Post by Ramen on Nov 1, 2014 19:41:49 GMT
Next week is going to be insanely busy, so this might be the last time that I get to comment in a post like this.
The best forecast for the House is that Republicans will pick up 7-15 seats. More than 15 seats would surprise me, but honestly I haven't seen any forecasts in the past month so things might have changed. Obama's approvals have slipped a point or so and the "generic matchup" polling question about control of the House has shifted a point in the GOP's favor. But nothing else seems to have changed.
For the Senate, without rehashing everything from the other thread, I'm going to simply start counting down the close races to get an estimate of how many seats Republicans are likely to hold. I think that everybody can agree with TheUpshot's take that there are 45 likely Democratic seats and 48 likely Republican seats (if you include LA and AR as likely Republican):
www.nytimes.com/newsgraphics/2014/senate-model/
That leaves seven potential close seats - NH, NC, KS, GA, IA, AK, and CO. Republicans only need three of those to gain control of the Senate.
I think we can safely move NH and NC to likely Democratic wins. Hagan in NC has held a consistent 2% lead in the polls. Shaheen has the same lead, but Scott Brown has been narrowing the gap. Still, it's not very likely that Brown comes out on top. So that brings the total to 47 Democratic seats that are likely.
I also think that we can safely move AK into the likely Republican column. Sullivan has held a consistent 3% lead in the polls for over a month now and nothing in that race shows signs of turning. Yeah, Begich is an incumbent, but he's in a deep red state during a year where Democrats are very unpopular. He'll probably lose. That's 49 seats for Republicans.
Georgia...probably Republican, but the polls are a lot closer than I would have thought. Half the recent polls show Perdue with a 2-3% lead (I'm discounting the Monmouth University poll as an outlier) and half show a tied race. That suggests a true Perdue lead, so I think the most likely outcome is a runoff where Perdue wins. That's 50 seats for Republicans.
Iowa is tricky as hell. The Republican has a 2% lead, but that race has been close all year and the latest polls are all over the map, with some showing Rep lead and others showing Dem lead. That one is a true toss up in my view.
Colorado is the same. The Republican has a 2% lead in the polls and has kept that lead for the past month and a half. That alone would normally be enough for me to put it in the GOP camp. However, it's Colorado. Three of the last four elections, the Democrats have outperformed the polls by about 2-3%. Bennett in 2010 was behind by 3% and he squeezed out a narrow win due to superior ground game. Same for Obama in 2012 - polls showed a tie and Obama won by 3-4%. On the one hand, I *know* that Democrats are running their unusually sophisticated turnout program (the "Bannock Street Project" named after the location of an office in Denver). On the other hand, I've seen some of what the Republicans are doing with turnout in Colorado, and it's as sophisticated and effective as anything I've seen. On the other other hand, this year Colorado is using all-mail ballots, which should increase turnout among lower income and minority groups, which could help Democrats.
If the polls are actually picking up the people who are most likely to vote, then Republicans win. What's unknown is if the polls are once again off by a few points. If Democrats are converting people who otherwise wouldn't vote into voters, but those people are not making it through the polling firms' likely voter screens, then Democrats have a shot.
Finally, I also put Kansas as a true toss-up. The independent is ahead slightly, but that's only a recent trend. The New York Times forecasting model has it as a 60% chance that the Republicans lose that seat, but the margin of error on the forecast is pretty wide. Who the hell knows with that one?
So best guess at this point is that Republicans are most likely to have at least 50 seats with 3 races too close to predict. A majority of 53 seats is very possible, as is a 50-50 split. That said, I think that it's unlikely that Democrats sweep all three of the toss-up elections, which means that the most likely scenario now is a narrow Republican majority in the Senate. It would take a shocker in Georgia or Alaska plus a full sweep of the three toss-ups for Democrats to retain the majority. I just don't view that as very likely.
What is known for sure is that the Senate elections this year are among the most interesting in recent history. At least for nerds like me.
The best forecast for the House is that Republicans will pick up 7-15 seats. More than 15 seats would surprise me, but honestly I haven't seen any forecasts in the past month so things might have changed. Obama's approvals have slipped a point or so and the "generic matchup" polling question about control of the House has shifted a point in the GOP's favor. But nothing else seems to have changed.
For the Senate, without rehashing everything from the other thread, I'm going to simply start counting down the close races to get an estimate of how many seats Republicans are likely to hold. I think that everybody can agree with TheUpshot's take that there are 45 likely Democratic seats and 48 likely Republican seats (if you include LA and AR as likely Republican):
www.nytimes.com/newsgraphics/2014/senate-model/
That leaves seven potential close seats - NH, NC, KS, GA, IA, AK, and CO. Republicans only need three of those to gain control of the Senate.
I think we can safely move NH and NC to likely Democratic wins. Hagan in NC has held a consistent 2% lead in the polls. Shaheen has the same lead, but Scott Brown has been narrowing the gap. Still, it's not very likely that Brown comes out on top. So that brings the total to 47 Democratic seats that are likely.
I also think that we can safely move AK into the likely Republican column. Sullivan has held a consistent 3% lead in the polls for over a month now and nothing in that race shows signs of turning. Yeah, Begich is an incumbent, but he's in a deep red state during a year where Democrats are very unpopular. He'll probably lose. That's 49 seats for Republicans.
Georgia...probably Republican, but the polls are a lot closer than I would have thought. Half the recent polls show Perdue with a 2-3% lead (I'm discounting the Monmouth University poll as an outlier) and half show a tied race. That suggests a true Perdue lead, so I think the most likely outcome is a runoff where Perdue wins. That's 50 seats for Republicans.
Iowa is tricky as hell. The Republican has a 2% lead, but that race has been close all year and the latest polls are all over the map, with some showing Rep lead and others showing Dem lead. That one is a true toss up in my view.
Colorado is the same. The Republican has a 2% lead in the polls and has kept that lead for the past month and a half. That alone would normally be enough for me to put it in the GOP camp. However, it's Colorado. Three of the last four elections, the Democrats have outperformed the polls by about 2-3%. Bennett in 2010 was behind by 3% and he squeezed out a narrow win due to superior ground game. Same for Obama in 2012 - polls showed a tie and Obama won by 3-4%. On the one hand, I *know* that Democrats are running their unusually sophisticated turnout program (the "Bannock Street Project" named after the location of an office in Denver). On the other hand, I've seen some of what the Republicans are doing with turnout in Colorado, and it's as sophisticated and effective as anything I've seen. On the other other hand, this year Colorado is using all-mail ballots, which should increase turnout among lower income and minority groups, which could help Democrats.
If the polls are actually picking up the people who are most likely to vote, then Republicans win. What's unknown is if the polls are once again off by a few points. If Democrats are converting people who otherwise wouldn't vote into voters, but those people are not making it through the polling firms' likely voter screens, then Democrats have a shot.
Finally, I also put Kansas as a true toss-up. The independent is ahead slightly, but that's only a recent trend. The New York Times forecasting model has it as a 60% chance that the Republicans lose that seat, but the margin of error on the forecast is pretty wide. Who the hell knows with that one?
So best guess at this point is that Republicans are most likely to have at least 50 seats with 3 races too close to predict. A majority of 53 seats is very possible, as is a 50-50 split. That said, I think that it's unlikely that Democrats sweep all three of the toss-up elections, which means that the most likely scenario now is a narrow Republican majority in the Senate. It would take a shocker in Georgia or Alaska plus a full sweep of the three toss-ups for Democrats to retain the majority. I just don't view that as very likely.
What is known for sure is that the Senate elections this year are among the most interesting in recent history. At least for nerds like me.
