Andrea Louise Campbell’s remarks about my book Reforms at Risk are satisfying to receive not only because of their generosity but because of their source. In the collective effort to bring attention to history and temporal forces within the orbit of mainstream political science, few scholars have made more important and creative empirical contributions over the past decade than Campbell. I especially appreciate her keen probing of the reach and limitations of my argument. Let me offer the following responses to the very stimulating questions that she raises.

Campbell’s first question is whether my analysis specifies necessary or sufficient conditions for reform sustainability. It does not. My analysis suggests that several factors shape how a reform evolves, including its relative timing, the policy legacies that the reform seeks to alter or displace, and the strategy and design of the reform itself. The burden of my argument, however, is that the two most important factors in shaping outcomes are institutions and interest-group feedbacks. I deliberately steered away from the language of “necessary and sufficient” conditions because my analysis presumes probabilistic, not deterministic, causation. The configuration of institutions and interests influences the susceptibility of a reform to downstream erosion, entrenchment, or reversal. It loads the dice. But it does not determine outcomes because, as I show, the fate of reforms is never completely settled, even if some pathways are largely foreclosed.

What I think my work does reveal is that different factors shape reform trajectories in different ways. Here Campbell’s reading of my work is particularly shrewd and insightful. Institutions matter in part because they affect the credibility of the political commitments underlying reforms. They establish the structural frameworks in which political contestation over reform outcomes takes place, channeling conflicts in some directions rather than others. It may be theoretically possible for other mechanisms, such as a deep consensus on the legitimate role of government, to serve this anchoring function. In practice, however, institutions are essential to governance and conflict management in our pluralistic democratic polity. Hence institutional configurations require close attention. Interest-group feedbacks matter because they directly affect the level and nature of support for the extension of a reform’s line of policy development. Rather than asking whether institutions or interests are primary, I find it more productive to examine their interaction effects. My analysis thus explores how institutions and interests may both cause one another and come into tension and conflict during the critical postenactment phase.

Campbell’s second question is whether sustainable reform is even possible in policy arenas that are central to legislators’ reelection needs. As I argue, the congressional penchant to “bring home the bacon,” “serve the organized,” and meddle with even relatively well-functioning markets can pose a major obstacle to the sustainability of general-interest reforms. Some arenas offer such rich opportunities for legislative credit claiming that any effort to improve public policy constitutes an explicit threat to the powers and privileges of Congress itself. Yet history suggests that some policy arenas that were once mired in reelection-driven, particularistic favor provision have been durably reformed. An example—which I did not give sufficient attention to in the book—is liberal trade policy. For much of the late nineteenth century, Congress served protectionist interests by maintaining high barriers to free trade. While Democratic Congresses sometimes voted to lower tariffs on particular goods, the interests of consumers and exporters were systematically underrepresented. But Congress was able to reconfigure trade policy through the enactment of the Reciprocal Trade Agreement Act of 1934. The RTAA employed several mechanisms to durably collapse logrolling coalitions supportive of high tariffs. These included the shifting of trade decisions from the tax arena (in which Congress considered each tariff separately) to treaty negotiation (in which the president negotiated comprehensive trade packages with other nations). The enactment of the RTAA was motivated by the desire of the Democratic Party leadership to “build support for free trade within the party and to insulate trade policy from a future Republican Congress.”Footnote ¹ This suggests to me that the partisan incentive to lock in future electoral gains can sometimes propel Congress to reform itself. That has not happened with respect to congressional control over the income tax code—and I suspect the pathologies of the federal tax system will need to get a lot worse before it does.

Campbell’s third excellent question concerns the relative importance of contingency versus policy design. Do reforms last because they are well timed or because they are well built? One way to address this question is to distinguish between two kinds of design failures. The first is the failure to design against predictable postpassage threats to sustainability. It was completely foreseeable that if commodity prices fell, the market transition scheme contained in the Freedom to Farm Act would not stick. Journalists wrote about this risk and elected officials and farmers’ groups openly discussed it. Yet Freedom to Farm contained no institutional mechanism to protect its subsidy retrenchment scheme from this uncertain but high-probability event. The reform’s failure was thus the product of bad relative timing only in the sense that a contingent but foreseeable event (a shift in agricultural market conditions) exposed the underlying weakness of the policy design itself. In contrast, policy designers deserve less blame (or credit) when unforeseeable events alter political incentives in ways that profoundly shift the context in which a reform is implemented. The success of welfare reform, for example, must be attributed at least in part to its fortuitous timing. During the debate over enactment, there was ideological conflict over whether the reform was “too harsh” to persist. What no expert or politician predicted was that the U.S. economy would grow so rapidly that the unemployment level would fall below four percent within the decade. This permitted many low-skilled, former welfare mothers to earn enough to make up for their lost benefits, enabling the reform to “take” before the next big recession hit.

Campbell’s fourth question concerns the period specificity of my analysis. All of the reforms I analyzed were adopted between the 1970s and 1990s. This was an era that featured growing concern about the costs of government regulation, budget deficits, a fraying safety net, and poor macroeconomic performance. I restricted my attention to this recent historical period in part for methodological reasons. I wanted to hold certain background political conditions, such as the strength of the party system and the role of think tanks and policy expertise, relatively constant, thereby facilitating analysis of policy-specific factors. I also had a substantive reason to restrict my attention to the modern period. I define reform to mean not simply major policy innovation (which has occurred through American history) but rather as a nonincremental shift in a preexisting line of policy development intended to promote general benefits. I thus distinguish the reform process, in which the federal government already has a big presence in a policy arena and is seeking to make its role more effective and efficient, from what Lawrence D. Brown calls “breakthrough” policymaking, in which the government explicitly penetrates a policy field or sector for the first time.Footnote ² Breakthrough policymaking still takes place, but it has become much less common as the American state has expanded.

A common reform thrust in recent American politics, grounded in the efficiency-oriented ideas of policy experts and think tanks associated with both parties, has been to use marketlike mechanisms and incentives (defined broadly to include a “user fee” approach to social provision, as in the Medicare Catastrophic Coverage Act). This neoliberal reform thrust shows up in virtually all my cases. But neoliberalism is certainly not the only available reform strategy in the U.S. setting, and had I canvassed more widely across American history, other approaches would have turned up, including (to list one approach that I am exploring in my current research) efforts to change the state’s role vis-à-vis the professions like medicine. Even in the period that I focused on, however, some major reforms emerged out of ideational and cultural paradigms at odds with neoliberalism. One example is the “new paternalism” in social policy, in which the government seeks to tighten supervision of the lives of poor or otherwise marginal citizens.Footnote ³ Certain policies toward drug users, the homeless, and fathers who have abandoned their children can be understood from this perspective, and the work-oriented aspects of welfare reform may fit this category as well. I have not explored these cases in depth, but I would expect my explanatory framework to provide significant insights into their trajectories.

Another extension of my explanatory framework would be to examine how breakthrough policies themselves evolve over time. An interesting case is Social Security. In her wonderful book How Policies Make Citizens, Campbell provides a compelling analysis of the powerful interest-group feedbacks that Social Security generates today, making senior citizens the most politically active age group.Footnote ⁴ But as Julian Zelizer and I point out in a recent conference paper, Social Security was not self-reinforcing at birth.Footnote ⁵ Indeed, Social Security’s original design was weak in key respects, including its failure to provide upfront benefits. These design weaknesses blunted the program’s capacity to build a constituency. Social insurance was not consolidated as a breakthrough policy until the Eisenhower administration (after the 1950 Social Security Amendments had substantially expanded benefit levels) signaled its endorsement.Footnote ⁶

Campbell’s final point is that I could have been less modest about the book’s theoretical contribution to the literature on policy feedback effects. I appreciate her recognition that my book helps open the black box around feedback effects and that it provides an analysis of the mechanisms by which policies shape subsequent politics. She is certainly correct that I could have engaged more directly the comparative politics literature, from which I have learned a great deal. That said, one of my goals in writing the book was to encourage political scientists who study the U.S. Congress, lawmaking, and representation to pay more attention to history and political development. As Paul Pierson has observed, getting mainstream, empirically-oriented political scientists (especially Americanists) to take policy development seriously is an uphill battle for a host of reasons.Footnote ⁷ I would be pleased if my book contributes to the advancement of that important project.