In this brief reflection I mostly set aside the empirical claims of the book, gratified that the commentators broadly agree that (with important exceptions) they have held up well. ^{Footnote 1} Instead, I want to focus on three core ideas from Dismantling that I believe retain importance for understanding the political economies of advanced democracies. In each case, however, my thinking has evolved over the two decades since that book appeared, in large part because of the work of other scholars.

The first idea, discussed at some length in most of the commentaries, concerned the politics of policy feedback. In Dismantling this was mostly an argument about institutionalization. In Tom Stoppard’s Arcadia (Stoppard Reference Stoppard1993, 5), the precocious Thomasina asks her flummoxed tutor to explain why you can stir jam into pudding, but you cannot stir it out. Politics does not stop when you adopt a law, but the law can change the politics. I learned this from reading the work of Theda Skocpol and Margaret Weir (Weir and Skocpol Reference Weir, Skocpol, Evans, Rueschemeyer and Skocpol.1985). It was the development of this theme that was at the heart of my claim that once welfare states had had decades to mature, the politics of social policy would be “new” in important respects.

I will say little about this here because the main issues are deftly discussed in the commentaries of Campbell, Gingrich, and Mettler. As they stress, excellent, nuanced work on policy feedback has proliferated. This research program has firmly established that a sophisticated and realistic institutionalism should consider the activist state, alongside formal political institutions, as a fundamental system of rules shaping political development. Mettler calls this the policyscape—a nice term. The emergence of the policyscape is a huge change in modern polities, but one that has been mostly ignored by overly formalistic versions of institutionalism. In a recent essay (Pierson Reference Pierson, Jenkins and Milkis.2014), for example, I argued that the expansion of the activist state in the United States has played a crucial and underappreciated causal role in generating and sustaining political polarization.

A second idea that I considered central to Dismantling was that of visibility. Scholars have long known that social policy structures vary dramatically in how they distribute costs and benefits. What was less appreciated was the extent of variation in how visible those costs and benefits were to different political actors, especially voters. This insight—which was of course at the heart of E. E. Schattschneider’s approach to politics—had been developed in the work of R. Kent Weaver (Weaver Reference Weaver1988) and R. Douglas Arnold (Arnold Reference Arnold1990). I had the privilege of working with Weaver at the Brookings Institution while I was completing my dissertation, and his thinking had an enormous impact on my work. It was not just that the visibility of programs influenced political mobilization. More important, visibility could be manipulated: policy makers could often design (or redesign) policies so that voters would see the things they wanted them to see, and not see what they preferred to leave hidden—such as particular changes that might cut social benefits gradually over time.

I always thought this emphasis on the manipulation of visibility was one of the most important arguments developed in Dismantling, but initially it did not gain much traction. Over time, however, interest in this dimension of politics has grown, from Christopher Howard’s The Hidden Welfare State (Howard Reference Howard1999) to Jacob Hacker’s The Divided Welfare State (Hacker Reference Hacker2002) to, more recently, Mettler’s The Submerged State (Mettler Reference Mettler2011). In my own extended collaboration with Jacob Hacker, ideas about visibility have been central. We maintain that there is often considerable potential for organized actors to use policy designs to increase their freedom of movement at the expense of the disorganized. Our own work has been part of a growing recognition among many Americanists of the limited capacities of voters (Bartels Reference Bartels2008; Lenz Reference Lenz2012) and the strategic capacities of elites (Bawn et al. Reference Bawn, Cohen, Karol, Masket, Noel and Zaller2012; Van Houweling Reference Van Houweling2012). In a recent essay, we suggested that students of American politics need to move from a “Downsian” paradigm oriented around voters and elections toward a more “Schattschneiderian” framework that directs attention to contestation (stretching well beyond the confines of elections) among the well-organized over control of governance (Hacker and Pierson Reference Hacker and Pierson2014).

Indeed, this focus on sustained contestation in pursuit of durable change was for me the third major idea explored in Dismantling. Reagan and Thatcher sought to use their electoral victories to achieve policy victories. In turn, they hoped to use those policy victories to create new political and policy opportunities down the road. Peter Hall deftly explores this point in his comment. Again, I believe that this remains an important argument for political scientists to pursue, potentially getting us closer to the heart of where political power lies in democratic polities and how it is exercised. At the same time, here I would probably be a little more critical of Dismantling than the generous contributors to this symposium have been.

Hall notes that governing for the long-term often means taking care of your own coalition and trying to cause trouble for competing ones. Exploring the means of accomplishing this was central to Dismantling, but in ways that I would now regard as insufficient for three interconnected reasons. First, its depiction of political coalitions is underdeveloped. Dismantling focuses on “retrenchment advocates” who were underspecified, and in practice remained limited to the cluster of top policy makers around the two conservative political leaders. What received too limited attention were the organized actors (“intense policy demanders” in the terminology of Bawn et al. Reference Bawn, Cohen, Karol, Masket, Noel and Zaller2012) who provide vital resources over time and recruit, monitor, and discipline politicians.

A second limitation, stressed by Hacker (Reference Hacker2004), is a failure to fully appreciate that retrenchment advocates often did not need to change the law; they could sometimes “win” simply by blocking reform. Hacker’s argument about “policy drift” has had a profound impact on my thinking about how to study political economy. ^{Footnote 2} The social world is always changing, and economic markets in particular often shift rapidly. Policies need to “keep up,” and often opponents can gain much of what they want if they can simply succeed in blocking efforts to combat inevitable policy drift. To be sure, Dismantling took note of this possibility (discussing strategies of “decrementalism”), but Hacker’s account was not just more analytically precise; he persuasively argued that this dimension of policy contestation was far more pervasive and important than I had recognized, especially in a veto-ridden institutional setting like the United States. As Campbell’s commentary suggests, even when the structure of a program may be durable, its social impact may diminish. ^{Footnote 3}

The third, linked limitation is that Dismantling focused only on a subset of issues in the broader political struggle between competing coalitions. “The welfare state” includes many very important programs, but it is only one part of the much broader set of policies that define a political economy and help determine the distribution of economic and political resources among contending interests. It was this broader set of distributional issues that was central to Reagan and Thatcher’s political project. True, the evidence now seems clear that welfare state programs have been quite durable (even as a profound and lasting economic crisis, especially in Europe, renders their future status uncertain). Yet in many places (and most certainly in the United States) that solidity now looks something like a Maginot Line. Rather than directly assaulting the welfare state, those seeking to remake the American political economy have mostly outflanked it, relying heavily on tactics of gridlock-inducing policy drift to produce major changes in taxes, industrial relations, corporate governance, and financial regulation that have been highly beneficial to the most affluent (Hacker and Pierson Reference Hacker and Pierson2010). As Hall argues, sustained efforts to shift mass and elite political discourse are an important part of the story—one that Hacker and I will address in a more sustained way in work in progress (Hacker and Pierson forthcoming).

Take these three points together and there is an irony. Dismantling argued that the shift in relative power resources from organized labor to business was less damaging to the welfare state than previous research would have suggested. The commentators (most systematically Stephens) argue that this claim of Dismantling holds. What I am suggesting is that more recent scholarship, including my own, points to the need for important modifications to this account. When we widen our gaze to consider a broader set of policies, and to recognize that statutory stability in a context of profound socioeconomic change is not really the same as “resilience,” our conclusions look different. To say that the assault on the welfare state largely failed is not to say that neoliberalism did. The irony, fully acknowledged in my recent collaborative work, is that the broader success of neoliberalism has depended heavily on precisely the shifts in the balance of organized power (with business increasingly effectively mobilized, and labor unions in steady decline) that were discounted in Dismantling.

At its best, scholarship is a progression in which each individual contributes something to our growing collective understanding of important phenomena. Although it sometimes feels like scholarship is going in circles or drifting sideways, I have no doubt that we know far more about the politics of social policy than we did a quarter-century ago. In this developing research agenda Dismantling provided neither the first word nor the last word. ^{Footnote 4} I am very gratified that such an impressive group of scholars feels that it is worth discussing two decades after publication.