Handicapping Long Shots and Late Shots

by Stuart Rothenberg October 19, 2006 · 12:01 AM EDT

If you make your living as a political analyst and handicapper (and the odds are very good that you don’t), you’re facing an interesting dilemma right about now. Do you add every possible House upset race that you can think of to your list of competitive races, or do you keep your list as small as possible, listing only those races that you truly believe might flip in November?

If you add longer-shot races to your list, you minimize the chance that you “miss” a race that changes party control. That’s good, since it’s always better to look brilliant than appear even a little clueless. But if you do that, you also give credence to long-shot candidates who probably won’t flip a seat. And worse, you flirt with intellectual dishonesty.

In the past, I invariably added races about which I was skeptical, preferring to have too many races on my list rather than not enough.

This year, with a Democratic wave brewing, the handicapping problem is particularly acute. Given that waves are inherently unpredictable and likely to sweep in at least one or two candidates who should have no chance of winning — look no further than Republican Steve Stockman, who toppled House Judiciary Chairman Jack Brooks (D-Texas) in 1994 — I’m tempted to add anyone who could possibly ride a wave to victory. It’s the easy way out, and I’ve always taken that route in the past.

But I’ve decided that I won’t do it this year — at least not yet. I won’t add Reps. Scott Garrett (R) in New Jersey’s 5th district, Mark Souder (R) in Indiana’s 3rd, Jim Ryun (R) in Kansas’ 2nd, Cathy McMorris (R) in Washington’s 5th or other potentially threatened Republicans just because their Democratic opponents have an allegedly encouraging poll or are up with TV ads. While I can’t completely rule out the possibility that some or all of them could drown in a Democratic tsunami, I believe the chances are so small that I can’t bring myself to put them on a list of endangered incumbents. As I said, at least not yet.

Take Garrett. His opponent, Paul Aronsohn, distributed a polling memo, based on an initial March survey and a late-September poll, that alleged his race against Garrett was “tightening up” and that a Democratic victory was “very possible.”

But New Jersey’s 5th district, represented by Garrett, is a Republican-leaning district that gave President Bush 57 percent of the vote in 2004. In 2002, Garrett drew almost 60 percent in an open-seat victory against a credible Democrat who outspent him, and two years later he won re-election with almost 59 percent.

Aronsohn’s own polling shows the Republican generic vote advantage has been cut in half, but it still stands at 7 points — a significant edge. Then there is the trial heat, which shows Garrett’s 25-point advantage in March having closed to 16 points in late September. But those surveys also show that Garrett’s share of the ballot test has remained unchanged at 49 percent of the vote.

Even though the Republican “generic” number has slipped, and even though the “wrong track” number and President Bush’s job approval rating have eroded, Garrett’s trial heat number has remained the same, just a whisker under the 50 percent mark.

Sure, the ballot test in the poll has tightened, but only because some of the undecided voters went over to Aronsohn, whose share of the vote went from an anemic 24 percent in March to an only somewhat less anemic 33 percent at the end of September.

I can’t say with absolute certainty that Aronsohn can’t, or won’t, beat Garrett next month, only that the chances seem so small that rating the race as competitive seems misleading. If, barring other evidence of an Aronsohn surge before Election Day, the Democrat does win, it will be a jaw-dropping upset, and I won’t mind missing it.

The situation is a bit different, but only a bit, in Colorado’s 5th district, where a media poll shows a dead heat between Democrat Jay Fawcett and Republican Doug Lamborn for the open seat of retiring GOP Rep. Joel Hefley.

This is a solid Republican district in which Bush drew 66 percent in 2004 and 63 percent in 2000. Recent party registration figures show an advantage for the GOP of 100,000 registered voters — 189,942 Republicans (46 percent), 89,182 Democrats (22 percent) and 132,606 independents (32 percent).

Like New Jersey’s 5th district, one public poll alleges that the race is competitive, but here it’s an independent Mason-Dixon survey. Moreover, Republicans had a bitter primary and have not united since. All of those circumstances argue for including the race on a list of competitive contests, at least for the moment.

But if you are a Democratic partisan looking for a way for Democrats to net 40 or 50 seats, don’t kid yourself. The fundamentals of Colorado’s 5th are so terrible for Democrats that it’s hard to believe Fawcett will win. Lamborn certainly created enemies within his own party, but he can afford to lose a bunch of them, given his edge in party registration.

For Democrats, the more interesting long shots may include Rep. Melissa Hart (R-Pa.), whose 4th district gave Bush 54 percent of the vote in 2004 and 52 percent four years earlier. Or Rep. Gil Gutknecht (R), who represents a competitive district in southeastern Minnesota. A true national wave could hit hard in districts like those, creating opportunities that Democrats ought not have in a normal year. That’s why both of these races now are on my list of Democratic challenges to watch.

I expect a major surprise or two on election night, with an unheralded Democrat or two pulling off an upset, possibly in a rock-ribbed Republican district. I’ll try not to miss that race, but I won’t be embarrassed if I do.