Thomas Ferguson and Jie Chen wrote an article tying the recent Massachusetts Senate election to national political trends:
Passage of the health care reform bill has convinced some analysts that the Massachusetts Senate election might be a fluke. In fact, polls taken after the legislation passed show Republicans widening their lead in fall congressional races. This paper takes a closer look at the Massachusetts earthquake. It reviews popular interpretations of the election, especially those highlighting the influence of the “Tea Party” movement, and examines the role political money played in the outcome. Its main contribution, though, is an analysis of voting patterns by towns. Using spatial regression techniques, it shows that unemployment and housing price declines contributed to the Republican swing, along with a proportionately heavier drop in voting turnout in poorer towns that usually provide many votes to Democratic candidates. All these factors are likely to remain important in the November congressional elections.
Ferguson and Chen write:
It seems more likely that the citizens rallying under the Tea Party banner are pretty much what they say they are. They are ordinary Americans hammered by the almost Biblical series of economic plagues that for most began in the fall of 2008, when the decision to let Lehman Brothers go bankrupt turned a looming economic crisis into a world historical disaster. They have been driven to the breaking point by watching their jobs and retirement savings melt away as banks hit them with steeply rising fees on their credit and debit cards while paying next to nothing on what is left of their cash holdings. . . .
We are thus driven to conclude that the sometimes wild assertions and arguments advanced by Tea Partiers largely reflect the poverty of economic and political analysis in the establishment media. Indeed, the U.S. case bears an unsettling similarity to the situation in many parts of the parts of the Middle East. Political establishments and governments refuse to countenance critical discussion of social and economic problems. They marginalize alternative views, while beating the drums unceasingly for orthodoxy. When a crisis hits, however, no one believes them. So disaffected citizens set to work with the only tools they have – bits and snatches of traditional economic and political thinking – to analyze their predicament on their own. . . .
The key problem confronting the [Democratic] party is the one that has defined it since the halcyon days of the New Deal: the terms of trade between high finance and the rest of the party. Long before Obama won the nomination, it was plain from analysis of his early money that financial reform would not be a priority of his administration. Now, however, it is hard to see how the Democrats can hope to do much more than squeak through without revivals in both employment and housing markets. Both of these, however, are probably impossible without facing down Wall Street. . . .
The situation is no better – in fact, it may be worse – in regard to housing. . . . nothing has replaced the disastrous “shadow banking” system . . . since the Democrats now own this crisis, it may be Great Britain, 1931, when another coalition of Liberals and Labor lacking both courage and convictions endured repudiation at the hands of an electorate mortally afraid of both unemployment and budget deficits.
In some ways, this sort of work fills in the gap between general findings of political science and the blow-by-blow, issue-by-issue discussion in the news. Both levels are important–economic trends influence votes, but this happens in the context of particular issues. I find it useful to see these connections being made.
I wouldn't have thought it possible, but the analysis you link to is even stupider than normal punditry. I don't know which is worse, the inappropriate use of statistical analysis or the complete inability marshal a logical argument, but the combination is fairly stunning.
I think the core idiocy is the author's equating the differences between the 2008 Massachusetts presidential electorate and the electorate in the special election for Senate seat with the reasons for the "Tea Party" movement. For example, they say:
There is actually zero evidence in their paper for that conclusion. They present no evidence that that was true in Massachussetts and no reason to think that, if that was true there, why it would be true everywhere. The authors seem to ignore or perhaps they are unaware that the comparisions they use are completely inappropriate. Even though turnout was very high in the special election, it was still a special election with a single race on the ballot. Turnout patterns will be very different than in a Presidential election and in the direction they found. The only mildly surprising finding is the drop-off in towns with higher African-American populations was smaller than elsewhere. In any event, arguing that the fairly small correlations they found portend the end of the American political system seems a bit overwrought to say the least. To think that it explains the "Tea Party" movement is just bizarre.
The simplest way to say this is this equation:
"People who voted for Scott Brown" not equal to "Tea Party Movement".
That simple fact means that the entire analysis is completely off-base. Even if that were true, I fail to see how the rest of the analysis would make any sense. The authors make no coherent argument for their proposed solutions.
[It really pains me to say all of this because I'm very much in favor of the authors suggestions for what the Democratic party should do. I just don't think that there is any evidence that it will cost the party when they don't.]
I must agree that this article is quite overblown and makes too much use of rhetoric. "Massachusetts earthquake", "almost Biblical series of economic plagues", "The One, who had towered over the political landscape since
the presidential primaries in 2008, suddenly looked like Nothing." I can't seriously read a paper that uses such silly and sloppy language.
It is even worse that the sloppiness extends to statements that are refutable even by the most casual of historical or statistical thinking. Just to take an example from the cited sections above, the authors assert that "the decision to let Lehman Brothers go bankrupt turned a looming economic crisis into a world historical disaster". Sorry to ask rhetorical questions, but did these authors not notice the myriad financial troubles across the world going back at least a year prior to that event? It is the attribution of complex outcomes to overly simply but easily observable events that characterizes much poor statistical practice. Also, why the use of a Hegelian term like "world historical" anyway?
Yes, I agree that the rhetoric is off-putting and also that there are gaps in the reasoning–they don't prove their case (hence phrases such as "our data are consistent with" rather than "our analysis shows"). Still, I think this sort of analysis is interesting, and I wanted to share it, despite its flaws.
Comment: Thomas Ferguson
Reading these tendentious misreadings reminds me of the inscription on Dürer’s famous engraving of Erasmus: “The writings are a better likeness.”
Let’s start with “William Ockham.” He asserts that there is “zero evidence” in our paper supporting the claim by the losing candidate’s pollster that the administration’s failure to move vigorously against the banks and the recession badly hurt her candidate’s chances. That’s silly: We reached the same conclusion as another paper we referenced, that turnout in poorer towns fell disproportionately from 2008. The plain fact is that Democratic base did not turn out; we can’t prove that the factors the pollster mentioned are causal, but our analysis is consistent with it – which is all we said. It is interesting that in January, 2010, a Wall Street Journal blog reported that Ben Bernanke’s poll ratings were at a level approaching what other polls suggest for Congress and that people with a high school education or less opposed his confirmation by a 2-1 margin.
And we took care of Mr. Ockham’s objection that “turnout patterns will be very different than in a Presidential election and in the direction they found” before he voiced it. We were careful to compare the drop off in turnout between 2004 and 2006 with that of 2008 to 2010. The relationship with town income is patent only in 2010.
The rest of his discussion is simply noise. Who would possibly identify everyone who voted for Scott Brown as automatically a member of the Tea Party? There is nothing in the essay that would justify such a claim and it’s obviously foolish. After all, there were Republican voters in the state long before the Tea Party. It was wise of Mr. Ockham to use a pseudonym.
Jeff Shrader’s strategy is not well designed to burnish his case. He begins by confessing that he couldn’t bring himself to read the paper seriously. If he had, he would have found the references to the three papers I wrote with Robert Johnson on the financial crisis. Two are in the International Journal of Political Economy; the third appears in the Economists’ Voice (here: http://www.bepress.com/ev/vol7/iss1/art1/ ) The last is entirely devoted to counter-claims about the shock advanced by John Taylor and John Cochrane and Luigi Zingales. The allusion Jie Chen and I made in our paper to the uniquely disastrous role of the decision to let Lehman go bankrupt is common knowledge, as today’s (April 16, 2010) New York Times column by Paul Krugman reminds us. A particularly elegant way of making the point, though, is Richard Koo’s Exhibit 15 of his presentation to the INET Conference at King’s College, Cambridge, available here: http://ineteconomics.org/sites/inet.civicactions….
Much political science discussion resembles an airless room. Our discussion of what the financial crisis has done to the American people may strike some people as colorful, but it is hardly extravagant. Economists have recently awakened to the truly gigantic costs financial crises impose. As Johnson and I wrote in our INET paper, only wars really compare with them. (The paper is available here, at Session 8: http://ineteconomics.org/initiatives/conferences/… ; though it presumes the reader is familiar with the statistical studies of the BIS, the IMF, and other organizations.) Our view is right in the mainstream. You will find essentially the same position in Joseph Stiglitz’s Free Fall, Simon Johnson and James Kwak’s 13 Bankers or Carmen Rinehart and Kenneth Rogoff’s This Time Is Different. It is not far fetched to compare what has happened to many ordinary Americans since 2008 to a series of Biblical plagues. And the gigantic surge in the federal deficit means that they are not going to get much help from the government over the next few years. It is easy to understand why public anger is surging; social scientists need to wake up.
The Democratic base didn't turn out for Martha Coakley. That much is true. But that is not all that you claim to show in your paper which is titled: 1,2, 3, Many Tea Parties?. I quote from the abstract:
You make some specific claims that tie your analysis to the Tea Party movement:
Now, what is that, if not defining the Tea Party movement as the same thing as voters for Scott Brown? (and I might add your analysis is completely contradicted by recent polling of Tea Party supporters).
Your closing argument says:
What other conclusion did you expect people to draw other than you think that the same type of people who voted for Scott Brown are who make up the Tea Party movement?
I stand by my statement that you have zero evidence in your paper to support the central claim that it is Obama's fault that Coakley lost. Your evidence is consistent with increased volcanic activity in Iceland as well, but there's obviously no casuality. The factors you studied concern effects of the crisis, not the administration's rhetorical response. Your evidence indicates that lower incomes and falling housing prices were associated with lowered turnout among the Democratic base. Interesting and not very surprising. I will agree that the stimulus should have been larger and that might have helped Coakley, but unless you are willing to argue that the administration could have done something to reverse housing prices before the special election, I'm not sure what you think the administration could have done. In any event, your paper doesn't support the assertions of Coakley's pollster at all. Those assertions were largely about rhetoric. Here's the comment I assume you are referring to:
On my point about turnout, you didn't deal with it in advance. You didn't deal with it all. John Kerry didn't run in a special election in 2006. You may think that there's no difference between a normal midterm and a special election, but I think there is.
Finally, I want to be clear that I don't have any problem at with your policy positions. I have a problem with the complete incoherence of your argument.
57 percent of Tea Partiers have a favorable opinion of former President George W. Bush. 5 percent of Tea Partiers are Democrats. 4 percent of Tea Partiers consider themselves liberal.
Gotta love how a group of diehard Republicans managed to brand themselves as a revolutionary force in American politics. They probably still think that George W. Bush is a Washington outsider — after all, he and his father only spent 20 years in the White House.
Thomas Ferguson Comments
It is probably best if readers simply read our essay and make up their minds about what we actually said on their own. William Ockham’s rejoinder just continues his tendentious misreadings. How little attention he’s paying is wonderfully shown by this gem:
“Your evidence indicates that lower incomes and falling housing prices were associated with lowered turnout among the Democratic base.”
That’s a reference to our spatial regression – but anyone can see that the regression concerns the difference in the Democratic share of votes actually cast, not turnout. The turnout discussion is in another part of the paper and relies on entirely different evidence.
Why he is so obsessed with Celinda Lake’s rather anodyne observations is beyond me. But I will aver that I agree with many, many other observers that the Obama administration should have mounted a much more vigorous effort on housing, perhaps along the lines of the New Deal’s Homeowner’s Loan Corporation, along with a much bigger stimulus.
His reference to the recent New York Times poll also bears a comment.
The actual Times poll, unlike the story, contains clear evidence that Tea Partiers detest bailouts. Differences between Tea Partiers and all respondents in their responses to questions about whether the bailouts were necessary and whether they did any good are large indeed – 20% or more. Look here: http://www.nytimes.com/interactive/2010/04/14/us/…
Why the article does not mention these differences is a puzzle, since their potential importance is obvious.
Note also that the poll, or at least the portions of it published on the web, does not give respondents a chance to comment on what they believe about the Federal Reserve’s role in the financial crisis. This “don’t ask, don’t tell” approach to attributing responsibility limits the poll’s usefulness as a guide to what both the general population and Tea Partiers actually think.