Numbers Look Grim for President Obama

Share on facebook
Share on twitter
Share on linkedin
Share on reddit
Share on delicious
Share on digg
Share on stumbleupon
Share on whatsapp
Share on email
Share on print

Superstorm Sandy has largely passed my area by, and Pepco has been spared another round of calamitous outages. Luckily for you that means I get to write a post digging deep into presidential election statistics.

Though the election polls have produced differing results, a general consensus has seemingly emerged. Mitt Romney is, at worst, tied with President Obama, and has upwards of a five-point lead. The Real Clear average of polls puts Romney up by less than a point. On the other hand, RCP has Obama up 201-191 in the electoral college, with a 290-248 edge in the “no toss-up” scenario. Obama has held a consistent edge in the battleground state of Ohio, though Rasmussen’s most recent poll now has Romney up by two.

In general, I agree with Jim Geraghty that it appears almost certain that Mitt Romney will win the popular vote. It takes polls with rather generous Democrat advantages (in the range of D+7 and up) to even get Obama tied. I trust Gallup’s likely voter screen more than other polls, and Gallup has had Romney with a steady advantage of three-to-five points.

It’s certainly possible that Mitt Romney could win the popular vote and lose the electoral college. It has happened to several presidential candidates in our history, and we are all familiar with what took place in 2000. What is fairly unlikely, however, is for Mitt Romney to win the popular vote by a substantial margin and still lose the electoral college. If Mitt Romney wins the popular vote by more than even just a percentage point, than he will be the next President of the United States. Of course we can never be certain in politics, but it seems like a safe bet that the electoral and popular vote winner will the the same person.

One of the reasons that an Obama electoral college victory in the face of a popular vote defeat is unlikely is that massive swings in national vote totals are reflected in all states. President Obama won the popular vote by seven percent over John McCain in 2008. Assume for the moment that Mitt Romney wins by just one percent – that would signify an eight point swing in favor of the Republicans. Such a huge shift in the electorate is not going to be limited to a small number of states. And as history has shown, when the incumbent party loses support, it loses support everywhere.

I have taken a look at each presidential election since 1976. Since that election, the incumbent has lost twice, the incumbent party has lost two additional times, the incumbent has won three times, and one time the incumbent party has won once. In all but two of the elections since 1980 there has been a net shift of at least eight percent. Let’s take a closer look:

In 1980, Governor Ronald Reagan defeated President Jimmy Carter by ten percent, winning all but six states and the District of Columbia. Carter had won election in 1976 by two percent over Gerald Ford. Exactly one state increased its Democrat vote share between 1976 and 1980: Vermont, which went from an 11-point Ford victory to a 6-percent win for Reagan. Vermont was in the midst of transforming from a deep red state to a deep blue state.* In fact, Vermont has voted Democrat in six presidential elections, with five of those years being the last five presidential elections.

* Since Republican states were traditionally designated with the color blue prior to 2000, technically Vermont has always been a blue state.

Four years later Ronald Reagan demolished Walter Mondale by 18 percent, a shift of 8 percent in the GOP’s favor. This time, seven states plus DC increased their Democrat vote share, though all voted for Reagan: California moved from a 17 to 16-point Reagan win, Iowa went from Reagan +13 to Reagan +7, Montana from +24 to +22, Nevada from +36 to +34, North Dakota from +38 to +31, South Dakota from +29 to +27, and Utah went from Reagan +52 to Reagan +50. Reagan also lost DC by 72 points instead of 62. So, essentially, there were no real major movements towards Mondale, though several states moved heavily further towards the GOP. This was especially true in southern states which Reagan barely won in 1980 or even lost, and which he won handily in 1984.

In 1988, George Bush defeated Michael Dukakis by eight percent. Though a fairly significant landslide, it did signify a loss of 10 percent by the GOP. Bush fared better than Reagan in exactly no states, though he lost DC by only 68 percent instead of 72. Most ominous of all were relatively narrow victories in California, Illinois and Michigan – all states that Republicans had done well in previously, but have not won since. 1988 also marked the final time that New Jersey would vote GOP in a presidential election.

Bill Clinton defeated George Bush in 1992, winning a plurality vote split between those two and Ross Perot. Clinton’s margin over President Bush was six percent, a net shift of 14 points towards the Democrats. The only state President Bush improved in was Iowa, which he lost by six instead of ten. Again, the movement of votes was total. Even strongly Republican states like Utah moved heavily towards Democrats. Obviously there wasn’t much room for Utah to move after the blowouts in 1980 and 1984, and Bush still won the state by 16, but the point is that all states moved in the Democrat direction both in 1988 and 1992.

We’ll get back to 1996 in a moment. In 2000, George Bush narrowly won election in the electoral college while losing the popular vote by some half-million votes.  Even though Bush lost the popular vote, he improved on Bob Dole’s performance by almost nine percent. As close as the final outcome was, not a single state in the union increased its Democrat vote share from 1996. On a side note, as much as we focus on Florida, it’s amazing how often we forget the nail-biters in Wisconsin (won by Gore by 0.2 percent), Oregon (won by Gore by 0.4 percent) and New Mexico (0.06 percent margin for Gore).

Again we’ll skip the re-election of the incumbent in 2004 and look at 2008. Obama defeated McCain by roughly seven percent, a nine-percent improvement over John Kerry in 2004. This time three southern states went in the opposite direction of the electorate: Arkansas went from a 10-point win for Bush to a 20-point win for McCain, Louisiana went from R+15 to R+19, and Tennessee jumped from a 15-point Bush margin to a 16-point McCain win. Otherwise, as was the case in the other five years in which the electorate moved sharply one way or the other, the movement was total. In heavily Republican states like Georgia, Idaho, Nebraska, North Dakota, South Dakota, Texas and Utah, Barack Obama improved upon John Kerry’s vote share by at least 10 percent. None of these states except Georgia was particularly close, but it shows that even the non-swing states move in the same direction as the electorate as a whole.

The two elections I skipped were ones in which the incumbent won but improved only marginally in his vote share. In 1996 Bill Clinton defeated Bob Dole, besting him in the popular vote by approximately nine percent, only a three percent improvement over his 1992 share. (His popular vote total actually increased by seven, but that was negated partially by the GOP vote increasing by about four percent). More importantly, nearly half the states actually increased their GOP vote share as three – Colorado, Georgia, and Montana – shifted from the Democrat to the Republican column (with Arizona and Florida shifting from Republican to Democrat).

In 2004, George Bush defeated by John Kerry by just over two percent, representing a nearly three-percent improvement. Again, 17 states actually moved towards the Democrats, though only two by more than two percent. Moreover Bush only lost one state he won in 2000 – New Hampshire, while picking up Iowa and New Mexico.

So the trends have been pretty clear over the past three decades or so. When there is significant movement towards one party or the other, almost no states buck the trend and move away from the rest of the electorate. Usually the shifts are consistent in all states – both swing and otherwise. The only elections that fluctuated involved incumbents winning by marginally higher vote shares than previously.

Notice that there is one pattern we have not seen: the incumbent winning by a lower margin than previously. That’s because in American history it’s only happened twice (excluding FDR’s third and fourth terms): James Madison in 1812 and Woodrow Wilson in 1916. In both cases, the president won with a lower Electoral College vote tally. Madison’s two elections came during a time when the popular vote tally was not always fully recorded. As for Wilson, he actually increased his popular vote tally from 1912 to 1916, but in 1912 he ran against two Republicans who split the electorate, and Wilson wound up with a much wider electoral college vote margin than was really warranted. In essence, NO president has won re-election with a lower popular vote margin than when initially elected. Barring something unforeseen, Barack Obama will certainly have a lower popular vote share than in 2008.

Now it’s true that Obama can afford to lose votes in Ohio, Wisconsin, Colorado, New Hampshire, Minnesota, and Iowa and still win each of those states. Obama carried all those states by decent margins (9-10 in each except Ohio, which he won by 6), and he could bleed several percentage points in each without losing them. But that would entail those states moving less strongly in the Republican direction than all other states.  Anything is possible in politics.

Possible. Not probable

More to explorer

PopeWatch: AntiChrist

Sandro Magister minces no words:   While the controversy rages on in view of the synod on the Amazon, which in reality


Good video featuring noted Civil War historian Allen Guelzo on Reconstruction.  Once Lincoln was assassinated Reconstruction was probably doomed.  He wanted to

State of the Race

    The race for the Democrat nomination for President is gelling faster than I expected.  The weaker Democrat candidates are pulling


  1. It seems like one of the big problems with just about all models is that presidential elections are fairly infrequent and so for any given comparison there are very few like situations.

    Right now, what I’m moderately confident in is that short of some big surprise in the near future Romney should manage a popular vote win, though possibly a narrow one. The big question is how this plays out in the electoral college. The general rule is that the two don’t split. But on the other hand, there also aren’t a whole lot of really close modern elections to go from.

    I fear that the basic structure of the map and votes are better for Obama than for Romney, since Romney needs to win Ohio plus one other state that isn’t already a moderately clear win for him (I’m counting FL, VA and NC as fairly clear wins, though it’s possible I’m being over-optimistic about VA.) There are several solid possibilities, but it means that Obama just has to play defense and hope that Romney doesn’t break through, while Romney needs enough of a wave of support that several states fall his way. Romney does indeed seem to be riding a wave, but at this rate it seems like we won’t know till election night if it’s big enough.

  2. I disagree with one aspect of your comment, Darwin. There seems to be the perception that Romney must win Ohio while Obama can afford to lose it and still win. I think that the reality is quite the opposite. The president is playing defense, and Ohio is his last line of defense. If that goes, he is done. While it would be difficult for Romney to lose Ohio and win the 270+ he needs, it is actually plausible that he can pick off enough states like Wisconsin, Iowa, and New Hampshire in lieu of Ohio (assuming other states like Florida, Virginia and Colorado come home).

  3. That’s because in American history it’s only happened twice (excluding FDR’s third and fourth terms): James Madison in 1812

    Umm. I do not think there was any popular balloting for electors prior to 1824.

    Re a disjunction between the popular vote and the electoral vote. It has happened 4x and in a fifth case (1960) the Democratic vote in two states was cast for an “uncommitted” slate rather than one bearing the Democratic candidate’s name, so there is some opacity about how to tabulate the popular vote.

    1. In one case you had a multiparty contest and six states held no popular balloting, having the state legislatures select the electors.

    2. In another case you had jagged state-to-state variations in the relative dimensions of the electorate and widespread fraud and (down South) intimidation (topped off with a finicky legal dispute over the decisive electoral vote, an elector having been disqualified because a federal employee).

    3. In a third case, you had some of the above and a popular vote margin under 1%.

    4. In the other two cases, the popular vote plurality was under 0.6%.

    If Obama wins the electoral college while losing by two or three million popular votes in a clean contest, it will be something without precedent (but, then again, a great many weird things have happened in recent years).

  4. is it assumed that Mitt has CO in the bag? If he wins OH but only takes back the South + IN (that one isn’t in doubt obviously,) he’d still narrowly lose.

  5. Colorado had looked pretty good for Romney for a while, although now I think it is back in the toss-up column.

    Umm. I do not think there was any popular balloting for electors prior to 1824.

    The system was complicated, and I’ll have to look back at my books, but I don’t believe that is totally correct. IIRC, most states allotted their electoral votes based on popular votes by this time, though not exactly through the winner-take-all allotment practiced in 48 states.

  6. * Since Republican states were traditionally designated with the color blue prior to 2000, technically Vermont has always been a blue state.

    And I thought I was the only one to remember that!

  7. G-Veg,
    It was never official, but blue for GOP had been the more predominant practice on TV network maps until Bush v Gore. I don’t recall what the networks did that evening, but the next day USA Today published a national map color-coded by county, which captivated American attention since it showed how how pockets of Dem support had overwhelmed a sea of GOP, and that map happened to use red for GOP and blue for Dem. After that, the colors became part of our national consciousness. Unfortunate if you ask me. Red is more appropriate for the Dems. That at least is my recollection.

  8. Darwin,

    My comment was simply that the Colorado model stated in August what many people at the time said was fanciful – Minnesota, Pennsylvania and Wisconsin would all go for Romney. Minnesota and PA were viewed as solid Obama. Yet here we are a week before and both campaigns have moved to those states. I don’t know any pollster or analyst who made a similar call.

  9. JDP,

    If Romney takes Ohio (assuming the South plus IN), then all he needs is any one of NV, CO, IA, WI, PA, or NH. While Obama could lose Ohio and still run the table in those states, it is not very likely.

    On the other hand, if Obama takes Ohio, Romney could still win if he takes (i) PA and any other state, (ii) WI and CO and any other state, (iii) WI and NV and any other state (except NH unless it brings along a single vote from ME), or (iv) various other combinations. While not easy, it is a better route than Obama’s.

    It is easy to see why both sides have to regard OH as key, but at the margin it is even more essential for Obama than Romney.

  10. I guess the reason I’d see Obama as having the easier time at the moment is that if he just wins all the states that he’s currently ahead in the polling in, he wins.

    I’d say that Romney has a fairly good path to 248 and Obama only has a fairly good path to 237, but of the remaining states (NV, CO, IA, WI, OH, NH) they’re all ties or moderately good “leans Obama” cases.

    My big hope is that the national polling (which is much more frequent and statistically rigorous) actually gives a far better impression of where the states are trending, and that VA is thus a lock and OH is tipping into Romney’s column right now (as the last couple state polls do actually suggest). If that’s the case, with one week’s runway left we could be popping corks early on election night. But right now I’m still worried.

  11. Back in the day, when broadcast networks were A) the only game in town and B) still more-or-less journalistically reliable, the colors actually were designed to switch every election. The incumbent was red and the challenger was blue one cycle, then vice-versa. This was originated in 1976 when NBC used a back-lit big board. But the scheme wasn;t close to nailed down yet – Republicans were usually blue because of the incumbancy-challenger cycle between 1980 and 1996. In 1980, the country looked, in the words of David Brinkley “like a suburban swimming pool.” CBS was reversed from the other two major nets, so even at that there was no real conformity.

    Eventually it settled into what it is now, for no real discernable reason, probably with the advent of CNN as a major player (GOP = red always) and because in 2000 the map was up for more than just election night as folks waited for the SCOTUS call. The two largest speculations are that Red and Republican both start with “R,” and that, in the eyes of liberal media types, blue is a peaceful and sophisticated color while red is angry and violent. Neither theory has been proven and both remain popular in various circles, depending.

  12. Darwin,
    I certainly agree that cork popping is very premature. But the most recent reports from the best poll unpackers in the business (Barone and Cost) are very favorable to Romney. Assuming nothing, I’m going to be content with my optimism till proven otherwise next week.
    More specifically, I think Romney will take the southeast, and he is ahead in OH and CO once the polls and their imbedded assumptions are understood. If he takes Ohio, we win. But if he falls short in OH, I think he has still has a decent shot given the other combinations.
    But you are right, the race is tight, and optimism is not money in the bank.

  13. Everyone, you might remember my post (linked below) earlier this month that stated why I believe the demographics point to a Romney victory in Ohio. I kept hearing from people in the know that everything was looking good and then the other day on CBS, the Ohio GOP chairman talked about the Ohio Groundgame being superior to the President’s, and the Democrats know it. As a matter of fact even Mark Halperin, no friend of the GOP, made a statement that his Democratic sources had never seen the Ohio Conservatives so organized.

    The Gallup early voting sample points to this as well. I believe Rush Limbaugh said something to the effect that this is what scares the Left the most. Below is also a link to a story of mine featured in the National Review, in which I talk about seven Ohio swing counties to watch on Election Night.

  14. WK,
    Thank you for your clarification of pre-1980 practices. Seems spot on right to me. I still think the USA Today map is the explanation post-1980.

  15. In terms of an electoral college/popular vote split, it’s worth noting that as of now several million people lack power due to the effects of Hurricane Sandy, and it’s not clear how far recovery efforts will progress before next Tuesday. Since the damage seems to be concentrated in blue states like New York and New Jersey, it’s possible that depressed voter turnout in these states due to the storms could help Romney secure a popular vote majority even if he fails in the electoral college (at the very least, I expect this to be cited as an explanation for the Romney popular vote win if the split does occur).

  16. Blackladder,
    That may be right, but unless there is some difficulty with the NOVA vote, do you agree that there would be no electoral college effect? If delayed metro-Philly returns are key to determining PA, it seems to me Romney has won.

  17. Disenfranchisement!M!

    Damn Neocons! Hurricane Sandy is a conspiracy I tell ya! We should delay the vote by a month so everyone has a chance to vote.

  18. BA says “It’s possible that depressed voter turnout in these states due to the storms could help Romney secure a popular vote majority even if he fails in the electoral college ”

    The only practical argument would be if Romney Won NY or NJ because of low turnout hurting the Dems And one or both of the states were critical to an Electoral College victory. That is not possible.

    Mike P There is no problem from Sandy in North or South VA affecting turnout. I don’t think it will have any significant impact in PA either. We are still a week away from the vote so I would only see parts of NYC and the Jersey coast being definitely impacted next week. I don’t think Obama will carry VA anyway.

  19. I think BA’s point is simply that lingering storm effects could impact vote turnout in New Jersey and New York so that while neither state will turn into Romney states, the overall impact would be to push the popular vote towards Romney even as the electoral college goes to Obama. I don’t think that the storm’s aftermath is likely to impact the vote that significantly – assuming 40% of that vote is going for Romney, there would have to be well over a million people who can’t get to vote in order to ignite that much of a shift.

  20. The loss of thousands of military absentee ballots doesn’t seem to bother the MSM a whit. However, the disenfranchisement of active duty military is deeply troubling, both because it seems immoral that we can’t seem to guarantee that privilege to those guaranteeing the privilege to us and because this could significantly affect states in play like PA.

  21. I believe the polls unless I have a reason not to.

    I don’t trust Gallup – Their screen is too tight. They had Obama +11 in 2008, which was farther off than anyone. Currently, their Romney +5 is an outlier and I believe for the same reason.

    The reason for the “skewed polls” is Republican leaning Independents. Polls that treat them as Independents show Romney crushing Obama among independents, but a D+8 electorate. Polls that treat them as Republicans show an even race among independents with even turnout from each party. Either way, this leads to a narrow Romney win in the popular vote. The “unskewed polls” are probably double-counting Republicans.

    I believe Rasmussen tracking is bouncing between R+3 and R+1. I would put it at R+2. His national polls are dead on, but his state polls are less accurate.

    R+2 is slightly less than Bush’s margin in 2004. Although Bush easily won the popular vote the second time around, the election was in doubt until the next morning because of the close race in Ohio.

    Right now, the polls show exactly that happening. Romney wins the popular vote by slightly less than Bush, but barely loses Ohio and the election.

    The difference is that Obama is focusing all his efforts on the swing states (like Kerry in 2004) and ignoring the safe states. Expect Romney to close the gap in the blue states and run up the score in the red states, due to GOP enthusiasm, but he’s still not polling at over 270EV.

    An Obama win without a majority vote would arguably be a “worst case scenario”.

  22. G Veg “However, the disenfranchisement of active duty military is deeply troubling, both because it seems immoral that we can’t seem to guarantee that privilege to those guaranteeing the privilege to us and because this could significantly affect states in play like PA.”

    Since the Repubs can’t be bothered to fight for it very much, I guess it’s not important to either party however immoral it is. This has been going on for many months and even goes back in some sense to Florida in 2000. Hardly a peep from Boehner McConnell and Romney.

  23. Thanks, Rozin. I figured Sandy would have no effect, but good to know.

    Paul, Blackladder, and Jim,
    I don’t see why the popular vote matters. It is not the system we have or the way the campaign is conducted. If Romney falls short, he falls short. I don’t think he will, but if he does his winning the popular vote would be irrelevant except to the extent it deprives Obama of any claim to a mandate.

  24. I take it as a given that Romney starts with a base of 257 electoral votes. This includes Colorado where the Republicans have the advantage in early voting.

    With that as a given Romney has the following paths to 270:

    1. Ohio-18 electoral votes.
    2. Pennsylvania-20 electoral votes.
    3. Michigan -16 electoral votes
    4. Wisconsin-10 electoral votes with New Hampshire -4 electoral votes
    5. Minnesota-10 electoral votes with New Hampshire-4 electoral votes
    6. Iowa-6 electoral votes-New Hampshire-4 electoral votes-Nevada-6-electoral votes

    New Hampshire I think is close to being a given for Romney. If Romney wins Ohio, Pennsylvania or Michigan he wins with no further states needed. With New Hampshire, Wisconsin or Minnesota can be Kingmaker states. If Romney loses all of the above states except New Hampshire, he still has a path to victory with Iowa and Nevada.

  25. Another reason for the reversal of red and blue in political symbolism could be that red lost its association with communism and leftism in general after the collapse of the Soviet Union and the end of the Cold War, while the wealthy, well-educated elites or “blue bloods” that used to be the backbone of the GOP became much more liberal in their leanings.

  26. Don is exactly right. Both candidates really need Ohio, but at least Romney has options if he loses Ohio. I don’t think Obama does.

  27. Mike, I agree that the popular vote doesn’t matter in the end. My point is simply that I don’t expect Romney to win the popular vote by more than a scant margin AND lose the electoral college. So the national polls do provide meaningful insight.

  28. NBC also had the blue = Reagan and red = Carter color scheme in 1980 as well (with Carter’s Georgia the only “red” state in the bunch shown here):

    Notice John Chancellor calling Reagan the winner at 8:15 p.m. Eastern Time (7:15 Central Time), which was somewhat controversial at the time since polls in some Western states hadn’t closed yet.

    Although I think Romney will ultimately win and by a larger margin than the media have led us to believe, I doubt the suspense will end quite as fast as it did that year.

  29. Mike P: [Romney] winning the popular vote would be irrelevant except to the extent it deprives Obama of any claim to a mandate.

    Wishful thinking I’m afraid. The media will talk endlessly of all the vote suppression by those evil Republicans that silenced the voices of so many.

  30. Paul, understood and agreed.

    Rozin, yeah they might try that but no one will take it seriously outside their echo chamber.

  31. “With New Hampshire, Wisconsin or Minnesota can be Kingmaker states.”

    Wisconsin maybe, but Minnesota? Really? As Paul said a couple of weeks ago, it’s like the flip side or mirror image of Arizona — a state that “should” be red but remains stubbornly blue/liberal. I know there are some polls showing it could be in play but I’m not getting my hopes up. And, wasn’t the unexpected Romney TV ad buy in Minnesota really intended to target western Wisconsin voters?

  32. The most recent poll we have from Minnesota Elaine shows it 47-44.

    Two things about this poll. First it is a Star-Tribune poll and they have a notorious history of greatly exaggerating Democrat numbers in their polls.

    Second, for an incumbent Democrat President to be at 47 a week out from the election in Minnesota is a definite warning sign that Obama is in trouble in this state. Minnesota hasn’t gone Republican at the Presidential level since Nixon in 1972. The Minnesota ad buy could do double duty. I doubt that the Romney campaign thought that Minnesota was really in play until they saw this poll.

  33. D McC I doubt that the Romney campaign thought that Minnesota was really in play until they saw this poll.

    Then they didn’t take the University of Colorado model seriously either. I have trouble believing that they would do this based simply on some media poll, no matter the results, when they have been notoriously inaccurate and volatile this cycle. I’m inclined to view it as following the same logic that Obama had in advertising in VA late in 2008.

    Mike P: I hope so but an Obama win would mean that their echo chamber is pretty large. Although at bottom Obama doesn’t care whether he has a mandate or not to do what he wants.

  34. A bit to my point, here’s a Rasmussen poll in Massachusetts that has Obama up by 19. He defeated McCain there by 26. There might be a bit of a home-state bias for Romney, but considering where his favorables were when he left office, maybe not so much. So there’s obviously no way Obama is losing, but if those numbers are accurate it shows how the overall tide is shifting. Those numbers are basically in line with what you’d expect from a 7-8 point swing in the electorate.

  35. Okay, trivia time. We’ve already covered Minnesota being the one state that has voted Democrat every election since 1976. Can anyone name the nine states that voted Republican in each of those elections? No cheating.

  36. You got 6/9 Don. South Carolina went for Carter in 1976, Montana for Clinton in 1992 and Arizona for Clinton in 1996.

Comments are closed.