Sundstrom on McCarty, Poole, and Rosenthal, _Polarized America: The Dance of Ideology and Unequal Riches_

[eh.net-review at eh.net]

Published by EH.NET (August 2007)

Nolan McCarty, Keith T. Poole, and Howard Rosenthal, _Polarized 
America: The Dance of Ideology and Unequal Riches_. Cambridge, MA: 
MIT Press, 2006. xii + 240 pp. $35 (cloth), ISBN: 0-262-13464-0.

Reviewed for EH.NET by William A. Sundstrom, Department of Economics, 
Santa Clara University.


The marketing material that accompanied this slender volume included 
a ringing endorsement from Paul Krugman: "Essential reading for 
anyone who wants to understand what's happening to America." Perhaps 
so; but the lay reader expecting a journalistic treatment of the 
issues is likely to find the book very tough going. Densely packed 
with statistical results based on layers of complex calculations, the 
book itself dances around an explanation of recent trends in 
political and economic polarization in the United States without ever 
quite getting there.

The authors begin by presenting evidence that both political and 
economic polarization have increased in the United States since the 
1970s. The recent rise in economic polarization (income inequality) 
and the evidence for it will be familiar to many, and the authors 
largely take the trend as a given. The claim of increased political 
polarization is more controversial. Some widely held notions about 
political polarization, such as the purported deep ideological divide 
between red states and blue states, or the alleged recent widening of 
political and cultural divisions among American citizens, turn out to 
be more myth than reality (see Fiorina 2005; Glaeser and Ward 2006).

McCarty, Poole, and Rosenthal, however, focus on political 
polarization not among the populace at large but among politicians -- 
specifically, polarization as revealed by the voting patterns of 
congressional representatives. By their measures, ideological 
moderates have become increasingly rare in Congress, and the parties 
have pulled apart along the left-right spectrum. As Fiorina (2005) 
points out, this increasing ideological polarization among political 
elites and between the parties can create the illusion that the 
voters themselves are moving apart, even when the distribution of 
ideological positions in the electorate is stable.

Using a methodology derived from their important earlier work on 
congressional roll-call voting (e.g. Poole and Rosenthal 1997), the 
authors assign each legislator to a position or "ideal point" in 
ideological space. A representative's predicted vote on any 
particular measure depends on where her or his ideal point lies 
relative to a cut point (or surface) specific to the piece of 
legislation. The best estimates of each legislator's ideal point as 
well as the cut points on specific votes are essentially the ones 
that best predict the observed voting outcomes.

Political economists wed to the notion that all politics is 
special-interest politics might suppose that the dimensionality of 
ideological space is enormous, with perhaps as many dimensions as 
there are constituent or district interests. But in fact, confirming 
what the authors have found in their previous work, it turns out that 
a very large proportion of individual legislator votes can be 
correctly predicted using a single "liberal-conservative" dimension, 
which the authors refer to as the legislator's NOMINATE score. The 
definitions of liberal and conservative here are data-driven: the 
liberal position is whatever policies liberals tend to vote for 
consistently. Very often this is closely related to political party: 
in the 108th House, for example, every Democrat was to the left of 
every Republican on the first NOMINATE dimension. At certain 
historical moments, a second dimension adds a little predictive 
power. For the period under study in this book, the significant 
second dimension is largely related to votes over civil rights 
issues. But by the 1970s the contribution of these second-dimension 
issues to predicting legislators' voting patterns was rapidly waning.

Most of the time, the authors interpret the liberal-conservative 
"first dimension" as a function of the distributional implications of 
policy: Republicans/conservatives tend to vote the interests of the 
rich, and Democrats/liberals side with the not-so-rich. This will 
hardly come as news to most readers, but the authors do provide some 
compelling evidence to back up what we thought we already knew. For 
example, a legislator's conservatism is significantly and positively 
related to the average income of her constituents. And survey data on 
voter party preferences also reveal an increasingly strong 
correlation between income and Republicanism. Exactly how some 
evidently non-economic issues, such as gun control or abortion, came 
to be bundled with the economic class interests in party positions is 
not at all clear, and the pure empiricism of the NOMINATE approach is 
not well-suited to figuring it out.

Having presented the evidence for polarization among political 
elites, the authors finally get to the question that really interests 
them: what is the relationship between political and economic 
polarization? They frame the question using a highly stylized 
median-voter model of redistributive policy. The core prediction of 
the model is that redistributive transfers will be a decreasing 
function of the ratio of median to mean pre-tax income. Or to put it 
the other way round, as the right skewness of the income distribution 
increased in recent decades, the median-to-mean ratio fell, and the 
political process should have responded by taking more from the rich 
and giving it to the poor and middle classes. The logic of the model 
is straightforward: as the median-to-mean ratio falls, the median 
voter finds that her marginal gains from redistributive transfers 
outweigh the aggregate efficiency losses of an increase in the tax 
rate.

And here we come to the paradox that lies at the heart of the book: 
Contrary to the simple model's prediction, federal policies have if 
anything become less progressive over the past three decades -- or so 
the authors claim. McCarty, Poole, and Rosenthal do not actually 
provide a comprehensive assessment of trends in the net 
distributional impact of federal policies, but rather point to a 
couple of key policy areas, including the downward drift in the real 
value of the minimum wage and recent changes in the estate tax that 
largely benefited the rich. These policy shifts are also consistent 
with a slight overall shift in party preferences toward the 
Republicans. While their assessment seems plausible to me, the 
selectivity of the evidence here is unsatisfying. Over the same 
period, for example, the generosity of the highly progressive earned 
income tax credit (EITC) increased substantially, bucking the trend.

Assuming they are right about the facts, what can account for the 
failure of the simple model? The book offers two explanations. First, 
a correct interpretation of the median-voter model is that 
redistribution should depend on the income of the median _voter¬_, 
not the median individual or family. And while it is true that the 
income of the median American family has deteriorated relative to the 
mean since the 1970s, the same cannot be said of the median voter's 
income. The rich, for one thing, are more likely to vote than the 
poor. Just as important, recent increases in immigration have 
resulted in a disproportionate increase in non-citizens (and 
therefore non-voters) in the lower half of the income distribution.

Second, political polarization in the legislature has contributed to 
policy gridlock, making it more difficult to pass significant changes 
in policy progressivity. A variety of mechanisms lead to this result; 
for example, polarization decreases the ideological range of 
legislation that can overcome filibusters or presidential vetoes.

The most disappointing aspect of the book is that McCarty, Poole, and 
Rosenthal have not provided a compelling explanation of the very 
trend that they highlight: the increased ideological polarization 
among political elites. Like Fiorina (2005) and others, they deny 
that underlying ideological trends among the electorate are a driving 
force. They also downplay changes in congressional or party rules. 
Clearly the massive party realignment of the South played a role in 
moving Republicans rightward and Democrats leftward on average, but 
as the authors acknowledge, political polarization has increased even 
outside the South. They do identify a suggestive trend in 
congressional voting patterns: namely, that the correlation between 
constituent income and legislator ideology (NOMINATE score) has 
strengthened over their period. That is, the voting behavior of 
legislators increasingly reflects the economic class of their 
constituents. But exactly why remains an open question.

The authors of _Polarized America_ have crammed a lot of interesting 
ideas and a huge amount of data analysis into a very small package. 
The result is a book that is at once too technical for the typical 
lay reader, but too schematic and incomplete to fully satisfy the 
social scientist. Some alternative explanations are posited and 
tested rigorously, others largely ignored. For example, increased 
immigration plays a significant role in their story, while declining 
unionization barely warrants a mention. The median voter model that 
frames their entire discussion of redistributive politics simply 
assumes that the revenue from a proportional tax will be 
redistributed evenly across the population. Would the implications 
change if low-income voters could direct government transfers toward 
themselves and exclude non-citizens? Are there other models of 
redistribution in the political economy literature that would be more 
consistent with the evidence?

These shortcomings notwithstanding, the book offers numerous insights 
into recent trends in American politics, and it presents a variety of 
methodologies and hypotheses that will be of potential use to 
economic historians interested in policy issues. One presumes that 
the many unanswered questions merely set the stage for the next 
installment in a research project that has already contributed 
greatly to our knowledge of the history and functioning of the 
American political process.


References:

Morris P. Fiorina (with Samuel J. Abrams and Jeremy C. Pope), 
_Culture War? The Myth of Polarized America_. New York: Pearson 
Longman, 2005.

Edward L. Glaeser and Bryce A. Ward, "Myths and Realities of American 
Political Geography," _Journal of Economic Perspectives_ 20:2 (Spring 
2006): 119-44.

Keith T. Poole and Howard Rosenthal, _Congress: A Political-Economic 
History of Roll Call Voting_. Oxford: Oxford University Press, 1997.


William A. Sundstrom is Professor of Economics at Santa Clara 
University. His current research interests include the history of 
racial discrimination in U.S. labor markets and the causes and 
consequences of local public library development in the United States.

Copyright (c) 2007 by EH.Net. All rights reserved. This work may be 
copied for non-profit educational uses if proper credit is given to 
the author and the list. For other permission, please contact the 
EH.Net Administrator (administrator at eh.net; Telephone: 513-529-2229). 
Published by EH.Net (August 2007). All EH.Net reviews are archived at 
http://www.eh.net/BookReview.