Many Americans were shocked and shamed by the televised images of New Orleans in the wake of Hurricane Katrina. Scenes of physical devastation were juxtaposed with even more disturbing scenes of human devastation. Thousands of residents, mostly Black and poor, seemed to be trapped in a Hobbesian state of nature, abandoned by government and civilized society. An untold number starved to death; for days their corpses floated in the streets.

According to one press report (DeParle 2005a), “What a shocked world saw exposed in New Orleans … wasn't just a broken levee. It was a cleavage of race and class, at once familiar and startlingly new, laid bare in a setting where they suddenly amounted to matters of life and death.” Another (Tilove 2005) asserted that “Katrina's whirlwind has laid bare the fault lines of race and class in America. For a lightning moment, the American psyche was singed.”

In the immediate aftermath of the storm, some observers suggested that the shock and shame of Katrina might “strip away the old evasions, hypocrisies and not-so-benign neglect” surrounding issues of inequality in America (Alter 2005). “For the moment, at least, Americans are ready to fix their restless gaze on enduring problems of poverty, race and class that have escaped their attention.” President George W. Bush promised “bold action” to “confront” poverty and racial discrimination. However, the hope for a silver lining proved to be short-lived. Less than a month later, amid reports of ideological wrangling over federal and local responsibilities, tax cuts and budget deficits, labor law and housing policy, the lightning moment seemed to have passed (DeParle 2005b).

Meanwhile, Robert Weissberg (2006, 33) pokes gentle fun at the “alarmist tone” of the public pronouncements issued by the APSA's Task Force on Inequality and American Democracy. “Have these distinguished professors possibly discovered a conspiracy to cancel elections, repeal the First Amendment, or otherwise impose a racist plutocracy? Are retired generals and multinational CEO about to launch a coup d'état? … Again and again, the message is unmistakable—deep economic disparities are killing democracy, and we better watch out or it might be the fire next time!”

Not the fire, Professor Weissberg, but the flood. And not a cataclysmic Katrina-style flood. I, at least, am not lying awake at night worrying about a coup d'état or a proletarian uprising. The real threat to democracy, it seems to me, is from constant, gradual seepage through the myriad cracks in the levees separating our putatively democratic political system from the powerful waters of a vibrant capitalist economy.

Are the cracks getting wider? We think so, but we don't know for sure. The sad truth is that we—political scientists—have not been sufficiently conscientious and clever to make the careful, repeated measurements that would be necessary to prove the point. We do have some basic training in plumbing, though. We have begun to do some hasty remedial probing of likely weak spots in the levees. And some of us have the common sense to worry when our feet are getting wet.

Of course, not everyone is alarmed. Weissberg's most fundamental complaint about the Task Force's work is that “The economic inequality/democracy nexus hardly constitutes a pressing public concern requiring disciplinary attentiveness” (36). I am not entirely sure whether that is intended as an empirical criticism or a normative one. If the former, it is certainly true that most Americans get through most days without seriously contemplating the moral or political ramifications of economic inequality. On the other hand, they probably spend even less time worrying about Weissberg's “far more urgent” issues like the Electoral College and non-citizen voting. So I suspect that Weissberg has some normative standards of his own for gauging which public problems are worthy of disciplinary attentiveness.

Weissberg does concede (34) that “everyone agrees that ‘responsive government’ is a splendid idea.” He goes on to add, however, that “the term invites so many murky meanings as to be analytically useless unless precisely defined.” Fair enough. Here is a stab at a precise definition. Let us stipulate that “responsive government” implies a systematic correspondence between the preferences of citizens and the policies adopted by their elected representatives. Let us further stipulate, following Dahl (1971, 1), that “a key characteristic of a democracy is the continuing responsiveness of the government to the preferences of its citizens, considered as political equals.” That would seem to imply that the democratic status of a political system could be compromised—perhaps even “threatened,” to revert to the Task Force's alarmist language—either by non-responsiveness or by grossly unequal responsiveness.

How does the American political system measure up to this standard? The Task Force's report makes brief reference to a series of empirical studies of disparities in responsiveness, including my own unpublished work on economic inequality and political representation (Bartels 2005). Figure 1—a slightly revised version of a figure presented in the volume of literature reviews supporting the Task Force report (Jacobs and Skocpol 2005, 126)—summarizes the key finding from that work.

The bars in Figure 1 represent regression estimates of the extent to which senators' roll call votes in three successive Congresses (1989–1994) reflected the ideological views of their constituents in the lower, middle, and upper thirds of the income distribution. (Detailed descriptions of the data and analysis are available from Bartels 2005.) The estimates suggest a good deal of responsiveness to the views of middle- and upper-class constituents; the implied difference in voting behavior between senators representing the most liberal and most conservative states is roughly equivalent to the difference between Democratic and Republican senators representing the same state. What is more striking, though, is that there is no evidence of responsiveness to the views of constituents in the bottom third of the income distribution. These data suggest, rather remarkably, that senators in three successive Congresses completely ignored the preferences of millions of constituents who happened to have modest incomes—not just gave their preferences less weight, but ignored them entirely.

Lest these results seem implausibly pessimistic, I should note that they are closely paralleled by the findings of a separate study, also unpublished, conducted by Martin Gilens (2005). Gilens's research design employed an extensive collection of national surveys on a wide variety of policy issues. For each issue, he examined whether a policy change supported or opposed by various segments of the public was subsequently adopted. He found a strong relationship between the views of affluent survey respondents and the subsequent course of public policy. For less affluent survey respondents the relationship was weaker; and when the analysis was limited to issues where rich and poor respondents had divergent preferences, Gilens found that the well-off were vastly more likely to see their views reflected in subsequent policy changes.

Of course, the fact that two unpublished papers say it doesn't make it true. Still, the coincidence of findings is striking. Let us imagine, for the sake of argument, that this empirical evidence of disparate responsiveness withstands rigorous scientific scrutiny to Weissberg's satisfaction. What would that suggest about the state of American democracy? Not that we are in the grip of a plutocratic cabal. Cartoon images of banana republics and ghostly power elites are beside the point. Political leaders appear to be responding significantly to the policy preferences of millions of citizens. Dahl's term—“polyarchy”—seems entirely apt as a description of this pattern.

On the other hand, the system portrayed in Figure 1, and by Gilens, comes nowhere close to approximating Dahl's democratic ideal of “continuing responsiveness of the government to the preferences of its citizens, considered as political equals” [emphasis added]. If we agree that such responsiveness is “a key characteristic of a democracy,” it seems less alarmist than complacent to “Warn of Threat[s] to American Democracy,” as the APSA's lurid media advisory did (APSA Press Release 2004). Gilens (2005) himself concluded that “representational biases of this magnitude call into question the very democratic character of our society.”

Is the water rising? There are some reasons to think so; but again, we don't really know. The pattern of responsiveness summarized in Figure 1 is for a single governmental institution in a single six-year period. Gilens's work is somewhat broader in scope, and he is currently gathering and analyzing data that may provide a clearer picture of how disparities in representation have (or have not) varied over decades. In the meantime, however, this much seems clear: if middle-class Americans have wet feet, millions more who are less affluent are already in it up to their necks.

Under the circumstances, it does not strike me as imprudent to inquire about flotation devices. Despite its “over-heated radical egalitarian tone” (Weissberg, 34), the Task Force report contains very little discussion of current political controversies and no specific policy recommendations.¹

It does include some stirring language about equalizing political participation; but Weissberg (37) worries that “shouting louder (‘voice’) is likely to be futile … for those mired in poverty.” Political activism, he writes, “does perform as advertised, but only sometimes, and even then usually for those who already enjoy many advantages.” On this point, at least, he and I are in agreement. But then what, if anything, is to be done? If economic inequality is a threat to democracy, perhaps the only feasible political solution is less economic inequality. However, that seems likely to require precisely the sort of “government fix” that Weissberg disdains.

A new study of income redistribution in OECD countries (Kenworthy and Pontusson 2005) has bolstered and extended the Task Force's observation that the U.S. has done less than other rich democracies to mitigate the effects of increasingly unequal market incomes in recent decades. Ten of the 11 countries for which data are available (all except the Netherlands) experienced increases in income inequality among working-age households in the 1980s and 1990s. Of these 10, nine engaged in more aggressive redistribution at the end of this period than at the beginning; in general, the increase in redistribution was greatest in countries where the increase in inequality of market incomes was greatest. The glaring exception was the U.S., which experienced a larger-than-average increase in income inequality but no corresponding increase in redistribution.

In attempting to account for American exceptionalism in this regard, the Task Force pointed to a variety of structural and cultural factors ranging from the fragmentation of American policy-making institutions to the ideological appeal of economic opportunity and limited government. One thing it did not do, however, was to focus attention on what seems to me to be a crucial factor shaping the political economy of American inequality—the contrasting policy choices of Democratic and Republican elected officials. In this respect, at least, I believe that the Task Force members took their “non-partisan” mandate a good deal more seriously than Weissberg gives them credit for.

What might a rigorous partisan examination of the American political economy show? As it happens, Hibbs (1987; Hibbs and Dennis 1988) provided a good answer to that question almost two decades ago. Figure 2 updates and extends that work, summarizing Census Bureau data on patterns of real income growth over the past half-century for families at various points in the income distribution. The figure shows separate average annual growth rates for families at each income level during Democratic and Republican presidential administrations.²

Under Democratic presidents income growth has been fairly egalitarian (at least in percentage terms), with average real growth rates ranging from 2.6% for families at the 20th percentile to 2.1% for families at the 95th percentile. Under Republican presidents the pattern has been dramatically different. Affluent families have fared equally well during Republican and Democratic administrations; but middle-class and poor families have fared markedly worse, on average, under Republican presidents.³

For working poor people (at the 20th percentile of the income distribution), Republicans have presided over average real income growth of 0.6%—less than one-fourth of the corresponding average growth rate under Democratic presidents.

Are partisan differences in the economic fortunes of American families really this stark? The arithmetic calculations from the Census Bureau data are straightforward. Their political significance can only be gainsaid by supposing that the apparent pattern is the result of a massive historical coincidence. Elsewhere, I have provided extensive checks on the robustness of the partisan disparity evident in Figure 2, including comparisons based on alternative economic units, time periods, and income definitions, statistical controls for historical trends, non-parametric tests, and the like (Bartels 2004). It seems hard to escape the conclusion that, over the past half-century, Republican presidents have been consistently bad for the economic health of middle-class and poor people.

Weissberg (35) is put off by the Task Force's nostalgia for the New Deal and the Great Society. He rejects the notion that democracy “requires massive fresh Washington expenditures for health care, education, subsidies for working parents, and the like.” He chides radical academics for ignoring the fact that “this costly ‘relief’ has historically proven ineffective while not always benefiting intended recipients.” The evidence in Figure 2 suggests precisely the opposite. Over a period of more than half a century the policies of Democratic administrations have been significantly more effective than the policies of Republican administrations in generating economic growth, and in distributing the benefits of economic growth broadly to people across the economic spectrum.

If there is a structural lesson to be drawn from Figure 2, it is a very simple one: politics matters. The rising tide of economic inequality in contemporary America has myriad economic and social causes, including technological change, demographic trends, and global competition. Some of these we can influence, others we can only adapt to. But however high the tide may rise, we are not condemned to wait behind our levees for disaster to engulf us. Even in their current soggy state, the institutions of polyarchy provide us with consequential choices. We can reinforce the levees; we can divert some of the fastest-running waters; and we can insist that the most vulnerable among us not be abandoned when the affluent flee to higher ground.