Response to Paul Pierson and Jacob Hacker

Inequality is a complex and multifaceted phenomenon. Jacob Hacker and Paul Pierson’s critique of Unequal Democracy seems to hinge in large part on a simple disagreement about which aspects of that phenomenon are most worthy of study. They write, “to us the critical rise in inequality is mostly at the top of the income distribution (where the top 1% has pulled sharply away from everyone else).” This is “the biggest fact about American inequality,” they say, the “crucial” trend; but my book “has very little to say about the spectacular rise of high-end incomes.” In their contribution to a recent special issue of Politics & Society, which expands on several of the points they make more briefly here, they complain about the “misplaced focus” of economic and political accounts, including mine, that emphasize “the broad spreading out of the income distribution, rather than the hyperconcentration of income at the top.”

To Hacker and Pierson, “Bartels’s argument about partisanship boils down to the claim that those on the bottom portions of the income ladder do much worse under Republicans than under Democrats.” Perhaps their diligent attention to the rarified realm of the top 1% has warped their sense of what it means to be “on the bottom portions of the income ladder.”

My tabulations showed that, since the late 1940s, the real incomes of families at the 80th percentile of the income distribution have grown 70% faster under Democratic presidents than they have under Republican presidents. The real incomes of families at the 60th percentile have grown more than twice as fast under Democrats. Families at the 40th percentile have experienced three times as much real income growth under Democrats.

These differences are enormous, and surprisingly robust given the limitations of the available data. Hacker and Pierson seem to want to dismiss them, asserting that “in his regressions, Bartels never finds statistically significant partisan differences in income growth above the 40th percentile.” That is a simple but rather crucial misreading; the t-statistics for partisan effects in my key table (2.3) are 2.9 for families at the 40th and 60th percentiles and 2.4 for families at the 80th percentile, all well beyond any reasonable threshold for “statistical significance.”

Cumulated over more than half a century, these partisan differences have been hugely consequential for the economic well-being of the broad American middle class. The partisan differences in income growth for families who really are near “the bottom portions of the income ladder” have been even larger in percentage terms—and even more consequential in human terms. To think of all this as a sort of sideshow to the real story of hyper-concentration of income going to the top 1 percent strikes me as an instance of “misplaced focus,” to borrow Hacker and Pierson’s phrase.

A second major theme in Hacker and Pierson’s critique is that my analysis is unduly focused on voters and elections—as they put it in their longer essay, on “politics as electoral spectacle” rather than “politics as organized combat.” This is a very plausible complaint, given the inevitable limitations of my own expertise. However, I think they push their point too far when they suggest that Unequal Democracy “tells us that voters are not driving policy outcomes, but it does not tell us who is.”

For one thing, there is ample evidence in the book that voters are often driving policy outcomes—by electing Democratic or Republican policy-makers. The movement to repeal the estate tax involved arduous effort by organized interests, but it is no coincidence that that effort only began in earnest when Republicans won the House in 1994, and achieved (temporary) victory within months of the election of a Republican president in 2000. Similarly, the Bush tax cuts were championed by powerful interest groups, but they never would have seen the light of day if a thousand more voters in Palm Beach had succeeded in voting for Al Gore.

In both these cases, it seems unhelpful to think of “politics as electoral spectacle” and “politics as organized combat” as mutually exclusive alternatives. The interplay of these factors is even clearer in my chapter on the eroding minimum wage. As Hacker and Pierson note, my analysis of minimum wage policy-making assigns a major role to the declining clout of labor unions. But partisan control of the White House and Congress also mattered substantially. Moreover, the terms of “organized combat” seem to have been significantly shaped by partisan politics: the estimated impact of labor union strength was almost twice as great under Democratic presidents as it was under Republicans, and the estimated impact of the partisan composition of Congress was three times as great under Democratic presidents as it was under Republicans.

There is, of course, much more to be learned about the role of labor unions, businesses, and other powerful organized interests in the politics of inequality, and in contemporary American politics more broadly. Thus, I enthusiastically second Hacker and Pierson’s closing call to “look more closely” at “the organized combat that is taking place in the trenches of American politics on a daily basis.” The primary reason why I haven’t done more of that myself is that it is tremendously difficult; the inferential challenges involved in using observations of “politics as organized combat” to “tell us who is” driving policy outcomes—and how those policy outcomes matter—are enormous. Fortunately, Hacker and Pierson’s own terrific new book, Winner-Take-All Politics, provides an excellent full-scale model of the sort of analysis they advocate, and an excellent sense of how richly analyses of this sort can contribute to our understanding of the political process.

3 Responses to Response to Paul Pierson and Jacob Hacker

  1. Sebastian December 8, 2010 at 10:59 pm #

    Hacker and Pieson (as well as Wilkinson) also challenge you on the mechanisms behind those effects you’re finding. I think that’s a major concern for your findings about the partisan effects of income growth in various income classes.
    I just find it very hard to buy those huge effects without a compelling explanation of how Democrats (and Republicans) actually cause them. I don’t want to rely on the old “correlation is not causation” canard, but without a plausible mechanism that really is what it comes down to.
    I was hoping you’d address that in your post, or maybe your upcoming one on Wilkinson.

  2. anon December 9, 2010 at 5:53 am #

    For those who don’t carefully read and remember every post on the Monkey Cage, in chronological order, here’s the link to Hacker and Pierson’s critique:

    (No offense meant, I just read you all somewhat irregularly, in Google Reader. So I was lost when I started reading your post.)

  3. Larry Bartels December 9, 2010 at 3:58 pm #

    Sebastian—I certainly agree that there is much more work to be done linking specific policies with their distributional effects. (I’d say the same about Hacker and Pierson’s book, by the way.) This is not a task for one scholar in a couple chapters of a single book that is mostly about other things; it will require substantial sustained investigation by people with much more policy expertise than I have. In the meantime, however, I think the tendency of some critics to portray the partisan pattern as merely a correlation plucked out of thin air is misleading. In particular, I refer you (and Hacker and Pierson, and Wilkinson) to (1) evidence that the partisan differences are not simply the result of exogenous trends or economic shocks or partisan turnover (pages 34-42); (2) evidence that the partisan patterns of income growth are mirrored in partisan differences in macroeconomic performance (pages 47-52); (3) evidence that the timing of partisan differences in income growth is highly consistent with a political explanation (pages 52-54); (4) examples of differences in macroeconomic policy under Democrats and Republicans over the past half-century that are alleged by policy analysts to have significant distributional implications (pages 42-47); and (5) evidence that there are significant partisan differences in the specific domain of minimum wage policy—differences large enough to account for some substantial fraction of the striking partisan differences I have documented in income growth for working poor families (Chapter 8). You might also consider whether policy developments in the two years since my book was published shed any additional light on the notion that Democrats are more likely than Republicans to produce income growth for middle-class and poor people. Alan Blinder and Mark Zandi estimated that the 2009 stimulus package increased GDP by 3.4%, reduced unemployment by 1.5%, and produced 2.7 million jobs. Do you suppose that we would have seen a similar stimulus package—or a strikingly progressive 2011 budget, or a massively redistributive health care reform—under President McCain?