Thursday, April 25, 2013

Tax policy colloquium, week 12: Larry Bartels' "The Class War Gets Personal: Inequality as a Political Issue in the 2012 Election"

This week, Larry Bartels presented the above paper, which relates to tax policy in the sense that it examines public views about tax progressivity, in the context of the 2012 presidential election.

Extra bonus fun: among the paper's data points was the impact of the Obama campaign's immortal "America the Beautiful" campaign ad, which depicted Romney crooning tunelessly while the words on the screen cited him for shipping jobs to Mexico and China, outsourcing other jobs to India, placing millions in a Swiss bank account, and investing in tax havens like Bermuda and the Cayman Islands.  The ad apparently had a very large impact on people right when they saw it - for example, inducing them to feel worse about equal opportunity in America and more negative about extending the Bush tax cuts for upper-income taxpayers.  But the ad's effects unsurprisingly degenerated over time (campaign ads have a short effective half-life).  Just seeing the ad again was a fond trip down memory lane for me.  Whether it mattered or not politically, it's truly an artistic masterpiece, and one could spend an entire seminar session(say, if my colloquium was "law and literature") unpacking its various levels of communication.

The paper's main focus, however, is on the following.  One natural way to view the 2012 presidential election is within a story frame of populist American responses to rising inequality that's accompanied by serious economic distress that arguably is the fault of the elite.  This is how many view the progressive and New Deal eras.  Then in 2012 we get the most anti-plutocratic presidential campaign in modern memory, with grade B Central Casting cartoon villain Mitt Romney starring in the plutocrat bad guy role.  The people rise up and endorse the Obama campaign's call for serious policy responses to rising plutocracy, including progressive tax changes.

This is a nice story, but is it true?  It's way too Frank Capra to be as convincing to the dispassionate observer as one might have liked.

In a way, this is what Bartels' paper examines.  It offers up two poster boys with alternative interpretations of the 2012 presidential election.  The first is Jonathan Chait, who said: "If there is a single plank in the Democratic platform on which Obama can claim to have won, it is taxing the rich."

The second is Kevin Hassett, who said: "I don't think the Obama victory is a policy victory ... In the end what mattered was that it was about Bain and frightening people that Romney is an evil capitalist."

Needless to say, both Chait and Hassett are playing the mandate game, because that's what people in Washington do.  Whenever you win an election, it gives you a mandate.  Whenever you lose an election, it turns out there wasn't any mandate.  Why people go to all this trouble is not entirely clear.  It's a nice talking point, but never really seems to matter much.  The Washington policy process is very substantially insulated from public sentiment.  Despite the high-minded idea that elected officials are agents of The People, broad public sentiment from the last election only matters if politicians conclude that it's stable and repeatable, and thus will drive results in the next election.  But as Harold Wilson once said, even a week is a long time in politics.  Two or four years are exponentially more so, and repeatability thus depends on inferring a long-term trend (a scenario that some Republicans appear to be at least evaluating with respect to immigration and gay marriage, but not tax progressivity).

Anyway, back to the paper.  Using survey data compiled from prospective and then actual voters during and right after the 2012 presidential campaign, Bartels finds that net support for the progressive tax changes that Obama endorsed was fairly slim and if anything narrowed during the course of the campaign.  He views this as rebutting the Chait position claiming a purposive public mandate, although one could also say that support for progressive tax changes "hardened" and "persisted."  But yes, it doesn't seem to have loomed all that large in voters' attitudes.

By contrast, a modified version of Hassett's thesis gains support from the survey data.  Where Hassett goes wrong is in suggesting (perhaps unsurprisingly for a disappointed Romney campaign adviser) that the Obama campaign was merely manipulating people by "frightening" them with a boogeyman story.  To the contrary, the public's view of Romney as a plutocrat who did not care about the interests of people outside his cadre was plain as a bell in late 2011, well before the Obama campaign started attacking him publicly.  They merely pointed to something that the public already believed. 

But "modified Hassett" still wins, in the paper's analysis.  The public responded very substantially to the view that Romney only cared about rich people, not anyone else.  Indeed, this much more deeply held version of the "evil capitalist" meme cost Romney, in the paper's best estimate, 2 to 3 percent in the popular vote.

This in turn suggests that, had there been no widespread public perception of Romney as a sneering, out-of-touch plutocrat, or if voters had not minded his being one (as indeed they might not have, in very different times), Obama's actual popular votye margin of 51.1% to 47.2% would have turned into something ranging from 49.2% to 49.1% for Romney to 50.2% to 48.1% for Romney.

I'm not sure I understand this counterfactual, in which the Republicans have a candidate with all of Romney's other perceived strengths and weaknesses (including, e.g., that he was non-ideological and a great "manager" who "understood creating jobs"), but not this weakness, even though it's arguably related closely to all of the other perceived attributes.

I also wondered why Romney does so well in this hypothetically revised version of the election, given that standard political science models based on the performance of the economy, especially over the last 6 months, tended to predict a small Obama victory.  But there may have been other factors leaning the other way.  For example, the public actually seems to have been worried about deficits and debt, and to have believed (I would say, in considerable tension with the actual campaign evidence) that Romney would have been more fiscally responsible than Obama.  So you could start from the prediction of the standard political science model, with Obama winning by a somewhat smaller margin than he did, than adjust his way for the plutocrat issue, then adjust back the other way for the issue of debt and deficits, leading to a net in which Obama wins by slightly more than expected.

But back to the points of more general interest here.  The paper suggests that revised Hassett beats Chait on the meta level as well as the direct one.  It finds that the evidence fails to support what I am calling the Frank Capra story - a fnding that I suspect might also hold if one could do a similar analysis of 1912 or the 1930s.  It finds instead that the public appears to have responded idiosyncratically at the personal level to Romney's trust deficit by reason of whom they thought he was, rather than what he was planning to do.

I tend to agree with this meta-conclusion, relative to the Frank Capra story in which an aggrieved public rises to demand serious policy responses to rising inequality and elite malfeasance.  But I was less confident that the paper actually provided strong evidence in favor of its meta-conclusion.  After all, one could imagine voter distrust of Romney the plutocrat as proxying for broader progressive views.  From the Capra perspective, why would rational voters, knowing how little in Washington politics they can actually control or even observe, nonetheless choose to focus on a particular issue, rather than on making judgments about the candidates' true characters and commitments?

In short, a U.S. voting public that attached a very high priority to addressing inequality arguably would have focused - just as the actual survey respondents did - on whose interests the candidates were inclined to serve.  The paper's meta-conclusion rests on a further assumption, which the data couldn't reach since the 2012 campaign only happened once, in which all that changes is the quirks of biography, and the public therefore doesn't reach the same conclusion about the Republican candidate even though he is the same as Romney in policy substance.

So in the end I agree with the paper's meta-conclusion because I agree with it, not because the adduced evidence seems conclusive in its favor.

No comments: