Not Perfect, But Close: My 2012 Predictions and Results

by | Nov 7, 2012 | Election Analysis

Not to throw my arm out of whack patting myself on the back, but I had a pretty good Election Night 2012.

As of today, I have correctly called all 49 of the states that have determined a winner in the presidential race. It looks like I am probably going to miss Florida, which I predicted Mitt Romney would win, but where President Barack Obama leads by about 40,000 votes with about 3 percent of the votes still to be counted, per CNN. Still, even 49 out of 50, if that’s where it ends up, puts me in very good company. Larry Sabato and his team at the University of Virginia Center for Politics—unquestionably among the best in the business—got 48 states right this year and, like me, picked Romney in Florida.

I’m not on the level of Nate Silver just yet—if Florida does go for Obama, as seems likely, he’ll be 50-for-50, and I tip my hat to him. He’s got an excellent system, far more scientific than mine, and he’s either 98 or 99 for 100 the past two presidential elections, depending on the final results in Florida. You can’t argue with those kinds of results (though many tried this year, and ended up with a plateful of crow).

I didn’t get every Senate race right—I missed Montana and North Dakota, as well as my prediction that Democrats would win either Nevada or Arizona, and my perception that Bob Kerrey had the momentum in Nebraska turned out to be badly off the mark. But in terms of the final spread, I came as close as a guy can get without actually being perfect: I said the Democrats would end up with a 54-46 majority in the Senate (counting independents Bernie Sanders of Vermont, who caucuses with Democrats, and Angus King of Maine, who is presumed likely to caucus with Democrats). Because Heidi Heitkamp eked out a victory by less than 3,000 votes in North Dakota, I missed the final total of 55-45 by a single seat—as I did in 2006 and 2008.

In the House, CNN reports that there are still eight seats that are too close to call. Currently the Republicans lead 233-194. If the Democrats hang on to win the six seats in which they lead, and the Republicans hold the two seats in which they lead, that will put it at 235-200—exactly the number I predicted last week, as recorded twice on this website in advance of the election.

I admit that I approach political prognostication as more of an art than a science. For much of my adult life, I have extensively studied statistical, anecdotal and historical data about the politics of the various states and their Congressional districts, and I follow polling obsessively. I try to put my political knowledge and my powers of observation together and arrive at the right result, and despite my own political leanings, I strive to be scrupulously objective in making election predictions.

Since 2006, I’ve come extremely close in three elections out of four. I blew it in 2010 (missing the House by 10 seats and the Senate by three), but I wasn’t alone. It was an odd year. In three elections out of four, my average House prediction has missed by two seats (2, 4 and 0, if the 2012 totals hold up), and my Senate prediction by exactly one seat each cycle. (Interestingly enough, in all four elections since 2006, I overcalled the Democratic House total two times, but I undercalled the Democratic Senate total in each case.)

I take this political prediction business seriously and do my best to be a credible source. One thing is for sure: there weren’t many this year who had a better night.

Time To Call The Presidential Race: It's Obama

by | Nov 4, 2012 | Election Analysis

I had planned to wait until Monday night to make my presidential prediction, but the way the polls have broken over the last week, President Barack Obama is a clear favorite to win reelection on Tuesday. Only a massive, last-minute reversal on Monday, or the unlikely and unprecedented occurrence of virtually all reputable polling being dead wrong, can alter this result.

For most of this year, I have expected the president to win reelection with 303 electoral votes to Mitt Romney’s 235, with Romney reclaiming Indiana, North Carolina and Florida for the Republicans, and Obama holding all the other states he won in 2008. As his fortunes have improved or declined, the number has gone up a bit or down a bit, but I seem to always come back to that total. With some hesitance about one state—Colorado—I am going to go with the number I’ve been predicting all along: Obama 303, Romney 235.

I arrive at this number by examining 12 states that Obama carried in 2008, 11 of which appear to be close in 2012.

Let’s start with the one state, among this dandy dozen, that will not be close: Indiana will not go into the Obama column again. This has been clear for months, if not a couple of years, and the Obama campaign wrote the Hoosier State off a long time ago. Its unexpected flip into the Democratic column in 2008 was a fluke, the by-product of a meaningful Democratic primary there, a bad year for Republicans and off-the-charts Democratic performance among young people and minorities.

Less certain are Romney’s bids to reclaim North Carolina and Florida. While it is likely he will slip by in North Carolina, Florida now appears to be up for grabs. It’s a tough call, but I think the dynamics favor Romney in Florida by a very slim margin.

Among the other swing states that I have largely expected Obama to hang onto, almost nobody is giving Romney a chance to win Nevada, and I concur. Obama has consistently led by small but clear margins, and his expected landslide among Hispanic voters seems to cement Nevada in his camp.

Although Romney’s team has made a desperate effort at the end to flip Pennsylvania and Michigan, there is no credible polling evidence that he will win either state. In my opinion, a Pittsburgh Tribune-Review poll today showing a 47-47 tie in Pennsylvania is not credible. The newspaper is owned by Richard Mellon Scaife, a conservative activist, and the poll is at odds with all other reputable polling.

In five of the remaining six states, the last week has seen clear movement in Obama’s direction, and the polling shifts in Ohio and Virginia appear decisive. Less certain, but unquestionably also shifting in Obama’s direction, are Wisconsin, Iowa and New Hampshire. If the preponderance of the polls and the polling trends are right, and I believe they are, Obama wins them all.

Tonight, the only state I don’t have a clear feel for is Colorado. By the looks of things, the president should hang onto it. The state has many Hispanic voters and many young, educated urban voters, and that’s in his favor. But he seems to have a problem in the suburbs, and the polling data from Colorado have been inconclusive. While the other swing states have all seen clear, even dramatic movement toward Obama in the last few days, Colorado is still close and doesn’t appear to be budging. I really feel that Colorado—along with perhaps Florida—is a coin flip, and candor compels me to admit that I can’t really call it.

If Romney wins Colorado, and everything else goes as expected, Obama would win 294-244 in the electoral college. If Obama wins Colorado and all the other states I project he will win, it’s 303-235. Because Colorado does appear to be a coin flip, I’m going to say that the Obama organization and the Hispanic vote will ultimately save the state for the president, but not by much.

The popular vote is a tougher call. Romney is going to absolutely slaughter Obama in most of the South, but then, Obama is going to rack up big margins over Romney in some of the nation’s largest states, such as California, New York and Illinois. I expect we will see third-party candidates such as Gary Johnson, Virgil Goode and others combine for about 2 percent of the total votes, perhaps slightly more, which is roughly in keeping with what normally happens. Based on the trajectory of the final national polls, all of which show Obama tied or slightly ahead by 1-3 points—and all of which have moved in Obama’s direction in the last week—I figure on Obama ending up somewhere between 49.5 and 50 percent, and Romney somewhere between 48 and 48.5 percent. Let’s call it at Obama 49.7, Romney 48.3.

So, to sum up:
Obama to win the electoral vote, 303-235, and the popular vote, 49.7-48.3
Democrats to retain control of the Senate, 54-46 (Democrats +1)
Republicans to retain control of the House, 235-200 (Democrats +7)

Election 2012 Predictions: Congress

by | Oct 31, 2012 | Election Analysis

In six years of electoral prognostications, I’ve done pretty well when it comes to calling the final tally in Congress. Over three election cycles, I have called the final composition of the U.S. House of Representatives within 5.3 seats, and the U.S. Senate within 1.7 seats.

I didn’t do as well as I’d have liked last time out. While I correctly foresaw the Republican wave of 2010, I still undercalled the 63-seat GOP pickup by 10 seats, as compared to my average of a 3-seat miss in 2006 and 2008. And I failed to foresee that what the Tea Party did for the GOP in the House would not carry over to the Senate; I overcalled the 6-seat GOP pickup by three seats, compared to my 1-seat misses in 2006 and 2008. It isn’t quite as hard for a radical/fringe candidate to get elected to the House as it is for the same type of candidate to get elected to the Senate, and that is as true for Democrats as it is for Republicans, who left three Senate seats on the table in 2010.

This year, most pundits thought the Republicans would likely capture the Senate, considering they needed only four seats, and the Democrats were defending 23 seats compared to only 10 for the GOP. But again, the Republicans have quite possibly punted away one or two seats by nominating candidates in Indiana (Richard Mourdock) and Missouri (Todd Akin) who have seriously harmed their own chances by espousing unpopular views on abortion in the context of rape-induced pregnancies. At this point, it looks like both are likely to lose, although they are fortunate to be in the states where they are. Both races will probably be close, but the advantage is with the Democrats.

It also turns out that Democrats have nominated a strong crop of Senate candidates, creating competitive races in three states the Republicans were thought certain to win: Arizona, Nebraska and North Dakota. Republicans may still win any or all of those three seats, but they all look to be heading down to the wire.

Finally, expected strong challenges to Democratic Senators in Ohio and Florida have not materialized; Ohio’s Sherrod Brown and Florida’s Bill Nelson, judging by polling trends, should both win reelection comfortably. Two other seats Republicans hoped to take, Virginia and Montana, both look like dogfights. Former Gov. Tim Kaine (D) appears to have a slight lead over former governor and Senator George Allen (R) in Virginia, although weather and electrical conditions could affect early voting, if not election day voting, and potentially alter the result there. In Montana, Sen. Jon Tester (D) is fighting Republican headwinds in his state and will struggle mightily to hold his seat against Congressman Denny Rehberg (R).

To predict the Senate, let’s take a look at the lay of the land. Taking into account the 33 seats that are up for election this year, the Republicans start with a 37-30 edge among the remaining 67 Senators who are not on the ballot until 2014 or 2016.

Among the 33 seats that are on the ballot next week, 13 are considered safe for the Democrats (including the Vermont seat held by independent Bernie Sanders, a self-proclaimed “democratic socialist” who caucuses with the Democrats). Six are considered safe for the Republicans, and in Maine, former Gov. Angus King, an independent, holds a strong lead over his Democratic and GOP challengers. That puts the spread at 43 Democrats, 42 Republicans and 1 independent, with 14 seats up for grabs.

Based on polling trends, it is clear that Democrats have opened clear leads in four of those 13 seats: Florida and Ohio and, belatedly, Massachusetts and Connecticut. Even though those last two are competitive, they seem to be headed into the Democratic column.

While Indiana and Missouri are close, they also appear to be tipping Democratic at this point, which leaves us at 49-43-1 for the Democrats, with seven true toss-ups: Virginia, Wisconsin, North Dakota, Nebraska, Montana, Arizona and Nevada.

Ultimately, it’s guesswork on these seven seats, but now that Wisconsin appears to be tipping back to Barack Obama, I’m going to figure on Wisconsin’s Senate seat going to Democrat Tammy Baldwin. I’m also going to give a slight edge to Kaine in Virginia, and a narrow edge to Rehberg in Montana, though those are both tough calls. In the end, I figure on the Democrats and GOP each taking one of those last two seats.

Democrat Heidi Heitkamp has run a great race in North Dakota, but in the end, I expect she comes up a bit short against Republican Rick Berg, so now I’m at 51-45-1.

In Arizona, Democrat Richard Carmona has consistently held a slight lead over Republican Jeff Flake, and in neighboring Nevada, Democrat Shelley Berkley has consistently trailed GOP Senator Dean Heller by a small margin. I figure on the Republicans and Democrats splitting those two seats as well, though the script could flip. A strong showing by Obama in Nevada could lift Berkley, and negative advertising against Carmona might ultimately sink him. And in Nebraska, former Democratic Senator Bob Kerrey, who was thought to have no chance just a few weeks ago, has been rapidly closing in on Republican Deb Fischer. Kerrey clearly has the momentum and I’m going to call for him to pull out the upset, for a final prediction of: 53 Democrats, 46 Republicans, 1 Independent.

In the House, I have carefully examined the map, and I’ve come to the conclusion there are really only about two dozen seats that are seriously in play. I won’t get into the rundown here, as this post is already long enough, but my call here is a small pickup for Democrats, but not nearly enough to reclaim the Speaker’s gavel. The race starts with a 242-193 advantage for the GOP. I’ve been going back and forth between a final figure of 236-199 and 235-200, but there are a couple unexpected Republican seats that seem like they might go. I’m going to split the difference and call it at 235 Republicans, 200 Democrats.

And that means at least two more years of divided government for whomever is president in 2013. Speaking of which, I’ll have a detailed prediction on the presidential race on Monday.

Two Weeks Out: Six States

by | Oct 23, 2012 | Election Analysis

With two weeks to go, it looks like the presidential race is effectively down to six states. While we will continue to hear about Florida, North Carolina and Nevada, my analysis is that President Barack Obama would need everything to go perfectly to win Florida or North Carolina, and Mitt Romney would need everything to go perfectly to win Nevada. Of the three, the state that is in the most doubt is Florida, but make no mistake, Florida will be a heavy lift for the president, and I expect he will only win it if there is a decisive break in his direction between now and November 6th.

If we look at the states that are clearly decided, it is almost an absolute certainty that Obama comes away with at least 200 electoral votes, and that Romney wins at least 180. Barring a decisive turn in Romney’s direction, there is almost no plausible scenario under which Romney wins Pennsylvania, Michigan, Nevada, or Maine’s 2nd Congressional District, which puts the president at 243. Barring an absolute collapse by Romney, it is extremely unlikely he wins fewer than 235 electoral votes.

So let’s look at what’s left, in order of likeliest wins for Obama to likeliest wins for Romney:

Ohio (18 electoral votes)
Wisconsin (10)
New Hampshire (4)
Iowa (6)
Colorado (9)
Virginia (13)

For President Obama, two states stand out in importance over all others: Ohio and Wisconsin. By all accounts, he continues to maintain small leads in both states. If he wins both of these states, and holds what he is expected to hold, he will take at least 271 electoral votes and win reelection.

It also appears clear that Ohio and Wisconsin are the two likeliest states of these six to stick with the president. In short, if Romney wins either Ohio or Wisconsin, he is likely taking Virginia, Colorado, Iowa and New Hampshire as well. Iowa and New Hampshire are probably complete toss-ups at this point, with the trends indicating Colorado and Virginia may be headed toward Romney.

The bottom line: if Romney wins either Ohio or Wisconsin, he will be the next president of the United States. Conversely, if Obama hangs on to either Virginia or Colorado, he will win reelection. Iowa and New Hampshire, right in the middle, will probably end up with the winner, but neither state is likely to be decisive.