Democracy requires that all citizens—rich and poor alike—have influence over the policies their government adopts.
|Reply: Martin Gilens|
Democracy requires that all citizens—rich and poor alike—have influence over the policies their government adopts. Of course, it would be unreasonable to expect everyone to have equal sway. Citizens differ not only in economic resources but also in time, knowledge, and interest in social and political affairs. Still, when influence becomes too skewed toward the affluent, when political power becomes too concentrated in the hands of a few, democracy itself is threatened.
If you accept those basic premises, then you have good reason to be worried about American democracy. That’s the conclusion I’ve come to after studying the relationship between public policy and public preferences as revealed in responses to thousands of questions from national surveys conducted between 1964 and 2006. If you judge how much say people have—their influence over policy—by the match between their policy preferences and subsequent policy outcomes, then American citizens are vastly unequal in their influence over policymaking, and that inequality is growing. In most circumstances, affluent Americans exert substantial influence over the policies adopted by the federal government, and less well off Americans exert virtually none. Even when Democrats control Congress and the White House, the less well off are no more influential.
The one bright spot in this unhappy tale of unequal influence is that political competition increases the responsiveness of policymakers to the views of the public and generates policies that more equally reflect the preferences of all Americans. When elections are near and when control of the government is divided or uncertain, parties broaden their appeal, and influence becomes more equal. So the core elements of democratic government—electoral competition and partisan rivalry—force policymakers to take public preferences more fully into account.
But competition is rarely strong enough to create anything remotely resembling equal representation, and growing income inequality is pushing that ideal even further off.
• • •
In a world of polarized politics, it is easy to forget that rich and poor Americans often agree on public policy. For example, most foreign military engagements—whether popular, such as war in Afghanistan, or unpopular, such as the Reagan administration’s Central America policy—have had similar levels of support across the income spectrum. Different income groups also express comparable levels of support for education spending, the War on Drugs, family-leave laws, childcare and job training for welfare recipients, and the Medicare drug benefit. And income has little effect on opposition to a federal sales tax and broad-based income tax increases.
In fact, about half of the proposed policy changes in the study show differences of less than ten percentage points between the preferences of poor, middle-class, and affluent Americans (I use the 10th, 50th, and 90th income percentiles to represent these three groups).
If Americans at different income levels agree on a policy, they are equally likely to get what they want. But what about the other half of the time? What happens when preferences across income levels diverge?
When preferences diverge, the views of the affluent make a big difference, while support among the middle class and the poor has almost no relationship to policy outcomes. Policies favored by 20 percent of affluent Americans, for example, have about a one-in-five chance of being adopted, while policies favored by 80 percent of affluent Americans are adopted about half the time. In contrast, the support or opposition of the poor or the middle class has no impact on a policy’s prospects of being adopted.
These patterns play out across numerous policy issues. American trade policy, for example, has become far less protectionist since the 1970s, in line with the positions of the affluent but in opposition to those of the poor. Similarly, income taxes have become less progressive over the past decades and corporate regulations have been loosened in a wide range of industries.
Nor do cross-class alliances work to dent the influence of the well off. When middle-class preferences align with those of the poor, responsiveness to the affluent remains strong while responsiveness to the poor and middle class is still absent. Low- and middle-income Americans have been united, for example, in opposing free trade agreements such as NAFTA and the General Agreement on Tariffs and Trade and in supporting abortion restrictions such as requiring the prior consent of the biological father. But the affluent tend to favor free trade and to reject these kinds of abortion restrictions. And the affluent few have gotten what they want.
What difference would it make if policy more equally reflected the preferences of all Americans? How would it change?
Greater representational equality would have a substantial effect on several important economic policies. We would have a higher minimum wage, more generous unemployment benefits, stricter corporate regulation (on the oil and gas industries in particular), and a more progressive tax regime. Some of these policies are favored by a majority of Americans at the 90th income percentile as well, but not with sufficient enthusiasm to overcome opposition from business and other interests. We would also see a more protectionist trade strategy and less foreign aid.
The views of the affluent make a big difference, while support among the middle class and the poor has virtually no relationship to policy outcomes.
We’d see little significant change in Social Security and Medicare, which enjoy strong support across the income spectrum. This doesn’t mean that these core welfare programs are beyond class-specific political influence, however. The small changes to which they’ve been subject—such as the increase in the Social Security retirement age and efforts to encourage Medicare beneficiaries to join HMOs—have been more consistent with the preferences of the well off.
The disproportionate influence of the affluent does not always move policy in a conservative direction. On moral and religious issues, the well off tend to be more liberal than the poor. More equal representation would consequently lead to greater restrictions on abortion, such as banning RU-486. There would also be tighter limits on stem cell research and more support for school prayer.
• • •
Democrats have traditionally been viewed as the party of the working class and Republicans the well-to-do. But my findings suggest that both parties are inclined to ignore the public. Both seek to control government, and strong partisan control by either leads to policymaking with little regard for the preferences of the governed.
During the Lyndon B. Johnson administration, for example, Democrats held strong majorities in the House and Senate, yet policymaking bore no relationship to the preferences of the poor and middle class, or, for that matter, any income group.
We remember the Johnson administration for its landmark domestic legislation, including the 1964 and 1965 civil rights bills, the War on Poverty, the establishment of Medicare and Medicaid, immigration reform, and expanded federal aid to education. Some of these programs were indeed quite popular. Medicare and federal aid to education were favored strongly by low- and middle-income Americans and only modestly less by the affluent. Public support for civil rights legislation grew during the early 1960s, and most of the civil rights bills had solid public support across the income spectrum by the time they were passed.
But the majority of Americans were opposed to many of the other domestic programs of the Johnson years. The Great Society and the War on Poverty were not responses to an upwelling of public concern for the disadvantaged or a desire to expand the role of government in addressing social needs. If anything, public support for government activism was declining. Strong majorities at all income levels opposed increased spending on aid to cities, low-income housing, and welfare or relief payments. Even more unpopular was the loosening of immigration laws in 1965, opposed by 90 percent of poor Americans and about three-quarters of the affluent and middle class. Johnson’s escalation of the war in Vietnam grew less popular over time, and strong majorities at all levels opposed the Vietnam War income tax surcharge adopted in 1968.
This is not to say that the reputations of the parties are totally out of step with reality. Downwardly redistributive economic policies, such as raises in the minimum wage, are more common when Democrats are in power, and upwardly redistributive policies, such as reductions in the estate tax, are more common under Republicans.
But this pattern—important as it is—does not reflect a broader representation of the middle class or the poor by the Democratic Party. As we’ve seen, lower-income Americans align with the traditional policy orientation of the Democrats only on economic and social welfare issues and prefer Republicans’ positions on abortion, school prayer, and gay rights. And over the past few decades, the Democratic Party has shifted strongly in a free-market, anti-regulation, free-trade direction—a shift more consistent with the preferences of the affluent than of the less well off.
Finally, public preferences on redistributive economic policies don’t break down as cleanly as one might expect. Proposals to raise the minimum wage, expand federal college assistance, reduce taxes for low-income Americans, and protect or expand government support for health care receive strong support from affluent voters. In addition some upwardly redistributive policies—such as imposing work requirements and time limits on welfare recipients, eliminating the inheritance tax, and cutting capital gains tax rates—receive strong support even from the least well off.
These findings may be disappointing to those who look to the Democratic Party as the ally of the disadvantaged. In some respects Democrats have in fact served this function in the social welfare domain. But in other domains, policies adopted under Democratic control are no more consistent with the preferences of the less well off than are those adopted during periods dominated by the Republican Party.
• • •
Whereas the Johnson years featured unified government and a certain disregard for the public will, the opposite conditions prevailed during the early years of the George W. Bush administration. Bush and a divided Congress hewed closely to voters’ preferences.
At least at first, the Afghanistan and Iraq wars and Bush anti-terrorism policies had widespread public backing. The administration also enjoyed support from Americans at all income levels for the Medicare drug benefit, No Child Left Behind (a long-standing Democratic agenda item on which Bush partnered with Senator Ted Kennedy), and the faith-based social services initiative. New federal regulations on funding stem cell research matched public sentiments. The income tax cuts and estate tax repeal adopted in 2001 and 2003 were supported by majorities of Americans at all income levels even though they clearly provided the greatest benefits to those at the top.
Democrats are seen as the party of the working class, Republicans the well-to-do. But both are inclined to ignore the public.
The sensitivity to public preferences during the early years of the Bush administration speaks to the importance of partisan competition. Bush had lost the popular vote and campaigned as a centrist. Meanwhile the 2001–’02 Congress was more closely divided than any in the previous half-century. The Republicans’ tenuous hold on power, their need for Democratic cooperation in the House and Senate, and the uncertainty of control after the next election encouraged both parties to court the public by pursuing popular, centrist policies.
The patterns of policymaking during the Johnson and George W. Bush administrations hold up through the decades. Strong majority-party dominance of Congress (especially of the Senate) is associated with lower responsiveness to the preferences of the public, while an even division of seats leads to a stronger association between public preferences and policy outcomes. Not surprisingly, influence remains greatest for the well off—the affluent always have greater influence than the middle class or the poor—but partisan rivalry still generates a considerable boost for the middle class and the poor.
Other aspects of political competition similarly serve to ameliorate inequality in responsiveness. For example, representational inequality is lowest during presidential election years. Indeed, these are the only years of the quadrennial federal election cycle in which there is any evidence whatsoever of receptiveness to the preferences of the poor.
Political gridlock also—somewhat perversely—reduces representational inequality. This may seem counterintuitive; after all, gridlock undermines the government’s ability to respond to public concerns. But gridlock is more effective at undermining unpopular policies than popular ones. The minority party has little incentive to cooperate with the majority in adopting unpopular changes, but standing in the way of popular ones exacts a political cost. Thus gridlock operates like a filter, thwarting the majority party’s efforts to achieve its non-centrist policy goals and strengthening the association between public preferences and policy outcomes.
The importance of political conditions is reinforced by comparing responsiveness during the first years of Bush’s first term in office and the first years of his second term. Congress was closely divided as Bush came into office, but after the 2004 election the Republicans gained unified control of the federal government for the first time in more than half a century. The consequence: the public lost the government’s ear. Policy outcomes in 2005 and 2006 bore no discernible relationship to public preferences at any income level. Neither affluent nor other Americans had any more influence over policymaking during this (brief) period of strong Republican control than they did during the height of Democratic Party control in the mid-1960s.
• • •
Across multiple presidential administrations and a wide range of political conditions, two patterns remain constant. First, the poor never have as much influence as the middle class, and the middle class never has as much influence as the affluent. Second, over the last four decades, responsiveness to the affluent has steadily increased while responsiveness to the middle class and the poor has depended entirely on the existence of the congenial circumstances just described.
Exactly what combination of factors accounts for the rising influence of the rich is hard to say. The growth of economic inequality, the mounting cost of political campaigns, the changing nature of the media environment, and politicians’ increased use of surveys may all have played a part by intensifying candidates’ need for money, concentrating the supply of that money even further in the hands of the affluent, and enhancing the role of policy considerations in elections.
The power of organized interest groups is often brought up as an explanation for the disproportionate influence of the well off. But while interest groups do strongly shape policy outcomes, they don’t account for the greater influence of high-income Americans. In popular discourse, interest groups typically bring to mind the Chamber of Commerce, the NRA, or the American Bankers Association. But the AARP, the AFL-CIO, and the National Governors Association are all powerful interest groups as well, and they and others like them are more likely to share the policy preferences of the poor and the middle class than of the affluent. In addition, industry lobbying organizations tend to side with the affluent in pushing for lower taxes and less regulation, but they align with the poor in favoring increased government spending on health care, education, mass transit, highway construction, and federal revenue sharing with the states.
Another seemingly plausible explanation for representational inequality is that policymakers are themselves affluent, and their personal interests and preferences shape their legislative efforts. Yet concern over the growing wealth of members of Congress is probably misplaced. Liberals and conservatives are equally likely to be found among Congress’s most and least wealthy members, and the substantial differences in economic status among members of Congress are not related to their preferences on economic policy issues or to their overall voting patterns.
Political competition increases the chances that policymakers will respond to the public’s views—but not by much.
The most straightforward explanation for representational inequality is that high-income Americans are more likely than are less well off citizens to vote, volunteer in campaigns, and make large political donations. All these political activities increase with income, but they do so in different ways. Low-income Americans are distinctive in their lower rates of voting and volunteering, while the differences between middle-income and affluent Americans are modest. But when it comes to campaign donations, high-income American stand out. Political donations, then, but not voting or volunteering, replicate the pattern of representational inequality described above: the affluent are distinctively influential while the middle class typically has no more sway than the poor.
Money is not the only valued commodity in politics. Groups that can mobilize large numbers of volunteers, such as labor and religious organizations, may exert influence that competes with that of the affluent, at least on particular issues at particular times. Nevertheless, electoral campaigns require money, and more and more of it over time. While the evidence is circumstantial, the overlapping patterns of policy influence and political contributions offer at least one highly plausible explanation for the representational inequality that my study reveals.
We might wonder, however, whether the 30 million Americans in the top income decile are really shaping political outcomes. Perhaps their apparent influence arises from agreement with a much smaller and more affluent circle that wields true sway over government policy. And even if the “merely affluent” do exert influence over political outcomes, the truly rich—the 1 percent—may dominate on those issues that most strongly affect their collective interests.
At the moment, it’s hard to tease apart the top 10 percent and the 1 percent. We know little about the policy preferences of very wealthy Americans both because there are relatively few of them and because polling organizations typically do a poor job of identifying them even when they are included in survey samples. Political scientists Benjamin Page and Larry Bartels (see his response on page 20) are currently trying to collect survey data from the top 1 percent of American households. If they succeed, we may be able to assess the extent to which the influence of the truly rich accounts for the responsiveness to Americans at the 90th income percentile.
• • •
Perfect, or perfectly equal, responsiveness to the public may not be ideal. There are good reasons to want government policy to deviate at times from the preferences of the majority: minority rights are important too, and majorities are sometimes shortsighted or misguided in ways that policymakers must try to recognize and resist.
But gross levels of inequality in influence are incompatible with the notions of political fairness that Americans embrace, and the increasing concentration of wealth at the very top is making that inequality in influence worse. While other nations have adopted policies to ameliorate the increases in economic inequality brought, at least in part, by changes in the global economy, America has not. Tax cuts for the wealthy, financial deregulation, attacks on unions, and cutbacks in funding for public education exacerbate economic inequalities. In the face of these changes and the continuing economic challenges facing less well off Americans, representational inequalities are all the more significant.
But political circumstances do matter, so there is room for optimism about reforms to reduce the impact of money in politics, shift the source of political money away from the most affluent Americans, and enhance partisan competition. All of these changes might help to bring government policy more in line with the preferences of all Americans. States have taken the lead on campaign finance reform. We don’t know whether these reforms actually reduce representational inequality, but my national-level results suggest that they very well may. Federal campaign finance regulations have been moving backward lately, but the recent Citizens United case was decided by a five-to-four vote. Perhaps a future court will be friendlier to efforts to control the flow of money in political campaigns.
In addition to campaign finance, competition-enhancing reforms such as nonpartisan districting and nonpartisan get-out-the-vote drives have been shown to produce more competitive elections and might thereby induce policymakers to attend more closely to the preferences of the public.
Finally, advocates can focus on those policies that are supported by the poor and the affluent alike. Even at the top of the income distribution, there is strong support for increases in the minimum wage, spending for education, means-tested job training and childcare, Social Security, and Medicare.
The obstacles to enhancing representational equality are considerable. Political reform is always difficult, and here especially because economic resources and the political influence that accompanies them continue to move toward the already advantaged. But the importance of elections and partisan competition in shaping responsiveness means that our political destiny is not predetermined.
American democracy is imperfect—so much so that its claim to that title is suspect. But our democracy was even more imperfect in earlier eras, lending some hope that, through their hard work, Americans’ attempts to build a more perfect union might not be in vain.