It’s possible indeed

by Henry Farrell on July 12, 2013 · 21 comments

in Campaigns and elections

This, from Megan McArdle:

My assertion that there’s a 70% chance that the GOP controls White House, Senate, and House in 2017 has attracted a lot of pushback. And it’s certainly possible that I’m wrong! Here’s my thinking, for what it’s worth: Since the Civil War, only two Democratic presidents have been succeeded by another Democrat. Both of them–FDR and JFK–accomplished this by dying in office. Since World War II, only four presidents have been succeeded by a member of their party. As I mentioned above, two of them accomplished this by dying in office. One of them accomplished this by resigning in disgrace ahead of his own impeachment. Only one of them, Ronald Reagan, left office at the end of his appointed term and was succeeded by a duly elected member of his own party.

reminds me of this classic cartoon by XKCD, which should be blown up into A0 format, and placed in a permanently visible position in front of the desk of every pundit tempted to make pseudo-quantitative oracular announcements about American politics (extremists might want to go the full Clockwork Orange with the eyeclamps but that strikes me as overkill).

Human beings are cognitively predisposed to perceive patterns in the world. Many, likely most of these patterns are garbage. Without good theories, and good ways of testing those theories, we’ll never be able to tell the garbage patterns from the real ones.


Chaz July 12, 2013 at 4:18 pm

This post shows us that the Democrats nominating Hillary Clinton or Joe Biden in 2016 would be a big mistake. If Rick Perry gets the nomination then the Dems’ only hope is to balance his k out with Cory Booker.

But I’m concerned that Munroe didn’t get the pattern right. It might be only first names ending in k. Then Rick Perry is unstoppable.

Chaz July 12, 2013 at 4:21 pm

Ugh, I misread the comic. We need Kory Booker.

Matt Jarvis July 12, 2013 at 6:13 pm

I’m confused….Eisenhower didn’t win in 1952 without winning the House or Senate. He had unified government for his first Congress.

Andrew Rudalevige July 16, 2013 at 10:01 am

Yes, the cartoon messed that one up. Ike had (small) Republican majorities in both chambers in 1953-54.

Kevin Drum July 12, 2013 at 8:12 pm

Henry, I think you’re being unfair here. Since WWII, the American public has shown a very distinct distaste for keeping a party in the White House for more than eight years. It’s only happened once, and it’s not some weird, invented stat. It’s one of the key features of postwar politics.

Obviously this doesn’t mean it will never happen again, but it’s hardly unreasonable to use it as a marker in favor of a Republican victory in 2016. In fact, I think several of the popular prediction models use it.

Nadia Hassan July 12, 2013 at 9:02 pm

There is a tendancy. It persists after controlling for approval and the economy, but it is diminished after controlling for the out party`s campaign strategy. The sample sizes are small, though. Still, even if the GOP has a strong chance to win yhe Presidenccy, the Senate is a less likely proposition. The GOP has a shot at the Senate in 2014, but the landacape in 2016 favors Dem gains. That is not a sire thing, but 2014 and 2016 seem to point toward even odds.

y81 July 12, 2013 at 9:35 pm

If it’s 50/50 who controls the Senate after 2016, and better than 50% that the Republicans control the White House and the House, then the Democrats would be awfully foolish to abolish the filibuster, no? Which was McMegan’s point (or McGarble’s, as Prof. Farrell would have it).

S. Tarzan July 12, 2013 at 10:01 pm


If you take out the examples that can’t be explained by a bad economy or an unpopular military, how elections does this ‘rule’ explain?

Nadia Hassan July 12, 2013 at 11:21 pm

2000, likely. One could arguably claim that Watergate was the cause of 1976 and Vietnam the key to 1968. And those events were likely consequential. But statistical analyses reveal that even after controlling for the economy, wars, and approval, parties that have held the White House for two or more terms are at a disadvantage. So, maybe the incumbent party could have held on in 60, 68, and 1976.

S. Tarzan July 12, 2013 at 11:56 pm


How large is the effect you’re talking about? 1960 and 1968 were both really close, after all, and 1976 wasn’t that much larger. Are we talking about something that could be swamped by campaign effects?

Nadia Hassan July 13, 2013 at 2:41 pm

It depends on what else is in the regression equation. Modelling by Larry Bartels, Alan Abramowitz, Carl Klarner, and John Sides & Lynn Vavreck indicates that holding constant the economy and approval, an incumbent party that has held the White House for 8 years will perform 4 points worse (in vote share) than one that has held it for 4 years. It is a large enough effect to decide races. The effects of campaign activity like ads and ground game are smaller.

In her study on the economy, the fundamentals, and campaign strategy, Lynn Vavreck found that the effects of incumbency and incumbent party tenure are smaller when including a variable for the out-party’s campaign, and the out-party’s campaign has a bigger effect than incumbency or terms.

David T July 13, 2013 at 11:36 pm

“Since WWII, the American public has shown a very distinct distaste for keeping a party in the White House for more than eight years. It’s only happened once, and it’s not some weird, invented stat. It’s one of the key features of postwar politics.”

At least once (2000) and arguably twice (1960, when JFK’s supposed popular plurality actually included votes for a mixed slate of pro- and anti-JFK Democrats in Alabama) “the American public” actually showed a preference for keeping the same party in the White House (in addition to 1988). It’s just that in 1960 and 2000 the non-incumbent party won the *electoral* vote. And in 2016 the Republicans are extremely unlikely to win the electoral vote if they don’t win the popular vote; even though Obama only won by 3.9 percent, to defeat him, Romney would have had to win at least one state Obama carried by 5.36% or more.

And of course choosing World War II as the starting point is very convenient but arbitrary. It allows one to ignore the Democratic years of 1932-52, the Republican years of 1920-32, and the Republican years of 1896-1912.

Henry Farrell July 12, 2013 at 9:28 pm

Kevin – there is some evidence that it’s difficult to get a third term. But that’s not what being argued in the material quoted (although I do think it gets referred to later in the post). For a ‘Democrats never are succeeded by other Democrats unless they die in office’ effect – what’s the causal story? For the ‘only four presidents since World War II have been succeeded by a member of their own party effect’ … oh dear. The point is that the tossing in of all sorts of purported historical trends and patterns in the antecedent reasoning leads to a completely incoherent mish mash of an argument. It’s doubtless a perfectly entertaining sport, invoking precedents as to how this or that hasn’t happened in the past. But without some causal explanation, and some commitment to somehow testing that causal explanation, the purported patterns in the quote are probably garbage. Or to put it another way: you can come up with similarly ‘convincing’ patterns to justify pretty well any prediction you might want to make. And pundits do.

Nadia Hassan July 12, 2013 at 11:54 pm

Andy Gelman has mentioned that 1968, 1960, 1976, and 2000 should be considered ties. Using never or rarely when some situations are 50/50.

TallDave July 12, 2013 at 9:29 pm

All of those XKCD panels are for criteria much more specific than “control of the government tends to oscillate between the major parties.” There are a lot of fundamental reasons why that trend exists (many of which those of us not married to a party tend to see as healthy for the nation).

dapoint July 13, 2013 at 4:03 am

I think all the commenters are missing the point of this post (actually acting like the pundits in question).

Bryan July 13, 2013 at 5:08 am

No one’s mentioning the landslide 1988 election. 3rd term for the GOP and would likely have been a 4th had the economy not gone into recession in 1991-1992.

jheartney July 13, 2013 at 3:51 pm

So long as we’re predicting things based on precedent, I think the controlling one here is “Megan McArdle is always wrong.”

matt w July 15, 2013 at 1:15 pm


Since I’m unwilling to give her the pageview, does she even bother to defend the assertion that the chance that the Republicans will control the Senate is over 70%? After an election where the Democrats are defending ten seats (every one in a state Obama won decisively) and the Republicans are defending 24 (seven in Obama states, at least three more in swingish states)?

Brett Champion July 19, 2013 at 10:27 am

But unlike some of the things in the XKCD cartoon, there might actually be a real world reason why the trend that McArdle identifies has a reason for existing other than simple chance. Presidential approval ratings have consistently been lower at the end of a president’s term than at the beginning. And generally not just a little bit lower either.

While I think one of the main drivers of lowered approval ratings is the public just being tired of seeing the president’s face all over the place for 8 years, that still has an effect on his party’s successor, who is almost of necessity going to be heavily associated with the sitting president, whether he or she wants to be or not. Fatigue with Obama seems to be setting in quite solidly now. If the Republicans put up a remotely fresh and reasonable candidate (e.g., Chris Christie), then the Democrats might be hard-pressed to shake the Obama-fatigue factor, especially if either Hillary Clinton or Joe Biden is the candidate.

David T July 19, 2013 at 10:16 pm

Obama’s ratings are similar to Bush’s in 2005–approval about equal to disapproval, or at most slightly underwater. Yet the GOP might well have won in 2008, “Bush fatigue” or no, if 2008 had not witnessed the worst economic collapse since the Great Depression. And “Clinton fatigue” or no, Gore actually won the popular vote in 2000…

Comments on this entry are closed.

Previous post:

Next post: