For Investors, an Election Day Toss-Up; Arends: The presidential election is much closer than the polls suggest, according to one famed pollster.

Investors have a financial interest in the presidential election.

Barack Obama and Mitt Romney offer different visions on taxes, domestic spending, and the Pentagon.

Also See

• Financial Advice for Presidential Candidates

The president has already said he wants to let the Bush-era tax cuts expire on incomes above about $250,000 a year. That would mean higher taxes on stock dividends and long-term capital gains for the highest-earning investors.

Mitt Romney wants to extend the Bush tax cuts above $250,000 as well. He also wants to slash taxes further and to ramp up defense spending - while spending less domestically than the president. If this would lead to higher deficits, and hence to a Keynesian reflation of the economy, then it would be better for stocks than for bonds.

So who's going to win?

Last week I sat down for lunch with David Paleologos, pollster at Suffolk University in Boston. Paleologos may be the best pollster working in America today. He famously called the 2008 New Hampshire primary for Hillary Clinton, when everyone else said Obama would win, and he was the first to predict Republican Scott Brown's January, 2010 upset win in the Massachusetts senate race. Paleologos is not skewed to either side.

His take now?

You should take these national polls, showing Barack Obama winning by six or seven points, with big fistfuls of salt. The November election could come down to the wire, Paleologos says. And Obama is much more vulnerable in the handful of swing states than most national commentators think.

"I think Romney's going to win Ohio and Florida," he says. Yes, polls show Obama ahead in both states, but most commentators don't understand the numbers, he adds/ In each case Obama's stuck at around 47%. And that's bad, bad news for an incumbent, because undecideds tend to break for the challenger. "If you're getting under 50% as an incumbent, you're vulnerable," says Paleologos. At 47% or below the numbers are grim.

On the other hand, he sees Obama winning Wisconsin and Pennsylvania, where he's at around 50% in recent polls.

Paleologos thinks the race will come down to six states: Colorado, Iowa, Michigan, Virginia, Nevada and New Hampshire. They're all too close to call. In each one, Obama's on around 48% to 49% - vulnerable, he says, but just on the cusp.

The big prize this year may be Virginia. "The whole presidential election might come down to four counties in northern Virginia," says Paleologos."If Obama can win 100,000 votes in northern Virginia, that could tip Virginia, and Virginia could tip the whole thing."

The debate is raging about Romney's choice for vice-president. Paleologos thinks Romney's best bet would be to pick a Hispanic or a woman, to give the ticket the biggest boost. That could mean Florida senator Marco Rubio, New Mexico governor Susana Martinez, or even New Hampshire senator Kelly Ayotte.

(Our conversation was just before the brief flurry of speculation in Condeleezza Rice.)

Oddly enough, according to Paleologos the more important choice may be Obama's. There was talk six months ago of swapping Vice-President Joe Biden and Secretary of State Hillary Clinton. It appears to have died down. For Democrats, that's a pity. "If Obama picked Hillary," says Paleologos, "the race would be over. She polls 10 to 15 points higher than Obama in the swing states."

Markets in the past haven't shown any preference for Democrats or Republicans in the White House. They do, however, get jittery about uncertainty. If this is a close race that could produce some bumps between now and November.