The False Premise of
Bartel's Unequal Democracy:
Setting the Record Straight
In 2008, Princeton University Press published Unequal Democracy: The Political Economy of the New Gilded Age by Professor Larry Bartels, then on the faculty of Princeton University and now at Vanderbilt University. The premise of the book was that the economy since WWII had fared better under Democratic presidents than under Republican presidents. Real economic growth had been greater, unemployment lower, and income inequality had not been as great during the terms of Democratic presidents. As promotional materials for the book put it, Bartels purported to show that "the gap between the rich and poor has increased greatly under Republican administrations and decreased slightly under Democrats, leaving America grossly unequal." This last economic effect was of particular interest to Bartels and raised an interesting puzzle that he explored at length in the book: if the economy is so much better under Democratic presidents and if the economy matters so much to voters, why don't voters (particularly middle class and poor voters) consistently elect Democrats as presidents? How do Republicans ever win?
To say that the book was well received is a gross understatement. Political scientists were extravagent in their praise. The APSA journal Perspectives on Politics ran a "Review Symposium" on the book that was wall-to-wall praise from eminent political scientists. Presidents Clinton and Obama also publicly praised the study. For a U.S. President to praise a work of political science is a highly unusual event. For public praise to come from two presidents is unprecedented. To top off Unequal Democracy-fest, the book won two highly coveted book awards. It was the 2009 winner of the Gladys M. Kammerer Award from the American Political Science Association for the best book published in the field of U.S. national policy and the 2009 Leon D. Epstein Outstanding Book Award from the Political Organizations and Parties Section of the APSA. Princeton University Press compiled more of the accolades at http://press.princeton.edu/quotes/q8664.html.
I was quite skeptical of Bartels' findings of a Democratic Party superiority in economic records and had a lengthy exchange with Alan Abramowitz at Emory University about them (first as conveyed by Michael Kinsley in a Slate article). This exchange was later picked up by Daron Shaw in his review of Bartels' book for The Forum. My initial analysis of the data indicated that Bartels' finding of a presidential party difference on the economy (whether growth, unemployment, or inequality of incomes) depended heavily on the choice of lags from when a president took office to when it could be reasonably said that he was responsible in some respect for economic conditions. Vary the lags a bit and the findings disappear.
Further investigation into this uncovered the reason why Bartels' results were so fragile. The reason is that each of the four new Republican administrations since WWII (Ike in 1953, Nixon in 1969, Reagan in 1981, and G.W. Bush in 2001) took office when the economy was heading into a recession. This was not true of new Democratic administrations. Thus, the presidential party difference that Bartels built his book around was NOT the consequence of policy differences between the political parties (with a mean old Republican Party driving the economy into the ground and bloating the economic gap between upper and lower income Americans). It was the result of the presidential parties INHERITING economies in very, very different shape.
Table 2 below from my Presidential Studies Quarterly 2012 article (an article I wrote in response to an article by Comisky and Marsh that went to great lengths in an ultimately unsuccessful attempt to sustain Bartels' conclusions) displays real GDP data on this score. Real GNP per capita data (the measure used in Bartels' original analysis) is presented in Table 6 of my 2011 Forum article (available below).
The two articles below provide the full reanalysis of Bartels' claims. I should note that the Forum piece uses data drawn from the same sources as those used by Bartels in his original study. My articles demonstrate, I think conclusively, simply, and without resort to arcane methodological contrivances, that the foundation of Bartels's analysis is NOT supported by a careful analysis of the facts. While he carefully probed his findings of presidential party differences, he did not go quite far enough in this probing. Once the effects of the inherited economy are properly taken into account, the presidential parties do NOT differ in their economic records--not in terms of economic growth, not in terms of unemployment, and not in terms of economic inequality.
I will spare you an account of the gross unprofessionalism that I met in the course of getting this research into print -- at two of the most highly respected journals in political science no less. I will say only that anyone who believes that the ideological hegemony in political science is not a serious problem is hopelessly naive. I cannot even imagine the treatment I would have received had my findings documented Republican presidents as having superior economic records. I will also note that, to date, Presidents Clinton and Obama have not yet endorsed these corrected findings and the APSA has not yet given them any award. On a positive note, I would like to thank the editors of The Forum and Presidential Studies Quarterly for publishing these two reports of this research and to Bill Keech for picking up on this research for his new edition of Economic Politics in the United States: The Costs of Democracy (Cambridge, forthcoming 2013).
Key Tables:
1. No Party Difference in Economic Growth Once Inherited Economic Conditions are taken into account
2. No Party Difference in Unemployment Once Inherited Economic Conditions are taken into account
4. Real GNP per capita growth rates around Presidential Party Transitions, 1948-2009
Full Articles Refutting Unequal Democracy:
The data for these articles are posted on my website.