Unequal Democracy is what good political science should be. It tackles an extraordinarily pressing problem with sophistication and subtlety, and casts powerful light on the barriers to effective expression of voters’ views in Washington. Based on extensive and very wide-ranging research, the book embodies an admirable conviction that on such a broad topic, multiple lines of inquiry and diverse sources of evidence are essential. The experience, creativity and insight of its author is on display throughout, and arresting findings practically spill off the pages. It is hard to imagine another social scientist pulling this off.
What’s more, the book’s disturbing bottom line – that voters face such difficulties monitoring and weighing in on policy and governance that they generally can provide only a very limited check on policy-makers – is persuasive and essential. Anyone who doubts that in the politics of inequality, political and economic elites generally have the upper hand will have to grapple with Bartel’s formidable book. And yet, we do think there are some major problems in the analysis offered in Unequal Democracy – inevitably so given the scale of the subject matter, but nonetheless with substantial costs to our understanding of the book’s vital subject. And since it is interrogation of the weaknesses in scholarly analysis and not just celebration of its strengths that leads to further scholarly progress, we will focus on the problems here.
In the interests of space, we mention just three (in ascending order of importance): citizen attitudes on taxes and economic policy, the role of partisanship in generating distributional outcomes, and the implications of voters’ limited capacity to constrain policy-makers. We provide much more detailed discussion of these and related issues in a recently published book.
First, as we have argued elsewhere, even as Bartels offers an extremely nuanced take on public attitudes towards taxes and other policy issues, he frames the matter in a way that suggests much higher levels of support for tax cuts than is justified. As Mark Hansen has argued, survey questions that simply ask whether voters favor a particular tax cut, without mentioning alternative uses of that money, are not very meaningful, because voters provide answers that simply don’t add up. Asked in this manner, voters favor virtually all tax cuts. Asked the same way, they also favor most spending increases. And they favor a balanced budget. To conclude based on this evidence, as Bartels seems to, that voters favored or supported a particular initiative (such as sharp reductions or elimination of the estate tax – one of the least favored tax reductions) seems suspect. To us, Bartels’s analysis highlights how critical the control of agenda-setting is, not that “Homer’s” desire for a tax cut had much to do with the 2001 debate.
We show in our 2005 book Off Center that Republicans were well aware of the need for such agenda control, with an internal White House memorandum stating flatly that voters preferred other policy ends to tax cuts and arguing that this meant that trade-offs should be taken off the table. Republicans’ success showed that if you can frame the question as “would you like this tax cut, yes or no?” you may win even if there are many alternative budgetary options that surveys suggest voters would prefer. Bartels’s rejoinder in Unequal Democracy that voters don’t get to decide how the agenda is structured suggests a much thinner view of democracy than we think he holds. And he sometimes veers into suggesting that voters don’t really have views at all. This was not true when it came to the priority they attached to tax cuts (the polls show remarkably consistency and clarity here). And it’s a position that seems at odds with Bartels’s arresting analysis elsewhere in Unequal Democracy when he shows that the opinions of less affluent Americans seem to have little influence on the votes of their senators. To look at the correspondence between public opinion and senators’ votes would seem to imply that people have opinions that senators can and should heed.
Second, we have real doubts about Bartels’s strong argument concerning the role of partisanship in the rise of inequality. It is not that we think the parties are identical on these matters – far from it – but we are skeptical about the huge effects that Bartels finds, and the main mechanisms he sees as driving the partisan divergence. Our concerns are two-fold. First, 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), and it has mostly occurred since 1980. Because Bartels relies on survey data on income that are topcoded, however, he has very little to say about the spectacular rise of high-end incomes. In fact, Bartels’s main measure of inequality is the 80/20 ratio, the ratio of income at the 80th and 20th percentiles. Yet the 80/20 ratio leaves out most of the story of rising inequality. According to the CBO post-tax and
transfer data, the 80/20 ratio rose from just over 3 in 1979 to 3.77 in 2005-an almost 25 percent increase. Over the same period, however, the ratio of the 99th percentile to the 20th percentile rose from 9.62 to 17.18—an almost 80 percent increase. And a very large share of rising inequality has actually occurred above the 99th percentile – the last generation has truly been one of winner-take-all.
As it turns out, when Bartels looks at the relationship between the 80/20 ratio and the partisan identity of the president, the association he finds is driven by the growth of income at the 20th percentile under Democratic presidents, not by the positive effects of Republican presidents on growth at the top. (In fact, the 95th percentile—the highest income group he looks at—seems to do about as well under either party, and in his regressions, Bartels never finds statistically significant partisan differences in income growth above the 40th percentile.) In other words, 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. This is an important insight—leaving open the question of whether it actually reflects the differing policies of Republican and Democratic presidents, as we discuss in a moment—but it does not directly address the issue of why growth has been so skewed toward the very top since the late 1970s.
Once we shift our gaze to the biggest fact about American inequality—the steady upward rise of the share of income going to the top 1 percent—a simple partisan story becomes much harder to sustain. Instead, something happened around 1980 that resulted in a fairly consistent upward trend in the fortunes of those at the very top, regardless of the partisan identity of the president (see Lane Kenworthy here). Bartels’s analysis has little to say about this trend, which we think is the crucial one for understanding the rise in American inequality. Not only does he focus on the 80/20 ratio; his analysis is also, by his own admission, much stronger at explaining the ebb and flow of the 80/20 ratio before 1980—that is, before most of the stunning rise in inequality. In other words, Bartels’s analysis is pitched at explaining a measure of inequality that systematically understates the growth of inequality at the very top, and his explanation works much better during the period before the big run-up in inequality occurred.
Bartels’s presidential partisanship story not only has serious problems explaining the sustained hyperconcentration of income since 1980, which has continued apace under both Republican and Democratic presidents. It also, we think, ascribes to presidents more independent influence over macroeconomic policy than they actually enjoy in our system of checks and balances in which the President must vie with Congress and the Federal Reserve for influence. Most important, Bartels’s does not, in our view, identify policies that can be plausibly linked the sustained run-up of top incomes. He suggests that the key tool presidents use to change the income distribution is fiscal and monetary policy. Democrats prime the pump, so to speak, with the consequence that working-class income growth is higher under Democrats than under Republicans.
Such contractionary and expansionary initiatives, however, seem a better candidate for explaining short-term fluctuations and patterns of income growth than long-term changes in the income distribution, especially sustained gains at the top rather than growth at the bottom. It is doubtful, for example, that the expansive fiscal and monetary policies that bump up lower-income workers’ employment and incomes could be sustained indefinitely. (Indeed, Bartels’s argument about Republican success in producing growth just before elections rests on exactly this kind of short-term dynamic). By contrast, long-term shifts in the income distribution are more plausibly linked to changes in the structure of the economy than to whether or not the economy is operating at peak output levels at any particular time. And yet the role of government policy in creating these larger structural changes – a role that we believe is very extensive (involving areas like financial regulation, corporate governance and industrial relations), and runs through both parties (though in different ways) is notably absent in Bartels’s account.
Third, and most important, we would stress that Bartels’ focus on voters can, at best, only bring us a part of the story. It tells us that voters are not driving policy outcomes, but it does not tell us who is. Here, Bartels repeatedly invokes “the immense significance of elite ideology in the making of American public policy” – but this turns out to be merely a residual category. While we think ideology is important, it is probably not the only or even the main thing driving elected officials. Instead, we would highlight the crucial role of organized interests in American politics, and in particular the mobilization of business and corporate groups from the 1970s on, the weakening of organized labor, and the rise of a large constellation of advocacy organizations on the left focused not on material, pocketbook issues but post-material, quality-of-life issues.
Indeed, we are struck by the very limited attention that Bartels and other Americanists have paid to organized interests in trying to explain the rise in income inequality – a stance that departs sharply from decades of work in comparative political economy. In Unequal Democracy (as well as McCarty, Poole and Rosenthal’s Polarized America), unions and corporations are hardly mentioned. Unions have four references in the index of Bartels’s book—two related to the minimum wage—while the National Election Study has twenty. Yet in Bartels’ striking chapter on the minimum wage – a chapter that fits somewhat awkwardly in the book – he concludes that union decline alone accounts for 40% of the real decline in the minimum wage. And this is just one policy area, and ignores the most powerful organized actors in the political economy – businesses.
Our three points are linked. They all suggest that the marvelous dissection of voters and public opinion in Unequal Democracy succeeds in explaining who isn’t driving policy outcomes. Now we need to understand who is. Here the work of comparative political economists, who have stressed the much wider interconnections between public policy and structures of economic rewards, as well as the centrality of organized economic interests to policy outcomes, needs to come to the fore. As Bartels’s brilliant book shows, political scientists have pushed the study of public opinion and individual political behavior to new heights of sophistication and insight. But from these heights, it is often hard to see the organized combat that is taken place in the trenches of American politics on a daily basis. It’s time for us to look more closely.