Time for Change Model (Implications for local politics)

This week Politico featured an article The End of the 2016 Election Is Closer Than You Think    The Politico article is a fantastic read, but doesn’t go into the particulars that I would like to explore.

Yes, the Politico article in someways scooped a theme I have been working on for sometime on this blog.  In the past I have been exploring the formation of political environments and asking “Do Campaigns Really Matter?”

Time for Change Model – Predicting POTUS elections

There are many reasons for developing and using models.  Often models are used to present a hypothesis in a clear manner.  We argue about models, back test them, refine them, and then use them to predict outcomes.

One of the most interesting models in Presidential elections is Alan Abramowitz’s Time for Change Model based on what is now referred to as the campaign fundamentals.   Abramowitz has since revised his model, and we will look at both versions of the models.

The first Abramowitz model was:

 PV=47.3+(.107*NETAPP)+(.541*Q2GDP)+(4.4*TERM1INC)

  • PV stands for the predictive share of the majority party vote of the incumbent president
  • NETAPP stands for the incumbent president’s net approval rating (approval-disapproval) in final Gallup poll in June
  • Q2GDP stands for the annualized growth rate of real GDP in the second quarter of the election year, and
  • TERM1INC stands for the presence or absence of an first term incumbent in the race

“This basic model has correctly predicted the winner of the popular vote in the last 5 presidential elections with an average error of 2 percentage points.” -Abramowitz

So, why change the model?

Because in the last 4 Presidential elections, the basic model overestimated the winning candidate’s share of the votes.   “This suggests that the growing partisan polarization is resulting in a decreased advantage for candidates favored by election fundamentals including first term incumbents.: – Abramowitz

The revised Abramowitz model is:

 PV=46.9+(.105*NETAPP)+(.635*Q2GDP)+(5.22*TERM1INC)-(2.76*POLARIZATION)

  • POLARIZATION – takes on the value of 1 when there is a first term incumbent running or in open-seat elections when the incumbent president has a net approval rating > 0; it takes on a value of -1 when there is not a first-term incumbent and the incumbent president has a net approval rating <0.

“Adding the Polarization correction to the model substantially improves its overall accuracy and explanatory power” – Abramowitz

FRIENDLY REMINDER:  The outcome is determined by the electoral college, not the popular vote predicted by this model.

Where does the model stand now?

NETAPP – the net approval of President Obama

If we look at Gallup’s polling results for President Obama’s Approval Rating – we can see currently it is +3%.

(What is surprising for most in Conservative circles is just how much the President’s net approval rating has been positive for his two terms.   His average approval from the beginning of his first term to the writing of this post is 47%.)

Q2GDP (annualized growth rate of real GDP in the second quarter of the election year).  The number obviously has not been released yet, but we can look at trends.

GDP (% change from Preceding Period in Real Gross Domestic Product)
2013 2014 2015
Quarter I II III IV I II III IV I II
GDP 2.7 1.8 4.5 3.5 -2.1 4.6 5 2.2 -0.2  TBD

TERM1INC – We know there is no incumbent in this election, so we KNOW this variable is 0.

POLARIZATION – We assume President Obama has a net + approval rating, so the variable will be set to 1.

Thought Exercise

While the model specifically states the GDP needs to be from Q2 of the election year (2016), as nerds we can have some fun.

If we were to perform the calculation now simulating an election this year with the following variables:

  • using the current net approval rating of +3,
  • the Average of the GDP Change over Obama’s term of +2.175%, (using the -.2 would be cruel)
  • TERM1INC =0, and
  • POLARIZATION = 1,

we calculate the incumbent party (Democrat) predicted vote % to be 45.8%.

 

Variables
NETAPP          3.00
Q2GDP          2.18
Q2GDP Increment          0.50  ignore
TERM1INC 0 Presence or Absense of Incumbent (1 Incumbent, 0 no incumbent)
POLARIZATION 1 1 – first time incumbent or in open seat incumbent appr >0, = -1 no first time incumbent and incumbent approval <0
PV=46.9+(.105*NETAPP)+(.635*Q2GDP)+(5.22*TERMINC)-(2.76*POLARIZATION)
PV=party of incumbent % 46.9 (.105*NETAPP) (.635*Q2GDP) (5.22*TERM1INC) (2.76*POLARIZATION) Estimate Q2GDP
45.8 46.9               0.3150             1.3811 0 2.76                      2.18

Abramowitz Model

Below is an interactive model that we can explore, plugging in your own variables.

Type in the yellow boxes and the sheet will give you the results – (incrementally increasing the GDP by the variable you provide)

The Abramowitz Time for Change Model is also inserted on a clean page all to reduce clutter for your enjoyment.

 

Ramifications

PV=46.9+(.105*NETAPP)+(.635*Q2GDP)+(5.22*TERM1INC)-(2.76*POLARIZATION)

As you can see, the model weights the GDP much higher than the net approval rating (by a factor of 6x), but the power of incumbency is considerable.economy-stupid

The model provides the incumbent party with a base of 46.9%, then adjusts for Net Approval, then adjusts for GDP, then adjusts for incumbency advantage/incumbency fatigue.

For example, if President Obama is +8 on Net Approval, the GDP change would need to be +8% change in Q2 2016 for the incumbent, Democrat party to break 50% with no incumbent running. (Note don’t forget to change TERM1INC variable to 0)

Admittedly, it is too early to use a model that explicitly states it is Q2 of the election year, but we can clearly remember Carville’s “It’s the economy stupid!”

If you believe the model, the next 12 month’s events are critical in determining the 2016 outcome and there is little the 40 people running for President can do about it – except drive down the net approval rating of the President.  Get ready!

Ramifications for Local Elections

As we observe with my prior posts, Politico’s article, and the exploration of Alan Abramowitz forecasting model, the central thesis is this:  the political environment is formed for success or failure well before any candidate announces a run for office.   Campaigns do not change political environments but rather are a product of them.

These macro fundamentals are largely out of the control of Presidential candidates and even more so out of the control of state and local candidates.

However, local political environments can be shaped with local issues with a dedicated and consistent effort.   Therefore, I’ll say it again: Attention all interest groups and political actors interested in the upcoming local elections: To be strategic in local elections,  the time to form the political environment is a year before the elections, not the 6 weeks of a campaign before the election date.

About Alex Patton