Historical Precedent For The 2016 Democratic Nominee

[Mister Mets]

One trend I've noticed in previous presidential elections (which obviously is limited to a rather small sample size) is that whenever a party seeks a third term in the White House they do worse than in the previous election.

This applies even if they win.

For example, Reagan won 58.8% of the popular vote in 1984, while Papa Bush won 53.4% in 1988.

If the dropoff in percentage of the popular vote for the next Democratic nominee is the same, it would suggest a popular vote of 47.27% for the Democrats. And I'm look at how earlier ratios of the party's strength in a reelection: the party's strength seeking a third term in the White House would fit Obama's 51.1.%.

In 2008, George W Bush got 50.7% of the popular vote, and four years later McCain got 45.7%. A Democrat who keeps this ratio gets 46.06% in 2016.

Sometimes, the difference is starker. LBJ won with 61.6% in 1964, while Hubert Humphrey won 42.7% four years later (you could also say that he lost the popular vote to Nixon by 0.7%.) If we give Humphrey a 49.6% result to make up for George Wallace's interference, that ratio gives the next Democrat 41.4% of the popular vote.

Nixon won reelection with 60.7%, and Ford went on to lose with 48%. A Democrat who keeps that ratio would get 40%.

Nate Silver said Hillary Clinton may be such a strong candidate that it would be like running an incumbent in 2016. When FDR sought a third term his 54.7% was impressive, but less than his 1936 total. A similar ratio to Obama's 2012 reelection would give a Democrat 46%.

Comparing FDR's 1944 results to Truman's 1948 gets slightly different numbers. In 1944, FDR got 53.4% of the vote. In 1948, Truman got 49.6%. A similar dropoff from Obama's 2012 total gives the next Democrat 47.46%.

The smallest dropoff was from 1996 to 2000. Clinton won reelection with 49.2% of the popular vote, while Gore got 48.4% in 2000. Of course, there was also Ross Perot getting just over eight percent of the popular vote, but a Democrat who stays at the Gore: Clinton ratio would get 50.27% of the popular vote, which will likely translate into enough states to win the White House.

It's just a trend I've noticed. It's possible that the numbers don't mean anything, and that I've noticed a pattern that has largely been driven by outside events, the result of coincidence.

It's also possible that we're in a different political era, and that observations about previous elections will be as meaningless in the next cycle as observations about 1928 were in predicting 1932. But it is an interesting precedent.

[Robert California]

No Washington-Adams, Jefferson-Madison, Jackson-Van Buren, Grant-Hayes, or Roosevelt-Taft?

[Orser67]

Interesting observation. Speaking of Nate Silver, he argues that incumbent presidents seem to be particularly strong when they run. So perhaps that contributes to the observation you noticed.

http://fivethirtyeight.blogs.nytimes.com/2013/07/18/the-white-house-is-not-a-metronome/?smid=tw-fivethirtyeight&seid=auto

"This contrasts with cases in which the incumbent party has won just one consecutive term in office. In those cases, since 1856, the incumbent party has won the popular vote in 14 of 18 elections and the Electoral College in 13 of 18, and its average margin of victory in the popular vote has been 8.5 percentage points."

[bedstuy]

Every presidential election is sui generis.

[Blue3]

Historical precedents are broken with every election.

Wasn't one broken in 2012 that re-elected Presidents usually have a larger percent of the popular vote in their second election than their first election?

[Mister Mets]

Every presidential election is sui generis.

That's not entirely true.

48 states voted the same way in 2012 as they did in 2008.

Historical precedents are broken with every election.

Wasn't one broken in 2012 that re-elected Presidents usually have a larger percent of the popular vote in their second election than their first election?

Yup. Though this isn't necessarily a good thing for Democrats, since it's an example of Republicans overperforming.

It was also partly due to the unusual circumstances of the 2008 election, which came two months into the biggest economic crisis since the Great Depression. That swung a few states to Obama.

[old timey villain]

Every presidential election is sui generis.

That's not entirely true.

48 states voted the same way in 2012 as they did in 2008.

Historical precedents are broken with every election.

Wasn't one broken in 2012 that re-elected Presidents usually have a larger percent of the popular vote in their second election than their first election?

Yup. Though this isn't necessarily a good thing for Democrats, since it's an example of Republicans overperforming.

It was also partly due to the unusual circumstances of the 2008 election, which came two months into the biggest economic crisis since the Great Depression. That swung a few states to Obama.

No, it was actually a case of Republicans underperforming, since the unemployment rate and generally low level of US satisfaction should have made it easy for them to defeat Obama.

If Republicans underperform in 2016, with an improved economy and a fresh new Democratic nominee they could lose again.

[bedstuy]

Every presidential election is sui generis.

That's not entirely true.

48 states voted the same way in 2012 as they did in 2008.

No, every election is sui generis.  The electorate and the country gradually changes so the results are unlikely to vary drastically from year to year. 

[President Tyrion]

Sui generis, yes.

If you wanted results that would actually be worthwhile, you would need to hold multiple elections at the same time, or with the same difference in time frame, because the gradual changes in the electorate are not homogenous over eras.

It's a good thought, but it doesn't withstand rigorous examination.

[Pessimistic Antineutrino]

Every presidential election is sui generis.

That's not entirely true.

48 states voted the same way in 2012 as they did in 2008.

No, every election is sui generis.  The electorate and the country gradually changes so the results are unlikely to vary drastically from year to year. 

That fact is actually an example of 2012 being sui generis.

2012 was the second election ever where only two states flipped since 1888, where the only states to flip from 1884 were New York and (coincidentally,) Indiana.

[Skill and Chance]

The problem with all of these comparisons is that if you go back further than FDR-Truman, the results change completely.  Hoover did better than Coolidge and Teddy Roosevelt did better than either of McKinley's wins.  And the biggest Republican win of the late 19th century streak wasn't until 4 elections in.  At the peak of their post-Civil War dominance, Republicans won 16 of the next 18 elections in the electoral college and were 14 for 18 in the popular vote.  

So either the 22nd Amendment was a game-changer (unlikely given that only FDR ever served 3+ terms and besides him only TR credibly ran for a 3rd) or it all comes down to chance.  Maybe the next Republican president will be elected in a landslide in 2016, or maybe he/she won't be elected until 2028 or 2032.  We don't have enough data to predict.  The best we can go on is economics and even that isn't perfect (Obama probably should have lost in 2012 and Gore should have won big in 2000).

[pbrower2a]

Circumstances are different -- boom vs. bust; inflation vs. deflation and steady price levels; military triumph and military debacle; war, peace, and the threat of war; races involving incumbents vs. open-seat elections; strong third-party nominees and their absence; differences of quality of opponents; political crazes and their absence; and legal changes in voting practice. Some elections have relatively little polarization in results between the states (the 2008 elections may be the most extreme manifestations ever*) and some have few.

The 1952 and 1956 elections and the 2008 and 2012 elections  are two of the most similar pairs involving the shifts of relatively few states aside from the 45+-state blowouts of FDR in 1936, Nixon in 1972 and Reagan in 1984.  

One of the most marked anomalies of 2012 was that President Obama won re-election with a near-mean percentage of electoral votes (61.7%). What is so strange about that? Beginning in 1900 -- when horses and buggies still vastly outnumbered cars and Marconi had yet to send information with radio waves -- no Presidential election had been won with between 57.1% (Truman 1948) and  65.3% (McKinley 1900) of the electoral vote. Presidential elections were either close or they are at least near-blowouts. America has changed in many ways with vast immigration and assimilation, urbanization with depopulation of rural areas, the rise of suburbia, technological change in communications and travel, the addition of five states, major wars, the enfranchisement of women and in practice minorities, and the huge shift of population.  (OK, had the Romney campaign hadn't come up with the silly campaign that tried to link President Obama to Hugo Chavez and Fidel Castro, Romney might have won Florida and stuck President Obama with 'only' 56.3% of the electoral vote and kept the pattern intact).  


*Jimmy Carter lost 18 states by less than 8% and won 13 by less than 8% in 1976 in a close election; in 2008 Barack Obama lost three states by less than 8% and won five by less than 8%.  

[Bull Moose Base]

The elections most similar to 2008, involving the fierce rejection of the incumbent party during a terrible economy, 1932 and 1980, were followed respectively by 5 and 3 White House terms for the new party. I think that's the most relevant precedent as far as third terms for a party but I think the most is probably 2012 where hardly any states switch from the preceding cycle.

[Stranger in a strange land]

Historical precedents are broken with every election.

Wasn't one broken in 2012 that re-elected Presidents usually have a larger percent of the popular vote in their second election than their first election?

[PolitiJunkie]

Historical precedents are broken with every election.

Wasn't one broken in 2012 that re-elected Presidents usually have a larger percent of the popular vote in their second election than their first election?

LOVE IT!

[Robert California]

Historical precedents are broken with every election.

Wasn't one broken in 2012 that re-elected Presidents usually have a larger percent of the popular vote in their second election than their first election?

James Madison vs. DeWitt Clinton, bro. Madison got freakin' 64% in 1808 and then like 53 or something in 1812.

[pbrower2a]

Historical precedents are broken with every election.

Wasn't one broken in 2012 that re-elected Presidents usually have a larger percent of the popular vote in their second election than their first election?

President Obama had some wiggle room for winning re-election -- and used it. He did not panic. Nickel defense -- make gains easy for the opponent, but nothing that can result in a victory for the team behind. He had to do so against a well-funded opponent with a ruthless campaign behind him.

.... Barack Obama has shattered so many precedents that 2016 is sure to be weird.     

[Mister Mets]

Historical precedents are broken with every election.

Wasn't one broken in 2012 that re-elected Presidents usually have a larger percent of the popular vote in their second election than their first election?

That's a great strip.

However, it is worth noting that while something may not be a certainty, it can be a strong probability.

[milhouse24]

I think those are very interesting numbers, and they prove that Hillary will have a very difficult time winning, especially if the GOP nominate a candidate younger than McCain and much friendlier than Romney. 

But Hillary is such a strange candidate.  On one hand she has been tangentially voted for president without actually ever being elected to the Presidency.  She's always maintained a high level of media exposure, but at what point does the public and media get tired of her and start a backlash. 

There certainly was a backlash against her "queen crowning" in the 2008 primaries, when Grassroots Democrats didn't want an Iraq-voting senator as the next President.  So perhaps, Hillary has no more "backlashes" to deal with, especially if she can manage to stay out of controversy (Benghazi). 

The question becomes what do voters want in the next president, and can Hillary deliver that?  Or will she simply run on the narrative as "first Female president who promises no changes."  Obama ran as a "Bi-racial candidate who promised lots of changes."  But there isn't much that Hillary can promise to "change since we have 2 democratic terms." 

Will the voters show up for Hillary in 2016 or will people tire of her, and think she is too elderly? 

People are always looking to the future, and Hillary might be talking about the past, or too old to tackle the exciting issues of the future.  After all, Hillary has to be expected to live longer than 2 terms (8 years) if voters really want to take her potential presidency seriously. 

I think Hillary might win, but it will be very close regardless of who the GOP nominate. 

[Mister Mets]

A related precedent I noticed is that the map does remain relatively stable, so as a party gains in the popular vote, they tend to keep the states they had in the previous elections and add new states.

I've noticed one exception for this in the 1996-2012 period.

In 2000, Bush adds about 11 states to Dole’s 19 states: Florida (with a caveat), New Hampshire, Missouri, Ohio, Nevada, Tennessee, Arkansas, Arizona, West Virginia, Louisiana and Kentucky. It's kind of obvious that if a party gains seven percent in the popular vote, they're usually going to keep the states they had earlier in the election and add a good amount to that, something we'll notice again in 2012.

In 2004, Bush does a little bit better in the popular vote, and takes most of his previous states, plus Iowa and New Mexico.
He loses New Hampshire (which is the one time this happens.) However, due to Nader’s 2000 showing, Bush’s share of the vote in the state is actually a bit higher in 2004 than in 2000 (48.87% VS 48.07%.)

In 2008, Obama gains Nevada, New Hampshire, Iowa, Colorado, Virginia, Ohio, Florida, Indiana, and North Carolina.

In 2012, Romney does better in the popular vote than McCain, and takes back Indiana and North Carolina.

There are some other exceptions in earlier elections. Two southern regional candidates (Stromm Thurmond and George Wallace) took states away from both parties. Dole peeled off Colorado and Georgia from Clinton, while losing Florida.

But this means for the electoral map is that any Republican who outperforms Romney in the popular vote will probably start with his 24 states.

The historical precedent is also that the next Republican nominee will outperform Romney, since parties have tended to peak and then steadily lose votes until the other side wins. Part of the subtext in discussions about how Hillary Clinton could perform in 2016 is that she's a better candidate than Barack Obama, which might allow her to defy history. Perhaps an underlying reason the Democrats steadily lost votes from 1936-1956 is that FDR was a better candidate than Truman, who was a better candidate than Stevenson. Eisenhower was a better candidate than Nixon, who was a better candidate than Ford. Reagan was a better candidate than Papa Bush. Bill Clinton was a better candidate than Gore.

If Hillary is actually is a better candidate than President Obama, the old rules might not apply. But I think Obama was a stronger nominee in '08 than she will be in '16.

[deliriumipa]

Wow, OP knows how to calculate an average.

[roadkill]

Historical trends have been used to correctly predict the Presidential election for years and years.  There's a very famous one that manages to get in the news every Presidential election cycle called The Keys to the White House model.  The problem with it and most historical forecasting models is that it can be subjective as compared to statistical models that have very little wiggle room for difference in interpretation. 

Still, they're mostly fundamentally sound.  Human behavior really doesn't change that much and for the most part will have reliable and predictable results for similar factors.

However ... Historical trends depend on that there hasn't been any kind of cultural or population shifts in the electorate.  I'm not sure you can rule that out here.