This is somewhat true, due to several different effects that operate on the system.
The two-term limit for a President since Truman means that every eight years the vote for continuity is a vote for a party rather than a person. That’s a subtle and complicated difference. Bush I in 1988 was seen as third-term Reagan rather than as the new Republican candidate. He wasn’t, obviously, and that contributed to his loss in 1992. If you consider 1980-1992 to be one bumped 8-year cycle, the longer cycle falls into place.
Presidents normally get a “honeymoon” period. People want them to succeed. And this also tends to insulate them from the bad effects of the previous president. Obama notably did not benefit from this in 2009. We have to see if this is a historic anomaly or whether the hyperpartisanship of the era has changed the dynamic.
The three big drivers of opinion - the economy, domestic social polices, and international affairs - seldom operate on four-year cycles. With an election campaign now starting a mere two years after inauguration, there isn’t enough time to see the results of whatever a President does. With low information at re-election time, people would rather stick with the familiar. As with Truman and Bush 1, problems sometimes do not become apparent until deep into a “third” term. The cycles of history override the cycles of the presidency.
People do indeed have a tendency to stay with the same party but that’s a subsidiary function of their wanting to stay with the same person. Taking that possibility out of the choice allows the counterpressures of wanting change to solve problems more success. The artificial two-term limitation makes change every two terms a greater possibility.
Ideally, one could do a regression analysis to see which of all the effects has the greatest weight in any election year. In 2008 and 2012 the answer was obvious. In 2016 the factors are more evenly weighted, I believe. I continue to weight demographics most heavily.
Keep seeing this “old white men” in the political threads. Are you quite certain there aren’t just as many or more middle-age (white and other) men and women? Sure, probably not many young, because idealism is a youthful dream, and the promise of a liberal utopia (that seems to make sense “on paper”) fits in with that. But most of us grow up and realize that the American taxpayer can’t afford to be the world’s welfare office.
Right. Those who think that aging leads to conservatism are basing the argument off of a selected group of historical accidents. The liberal baby boomers are not getting more conservative; they’re a large audience for Bernie Sanders.
Younger voters tend to reject today’s conservatism because it supports too many things they feel are inherently wrong. That will not change. Acceptance of intolerance toward gays, women, Muslims, blacks, Hispanics, Asians, and all the other straight Christian non-white men will never return in its present form and those who support it will always be reviled. (Even young evangelical Christians are moving away from their elder’s intolerance. Prediction: atheists are the next category to be accepted.)
Older people do tend to vote in greater percentages and in a greater number of elections than younger ones. That’s why mid-term elections are no predictor of presidential elections. Much of the conservative political discourse has depended on this fact. But the oldest voters are both the most conservative and dying off the fastest. There is simply no evidence that the middle-aged people of today will become that conservative. And the youngest and most liberal demographic will be aging into middle-age and likely voting more. Conservatism as we know it today is hanging by a thread over the next few decades.
How do you have more instances than his 11 or 12 starting with an even earlier year? That’s the main thing that seems to lead you to a different conclusion. You have a lot more switches than he does.
1904 - No switch
1912 - Switch
1920 - Switch
1928 - No switch
1932 - Off-schedule, resets the cycle
1940 - No switch
1948 - No switch
1952 - Off-schedule, reset
1960 - Switch
1968 - Switch
1976 - Switch
1980 - Off-schedule, reset (this one switched early, by your projection)
1988 - No switch
1992 - Off-schedule, reset
2000 - Switch
2008 - Switch
Dropping the off-schedule switches, as they don’t fall in the cycle projected by the perceived trend:
No switch - 5
Switch - 7
In your selected sample of 12 expected switch years, a switch occurred ~58.3% of the time. You argue that when a switch does not occur on schedule, it’s due to special circumstances–how special do they have to be when they obtain nearly half the time?
I agree with you quite often, on many things. In this case, however, it seems to me that you’re applying hindsight to explain “exceptions” to a trend that can’t reliably be established in such a small sample. Is there a trend? Maybe. I’m not saying there isn’t. I’m just saying that we can’t tell enough about it to give it any real predictive value.
I think this is 1988 in reverse; in 1987 just like 2015, the incumbent party president had middling approvals but since the party out of power still had fielded front-runners with out-of-mainstream views, Dukakis on death penalty and welfare, most of the GOP field on immigration and gay marriage, on top of the fact that by the time the incumbent is on his way out, the economy will have been better than when he took office, the incumbent party will stay in power.
If a two term President has a 60+% approval rating, he can usually see his successor elected. If he’s not that popular, probably not.
But sure, the sample size is small, so there might not be a historical trend. But I don’t agree that the historical trend doesn’t matterr becuase there’s a “new normal” of greater polarization. It’s more a case of it exists or it’s just been coincidence.
My vote is for “it exists” and polls are showing Republicans doing pretty well. Plus there are other issues to consider, like the fact that faith in the executive branch is at Watergate levels. Obama has a lot of personal popularity, but unlike Ronald Reagan who was also personally popular, he hasn’t made Americans have more confidence in their country or their government. Which means that a Democrat without his personal charisma is going to have a hard time overcoming those headwinds.
I look at this as being explanative. (Yes, spell checker: that’s a real word.)
If you want to say that every presidential election is a special circumstance, you could make a good case. Yet, most elections are not that hard to predict. A few are hard because they are close; very few are hard because they totally defy expectations. (After Dewey losing to Truman, how many can you name?)
As Chronos indicated, the historic results do not match a random allotment of wins. Any theory that had Obama losing in 2008 and 2012 is a bad theory. There are good reasons when a party retains power and good reasons when it loses it. These are almost always obvious well ahead of the election.
I think Silver does important work and I agree completely with his assessment of most talking head pundits as worthless gasbags. Horse race journalism, which unfortunately dominates this forum as well, is a pain. And those who “unskew” polls are beneath contempt.
For me, explanative examinations are better, because they deal with more than the one-day results; they tell you about an era. “Why” is a critical question. Polls don’t deal well with it. Nor is there a golden formula in which to plug numbers and spit out an election return. Political science and history are the tools for “why” questions. They happen to be my fields, so I look at things through those lens. I think the results are pretty good overall. I’m not going to defend myself past that.
The relative ease of predicting elections is one of the things I generally agree with you on. However, those predictions are made knowing many of the circumstances of the upcoming election. If the pendulum were the dominant factor, you could use it to predict several cycles in advance. The set of elections we’ve examined suggests that’s not the case.
My take on the pattern–and again, I’m not saying it doesn’t exist, just that it’s less clear than people make it out to be–is that it is more symptom than cause. Two terms represents a long enough span for a party to get the message that something went wrong, and to adjust positions, strategy, and branding to try to address it. It’s also time for a great many new voters to enter the rolls, who may not share the impressions and habits of those who were already voting. Yes, a “throw the bums out” impulse is probably part of it, but I think there’s more to it than that.
In my view, the strength of the pattern relies on the parties responding to repeated electoral loss by making successful efforts to increase their appeal with both new voters and existing mainstream voters.
This seems like a good paragraph to reiterate the title of Silver’s book The Signal and the Noise.
I’m with you. It is easier for an incumbent to win than a challenger, of course. I think that explains most if not all of the eight year ‘pendulum’ pattern people think they see. The past 3 times that both candidates have been challengers, the White House flipped parties, but there’s just not much data.
I’m not saying that there’s not something to this. I am saying that, “The biggest obstacle facing the 2016 Democratic nominee: Fighting the historical trend,” or statements similar to the title of this thread are vastly overstating the strength of this trend.
I lean towards every election being a special case. Therefore I don’t think a assigning a “why” to, for example Kennedy beating Nixon in 1960, provides a ton of insight into 2016. That was a very close election and an equally strong case could be made for any number of “whys”. Chalking it up to a historical trend seems a little lazy and ignores a bunch of political science and history.