Much ado about nothing: Obama will be president, again, in 2013

Many of us have watched transfixed as the Republican Party participates in a ‘no holds barred’ fratricide primary season. The heir apparent—Mitt Romney—is a weak front runner, at best, but has taken shots from all the conservative pretenders that have tried to supplant him—first Bachmann, then Perry, Cain, Gingrich, and finally Santorum. However, the smart money remains on Romney to take the nomination because he leads in all the important leading indicators: money, delegates, electability, perceived leadership ability, etc. Intrade, that handy barometer of conventional wisdom, has Romney’s chances of winning at 93%.

Ultimately though, does any of this drama really matter? Will it really affect the outcome of the election in November?

The short answer is no. This primary doesn’t mean much. Indeed, the Republican primaries are nothing more than an interesting side-show to an eventual Obama victory. Our own political forecasting model, based on hundreds of elections around the world, indicates that Obama has about an 85% probability of winning reelection and re-taking the White House. Yes, there is a chance that Obama might falter, but it is a small one and the Republican candidate has little bearing on it.

Why? Let me explain the logic.

(1)    Presidential elections at their core are about voters’ relative ‘desire for change’. When things are going well, voters want ‘continuity’ and typically go with the government (party-in-power) candidate. I call these continuity elections. In contrast, when things are going poorly, voters normally want a ‘change’ and turn to the opposition (party-out-of-power) candidates. I call these change elections.

Change Elections

(2)    Continuity elections usually happen when the economy is doing well, the administration has avoided political scandals and military actions have been successful or brief. Change elections are those where the economy is in the gutter, body counts are increasing, or political scandals have just exploded.

(3)    About 30% of all elections are neither change nor continuity but something in the middle. I call these middling elections. Middling elections are characterized by marginal economies and discontent with foreign affairs. Middling elections are a bit trickier because other factors play a role, such as incumbency, power of personality, and the efficiency of marketing and communications.

(4)    Roughly speaking, desire for change can be measured in any number of ways including right track/wrong track questions, presidential approval ratings and economic optimism. If we use Presidential approval ratings as our indicator, change elections correspond to approval ratings below 40 points; middling elections range from 40 to 50 points, and continuity election 50 points or more.

Change Elections Middling Election Continuity Elections
Approval ratings 40 or less Approval Ratings 40 to 50 Approval Ratings 50 or more

(5)    Incumbent candidates have a decided electoral advantage. Indeed, our data shows that, on average, incumbent candidates have almost 4-fold (3.6) advantage over opponents.

Probability of Incumbent winning by type of elections

*employ estimated probabilities from a logistic regression model, and not raw percentages, in our analysis. Inferentially, both the raw numbers and estimated probabilities show the same trend lines. To do this, I employed a simple three-variable model where the dependent variable was binary (whether or not the government candidate won the election) and the independent variables included (a) our three-category classification scheme (operationalized as dummy variables) and (b) whether the government candidate was an incumbent or a successor (See equation below). Pr(victory) = Dummy (Change election) + Dummy (Continuity election) + Dummy (Incumbent).

SOURCE: Ipsos modeling based on 187 executive branch elections in 35 countries; percents are estimated probabilities*

So, given this context, how does Obama stack up against this model? Quite well.

If current conditions maintain, the November 2012 election will be a middling scenario. Presently, Obama has an average approval rating of about 47 percentage points (RealClearPolitics.com, 3/22/12), which puts this election in the upper part of the middling scenario range. If his approval numbers were to hold at this level, Obama would have about an 85% change of winning in November. If the winds of political and economic fortune push his approval rating above 50 percentage points, Obama’s chances of winning go to almost 99%. Many economic forecasts predict continued moderate growth in 2012 making this scenario more probable (http://www.conference-board.org/data/usforecast.cfm; http://macroadvisers.blogspot.com/search/label/GDP; http://www.kiplinger.com/businessresource/economic_outlook/).

So what are the implications of this?

With a likely Obama victory, our analytical focus should be on what an Obama government might look like. Raising important related questions, including:

Will the Republicans keep the House?

  • Will the Democrats hold the Senate?
  • What will the policy Agenda in 2013?
  • Ultimately will Obama be able to govern or should be we expect gridlock for the next 4 years?

Stay tuned for answers to these questions.


*we employ estimated probabilities from a logistic regression model, and not raw percentages, in our analysis. Inferentially, both the raw numbers and estimated probabilities show the same trend lines. To do this, I employed a simple three-variable model where the dependent variable was binary (whether or not the government candidate won the election) and the independent variables included (a) our three-category classification scheme (operationalized as dummy variables) and (b) whether the government candidate was an incumbent or a successor (See equation below). Pr(victory) = Dummy (Change election) + Dummy (Continuity election) + Dummy (Incumbent).

 

Comments are closed.