My Biggest Concern, Allayed Slightly

My biggest concern about this year’s Presidential election grows out of Minnesota’s experience six years ago, when Jesse Ventura was elected Governor. The polls failed to foresee Jesse’s victory. What happened was that during the last week or two of the campaign, electing the pro wrestler became a fad among normally-apolitical Minnesotans, especially young men. A wave of previously unregistered first-time voters swept Ventura to victory in a three-candidate field.
Could something similar happen next month on a national scale? The media are reporting large numbers of first-time registrations. The Democrats have made huge efforts both to register young voters, and to panic them with false rumors about the draft. I’ve seen data suggesting that Kerry may have as much as a twenty-point lead among these newly-registered young voters. So my fear is that the current polls, especially “likely voter” polls, could be missing a groundswell for Kerry among new voters, especially young voters, who would not normally be considered “likely.” Of course, as a practical matter it is hard to separate new but legitimate first-time registrants from the outright voter fraud that the Democrats will also perpetrate.
This morning’s Minneapolis Star Tribune contains some data that might support my concern, albeit rather weakly:

Fueled by the passions of the presidential race, voter-registration campaigns have added tens of thousands of Minnesotans to the election rolls from predominantly Democratic cities, nearly twice the new registrations gathered in recent months from heavily Republican areas of the state.
Since June, there have been more than 28,000 new registrations in Minnesota cities that voted decisively for Al Gore in the last presidential election and more than 16,000 registrations in cities that voted decisively for George W. Bush, a Star Tribune analysis of state records found. The newspaper defined a decisive margin as 10 percentage points or more.
In all, more than 72,000 new registrations (not counting people who moved or changed their names) were filed between June 1 and Oct. 5, before the final push for preregistration efforts. About 2.5 million Minnesotans voted in the 2000 election, meaning those new registrants, if all of them voted, could constitute 2 percent to 3 percent of the electorate.

The numbers the Strib is talking about are not as large as I might have feared, but of course even these modest numbers could be decisive if the election is close enough.
A dissenting view comes from blogger Jay Cost, who has posted a letter to the Strib on his site. Jay writes:

You note that Gore counties have gained more registered voters between June, 2004 and October, 2004 than Bush counties. However, you fail to note that, since the 2000 General Election, the registered voter rolls in those Gore counties have declined at a much greater rate than the registered voter rolls in those Bush counties.
Gore won 8 counties by 10% or more in 2000…As of 11-07-00, those counties had 1,319,026 registered voters. As of 10-13-04, those counties had 1,157,187. In other words, they had lost 161,839 registered voters.
Of all the counties that Gore won in 2000, there are 176,933 fewer registered voters.
Bush won 25 counties by 10% or more…. As of 11-07-00, those counties had 483,617 registered voters. As of 10-13-04, those counties had 456,868 registered voters. In other words, those counties lost 26,749 registered voters.
Of all the counties that Bush won in 2000, there are 153,965 fewer registered voters.
Thus, as of October 13, 2004 — those 8 strong Gore counties had lost 135,090 more registered voters than those 25 strong Bush counties. All Gore counties have lost 22,968 more registered voters than all Bush counties.

Interesting. If this is right, the state as a whole has lost approximately 325,000 registered voters since 2000. I don’t know of any reason why that would be true. I do think, however, that demographic trends favoring Bush in states like Minnesota and Wisconsin could outweigh whatever advantage in new registrants the Dems may have. Of course, those demographic trends are presumably reflected in the polls, while the impact of first-time voters might not be.