peace dividend; it did not, as it does now, overshadow the domestic political environment.

Bloomberg wants something different. He's hinted at broad, programmatic solutions to intractable problems. He is, at the same time, a cultural liberal with an anti-libertarian streak; ask any New Yorker who smokes, or who enjoyed their trans-fats cheesburgers whether Bloomberg left them alone. Health care, education, the environment -- Democrats have an enormous edge right in attracting voters who care about these issues. If Bloomberg somehow offers an alternative, it's not hard to suppose that many Dem-leaning independents will be attracted to his candidacy.

Point two: independents are overcommitted to Democrats right now, giving them as much as 70 percent of their generic presidential support in certain surveys. A conservative Republican candidate wouldn't attract too many independents anyway, in theory, and Democrats have many independents to spare, and lose. Why assume that Bloomberg wouldn't take more votes from the Democratic candidate?

Point three: A Bloomberg candidacy could hurt Democrats the most in one very reliable (at least recently) must-win state: California. The more independents he steals, the more he hurts the Democratic candidate there. Also: Bloomberg might steal Dem-leaning independents in Southern states, states where Perot polled strongly among cultural conservatives and economic conservatives.

Now -- here are some counter-arguments to my arguments.

One is that Bloomberg could help deliver New York to the Democrats even if Rudy Giuliani wins the GOP nomination. In a quirk of politics, Bloomberg, who is more popular in New York City right now than Hillary Clinton is, could outpoll Giuliani at home. Republicans would need an enormous margin upstate to win, a margin they've never been able to produce. By contrast, Clinton's electoral base in New York state seems solid, a result of her close attention to the care-and-feeding side of her job.

Also unclear is what Bloomberg will say about Iraq, which is and will be the sine qua non of the 2008 race. If he runs without clearly distinguishing himself here -- and Perot, in 1992, knew it was the economy, stupid, too -- then it's hard to predict whether Bloomberg will be enough of a force to be taken seriously by voters in the end.

We want to hear what you think. Submit a letter to the editor or write to letters@theatlantic.com.