Monday, February 25, 2013

Political Hysteresis: Part Deux

[Guest Contributor - Donald G. Mutersbaugh, Sr.]

In a previous blog titled “The Great Republican Hope: Political Hysteresis” (, February 11, 2013), I presented the possibility that by just looking at the statistical relationships between Republican and Democratic presidencies over the years produced the possibility that a Republican would win the presidency in 2016. I began thinking about possible relationships between Democratic and Republican control of the Senate over the years. I decided to go back and analyze the data from 1861 to present regarding which party controlled the Senate. I once again have good news to report to the Republican Party: it may be their turn recapture control of the Senate in 2014 – at least statistically.

I previously decided to call this behavior Political Hysteresis: the tendency of an electoral outcome to vary about the central tendency of moderation based upon past and current candidates for the office of President. I am extending this definition now to include candidates for the Senate as a cohort analyzed by party (i.e., the Democrats and the Republicans). Granted, it is the candidate’s position and the Party's platform that drive the election. However, I would like to reiterate a simpler reason: hysteresis. “Hysteresis is the dependence of a system not only on its current environment but also on its past environment….To predict its future development, either its internal state or its history must be known.”]

Consider the following summary detailing which party was actually in control:

The matrix shows that Republicans have an average of 9 years of continuous term control; Democrats average 8 years (black is a correct prediction; red is an incorrect prediction). If you start in 1861 and add the average of 9 years to the ending year of control, you can correctly predict the Republicans in control subsequently 5 times, or 56%; using 8 years, you can correctly predict the Democrats in control subsequently 6 times, or 67%. (Note: I used ending years to predict the beginning year of control.)

However, if you start in 1861 and come forward in nine year increments, as a statistical projection, the Republicans will win in 2014 (i.e., 1861, 1870, 1879…2014). Unfortunately, the prediction accuracy is only slightly over 50%. The good news is that the Democratic prediction accuracy is 67%; it correctly predicted that the Democrats would win in 2013 (i.e., in 8 year increments 1869, 1877, 1885…2013) – but not in 2014! (Note: The years are slightly overlapping which produces 2013 vs. 2012.)

Now, to the interesting part; consider the following statistics. When there is a Democratic President and a Republican Senate, the Republicans controlled 52 seats and the Democrats controlled 47 seats (vacancies cause the unequal numbers). When there is a Republican President and a Democratic Senate, the Democrats controlled 55 seats and the Republicans controlled 43 seats. In the 113th Congress (2013- 2015), with a Democratic President and a Democratic Senate, the Republicans control 45 seats and the Democrats control 55 seats (the two independents are considered Democrats for this analysis). But this brings up an interesting question: since the Republicans are statistically projected to win in 2014, how will this occur since they will need to keep the existing numbers and win at least six more seats?

To answer this question, I did some research and found the following discussion of extreme interest: Twenty-one of the 35 seats up for election are now held by Democrats. Moreover, most [sic] the states that will be casting ballots for the Senate in 2014 are Republican leaning: 7 of the 21 Democratic-held seats are in states carried by the former Republican presidential nominee, Mitt Romney, while just one of the Republican seats is in a state won by President Obama.

I would encourage everyone to read the complete analysis by Nate Silver. It is excellent! He continues:

Are the conditions favorable enough to make Republicans odds-on favorites to gain six seats and win the Senate majority? Not quite. Six seats are a lot to gain, and Republicans are at risk of nominating subpar candidates in a number of races. But it would not take all that much to tip the balance toward them….

Summing up the possibilities across all 35 Senate races yields a net gain of four to five seats for Republicans, just short of the six they would need to win back the majority.

However, the margin of error on the calculation is very high at this early stage…. If Republicans swept all the “lean” and “tossup” races, they would gain a net of eight seats from Democrats, giving them a 53-to-47 majority in the 114th Congress. If Democrats swept instead, they would lose just one seat and would hold a 54-to-46 majority. Considering the uncertainty in the landscape, estimates from betting markets that Democrats have about a 63 percent chance of holding their majority appear to be roughly reasonable.

It is more than just a little bit exciting to see Political Hysteresis at work projecting a 53 to 47 Republican majority when one of the greatest American statisticians and psephologists, Nate Silver, opens the door to the possibility of a Republican win in 2014 – by the same numbers!

No comments: