It's Not "Just the Economy, Stupid" This Time

It's Not "Just the Economy, Stupid" This Time

By Sean Trende - July 13, 2010

It’s almost certain that the Democrats are heading for a terrible election in the fall – the only real dispute is whether it will be merely a bad election or a truly horrendous one.  And so, the pre-mortems have already begun.  Republicans are generally arguing that it is all about Obama’s policies, while Democrats argue that, like 1992, it is just the economy, stupid.

Of course, what we’re really arguing about here is presidential job approval, since it is the main driver of midterm elections.  Democrats’ strategy is simple:  appeal to the authority of political science models that supposedly make clear that the economy is the overriding factor in explaining presidential job approval.  Take, for example, The New Republic’s Jonathan Chait, who writes: “Right now, President Obama's approval rating is hovering just below 50%, about even with his disapproval rating. Given the state of the economy, that's not low. (I don't have any models handy.)” You can see The Washington Post’s Ezra Klein making a similar argument here.

There is no doubt that, as a general matter, the economy is an important factor in driving a President’s approval (this is also true for midterm elections, see my writing here).  But it is far from clear the economy is what is principally responsible for driving down President Obama’s approval rating and engendering a Democratic debacle in the fall. 

Setting aside the political science models for now, let’s just apply a little bit of common sense to the data.  People don’t approve of the way Obama has been handling the economy, and haven’t for quite some time.  But consider question 17 in this April CBS/NYT poll.  It asks who people blame mostly for the state of the economy.  Four percent blamed the Obama Administration, while a sixty percent blame the Bush Administration, Wall Street, or “someone else”.  It seems highly unlikely that when even the vast majority of Tea Partiers (10% blame Obama) refuse to lay blame for the state of the economy at the Administration’s feet, that it is what is dragging down his approval rating.

Let’s also take a look at the latest poll from Democratic polling firm Democracy Corps.  They allowed people who disapproved of the President (50%, in their poll) to give an open-ended response as to why they disapproved of the President, and created a word cloud for the responses:

As you can see, health care overwhelms the economy in these responses, and policy-driven answers such as “spending,” “oil spill,” and “everything” are about as salient as “economy.”  Indeed, on page 22 of the survey, the top reasons given for disapproving of the President were fiscal irresponsibility, health care, and the oil spill.  More respondents expressed concern that the President was a socialist/communist (8%) than were upset about jobs and unemployment (6%).

But since many of the “it’s the economy, stupid” arguments rely on an appeal to economic modeling to make their point, let’s examine an actual model.  There are a large number of models from which to choose, so to try to insulate myself from accusations of cherry-picking, I’ll use a piece Brendan Nyhan referred me to in an earlier back-and-forth: a 2002 article by Professor Brian Newman of Pepperdine. 

Newman’s model uses two economic indicators – inflation and unemployment – to predict average monthly Presidential approval.  It also utilizes variables for both positive and negative personal, domestic, and foreign policy events.  The coefficients for the statistically significant factors in Newman’s model, which explains about 90% of the variance in Presidential Approval ratings, are as follows:

Right from the start, we can intuit that economic variables are an important part of the equation, but not overwhelmingly so.  A positive domestic event can cancel out the effects of 11% worth of unemployment altogether.  Similarly, a negative domestic event is the equivalent of 6% of unemployment.

There are certainly examples of presidencies where a purely economic explanation of job approvals performs well.  Consider the front half of President Clinton’s second term.  If we run the equation assuming those two years had been completely neutral on the events front -- no positive events of any kind, nor any negative events – and thus isolate the economic variables, we get a chart that looks like this (as a methodological aside, I use the previous month’s predicted approval, rather than the actual approval (which Newman uses), because he and I are testing slightly different things):

We can get closer by accounting for events, but the purely economic model gets us within about 2 points on average.

But other Presidencies are not so well explained by the purely economic explanation.  Let’s take a look at the President to whom Obama’s plight is most frequently compared by the left, Ronald Reagan:

Here, the model misses by 6 points on average.  Reagan gets a boost in March of 1981 when he is shot (Gallup was in the field the next day), but even eight months later, he still performs about seven points ahead of where the purely economic model would predict.  And really, this just proves the point about events – and the way Presidents respond to them – being important, and even able to overwhelm poor economic conditions.  To prove the inverse point – that poor performance can override favorable economic conditions, consider the following:

What of Obama? Here’s the “pure economic” explanation charted out:

As you can see, from June onwards (perhaps not coincidentally, about when the focus began to fall on cap-and-trade and the health care bill), the purely economic explanation systematically predicts an approval rating for Obama that is higher than what actually exists.  The difference peaks at about 10% in September, then gradually narrows to about a three point difference today.  In other words, Obama has pretty consistently performed somewhere below where we’d expect him to perform given the state of the economy, sometimes substantially so. 

And again, this assumes that the public is treating a point of unemployment in March of 2009 the same way it treated it in, say, March of 1992.  I’m guessing the true gravitational pull of the economy early in Obama’s term was actually much less than the model suggests (and arguably still is today).  In other words, the prediction is probably a bit under-generous to Obama.

Now, let’s work in Obama’s domestic initiatives, while keeping our economic variables in place.  Let’s first put in a moderately negative narrative of his Administration’s initiatives:  we’ll assume that the public was largely positive on the stimulus, but was negative on the cap-and-trade vote, and was generally negative as the health care debate began and ran into the August recess and town hall meetings.  There was a break in the action in September and October, and then the public viewed the House and Senate votes on the health care bill as negative events.  Finally, the final passage of health care and the President’s response to the oil spill were viewed negatively.  Our prediction would look like this:

In other words, this looks pretty close to reality.  We end up somewhat lower than we would expect but the shape for February through June is pretty spot-on; if we inserted a positive domestic event around February (the health care summit? Obama’s appearance before the Republican caucus?) the model actually explains his approval ratings for the past three months within a point.  

We can quibble about whether we should include a positive event here (getting the Nobel in July) or a negative event there (the election of Scott Brown in January), but the point is that these policy issues – which I’ll add can best be evaluated qualitatively – are a very important part of explaining the course of Obama’s Presidency.

Finally, let’s see what the model would predict Obama’s approval ratings would look like if his governance had been well-received, even in spite of the poor economy.  We’ll assume that he enjoyed a typical honeymoon and that his early domestic agenda early had been quite popular*. Let’s also assume that Congress had quickly passed a modest health care bill in November of last year with some bipartisan support, that a modest energy bill passed in March, and that his response to the oil spill had been stronger:

If Obama’s Presidency had been marked mostly by achievements that were perceived positively by the public, he would be viewed quite positively today, even with a faltering economy.  We can also conclude, by the way, that the conservative narrative of Obama’s Presidency as an unmitigated disaster in every month is problematic as well; he’d be somewhere in the mid-thirties if that were the case, at least under the terms of this model. 

In other words, even accounting for the state of the economy, the nature of the policy narrative surrounding Obama’s Presidency is the difference between an approval in the high 50s and an approval in the high 30s.  That’s obviously a substantial difference.  

Now, you may come back with any number of other models that purport to show things differently.  For example, Professor Cuzan’s recent work tracks Obama’s approval nicely without accounting for any non-economic events during Obama’s administration (though one could argue that policy inputs are the reason that there’s a statistically significant difference between Republican and Democratic Presidents).  On the other hand, Newman’s appendix gives the results for different approaches to his data; some show the economic variables not having any significant effect.

The truth is that political scientists are a lot like economists:  They rarely agree on anything!  Sure, there’s some broad agreement as to the general notion that approval is driven by the economy, but we still haven’t sorted out the details as to which particular variables are the salient ones, and to what degree they drive the economy.  Indeed, with about fifty models out there using fifty different variables, all of which claim to explain 90% of the variance, I’m relatively certain that there will never be consensus on this.

The bottom line is that Obama’s approval ratings are almost certainly influenced by economic conditions.  But a controversial energy bill, a prolonged, contentious fight over health policy, and yes, even a “snakebit” response to the oil spill, have had a substantial effect on the President’s approval ratings.  If missteps continue, it could make the difference between a bad and awful midterm election for the Democrats.


*One way to do this might have been to break the stimulus up into several chunks – pass a tax cut bill, an unemployment insurance bill, an infrastructure development bill, etc.  He’d have gotten most everything passed, and probably would have gotten some Republican support for most of it; instead it was as if FDR had lumped together the FDIC, CCC, AAA, TVA and NIRA into a single Omnibus Recovery Act passed on the first day of the
73rd Congress, and then moved on

Sean Trende is senior elections analyst for RealClearPolitics. He is a co-author of the 2014 Almanac of American Politics and author of The Lost Majority. He can be reached at Follow him on Twitter @SeanTrende.

Why We Can't (Or Won't) Govern
Robert Samuelson · November 17, 2014
Marriage Is Pro-Growth
Larry Kudlow · November 15, 2014
Obama Touts Economic News as He, GOP Leaders Meet
Alexis Simendinger · November 8, 2014
A President Who Is Hearing Things
Richard Benedetto · November 12, 2014
Bret Stephens' Call for Robust U.S. Foreign Policy
Peter Berkowitz · November 16, 2014

Sean Trende

Author Archive

Follow Real Clear Politics

Latest On Twitter