The 2017 GOP drive to “repeal and replace” the Affordable Care Act (ACA) arguably constituted the most ambitious effort to dismantle a social program in American history. Certainly it was the most ambitious to come so close to enactment, falling just three votes short in the Senate. According to an extensive body of scholarship, this near miss should have been nearly impossible. The political fallout associated with dismantling social programs, as well as their entrenchment in social and economic life, should deter electorally sensitive politicians from undertaking retrenchment of the scale Republicans sought. To unravel this mystery, we explain how the radicalization of the GOP is rooted in distinct electoral and organizational realities that have simultaneously increased the party's desire to move right and its capacity to do so—even when the policies that Republicans seek to enact are distinctly unpopular. We also explain why the ACA was underentrenched and why Republicans believed they could use legislative procedures and policy design to minimize the risks they faced. Finally, we highlight some new barriers to retrenchment in this age of “asymmetric polarization.” The hurdles facing retrenchment—in particular, the sway of national public opinion—have not simply diminished; they have changed, in ways that should be incorporated into our theories.
On July 28, 2017, Senator John McCain approached GOP Majority Leader Mitch McConnell in the well of the Senate. The moment was freighted: The Arizona Republican who had lost to Barack Obama in the 2008 presidential race had recently announced he suffered from a deadly form of brain cancer. With Republicans enjoying a slim majority, McCain had also become the pivotal vote in the party's effort to repeal the 2010 Affordable Care Act (ACA), President Obama's signature legislative achievement. Under the rules of the budget process that McConnell was using to pass the repeal legislation, Republicans needed just fifty votes to succeed. If McCain voted no, the bill would die.
McCain turned down his thumb. Gasps and scattered applause could be heard. McConnell's head fell. The legislative drive that had started with the inauguration of President Donald J. Trump in January and continued through the House GOP majority's passage of a repeal bill in May had ended with Senate drama worthy of a Hollywood script. Despite unified Republican control of Washington, the legislative campaign to repeal “Obamacare” had fallen short. Although McConnell tried again with another bill in September, the last few GOP votes proved just as elusive as they had in July. While President Trump vowed to use his executive authority to undermine the ACA, congressional Republicans were forced to move on.
By no means is the fight over. As the legislative clock ticked down for 2017, Republicans managed to include in their big tax bill a partial repeal of the ACA's unpopular requirement that individuals show proof of health insurance or pay a fine (the so-called individual mandate). Nonetheless, the collapse of their health care bills during what was supposed to be a triumphant first year of unified Republican control was a stunning defeat of a top legislative priority that GOP candidates had campaigned on for four election cycles.
But was it surprising? A generation of political science research has shown that dismantling large-scale social programs is hard. The status quo bias of political institutions, the well-known tendency for individuals and organized groups to fight fiercely against losing what they already have, and the increasingly understood processes of entrenchment that occur as programs become integrated into economic and social life all mean that “frontal assaults on the welfare state carry tremendous political risks” (Pierson 1996: 177–78).1
Yet Republicans came remarkably close with their “frontal assault” on the ACA. Though the exact legislative ammunition differed from bill to bill, the goals of GOP leaders were consistently ambitious: up to $1 trillion in cuts in Medicaid over a decade (with the size of the cuts growing even larger after that) and evisceration of many of the subsidies, regulations, and revenue sources in the ACA. According to the nonpartisan Congressional Budget Office (CBO), these changes were likely to reverse all or most of the coverage gains under the 2010 law, which had reduced the number of uninsured Americans by roughly 20 million (CBO 2017a, 2017b, 2017c, 2017d). Nor did “electoral risks” appear to be a major consideration—except Republicans' fear that primary voters would punish them if they failed to pass their bill.
To be sure, the GOP effort ultimately faltered, and perhaps it has failed for good. But it is nonetheless remarkable that it made it as far as it did. With apologies to Sir Arthur Conan Doyle, this is a case of “the dog that almost barked.” In light of what we know about retrenchment, and given the extremely close divide between the parties, the mystery is not why Republicans came up short but why they barreled forward despite the risks and came within a few votes of victory despite the obstacles.
To preview our answer to this puzzle, the Republican repeal drive is perhaps the strongest evidence to date of the “asymmetric polarization” of American politics—the sharp movement to the right, away from cross-party compromise, and toward antigovernment fundamentalism within the GOP, without a corresponding movement (at least until 2017) of Democrats to the left (Hacker and Pierson 2005; Mann and Ornstein 2013). These developments are rooted in, and have in turn furthered, a growing disconnect between ordinary voters and their putative representatives, creating a puzzling situation (at least to those holding fast to traditional theories of American politics) in which Republicans have expanded their political sway even while pursuing policies with little public support.
At the same time, asymmetric polarization does not wipe away the underlying political dynamics that make retrenchment a perilous and typically ill-fated exercise. Indeed, we can understand the policy design and procedural course of the GOP effort only in light of some of the major lessons about the resilience of social programs that four decades of failed full-scale retrenchment have taught. For one, the ACA proved to be less firmly entrenched than many advocates hoped (in part because of Republican resistance). For another, Republicans tried to exploit procedural and policy strategies designed to overcome the remaining barriers they faced. Finally, the 2017 debate highlights some new barriers to retrenchment that have arisen in the age of asymmetric polarization. In other words, the hurdles to retrenchment have not simply diminished; they have changed. The dog that almost barked stayed silent for some unexpected reasons.
Our journey into this new political environment begins with a review of the two countervailing forces shaping it: the long-standing barriers to retrenchment and the growing incentives and opportunities for Republican radicalism. We then explore the limits of the ACA's entrenchment, many of which also turn out to be rooted in the asymmetric polarization of the parties. Next, we turn to the Republican repeal strategy and why it came up just short. We conclude with some reflections on the new new politics of the welfare state in an age of asymmetric polarization.
Why Retrenchment Is So Hard
Today, it is conventional wisdom to describe major social programs as a “third rail”—touch and die. But it was not always so. Many expected the Reagan presidency to result in a substantial downsizing of the American welfare state. Reagan entered office on the heels of the economic turmoil and tax revolts of the 1970s. He had a Senate majority. And he had many conservative Democrats to work with in the House. In the end, however, Reagan had a limited direct effect on the major elements of the American welfare state: Social Security; Medicare and Medicaid; tax breaks for private social policies, such as employer benefits; and even most antipoverty programs. Most of these policies survived, continued to grow, and in some cases were notably strengthened (Pierson 1994).
The literature on welfare state retrenchment attributes this resilience to two main factors. First, social programs are popular, and it is hard for elected representatives in a democracy to do unpopular things. The popularity of the welfare state makes retrenchment an exercise in “blame avoidance” (Weaver 1986) rather than “credit claiming” (Mayhew 1974). To avoid electoral disaster, politicians must pursue various strategies designed to diffuse or occlude responsibility: political strategies that lessen the risks to a particular party, such as oversized bipartisan majorities; procedural strategies that limit the degree to which voters can identify cuts with particular politicians; and policy strategies that spread costs in the future or hide their magnitude through other means (Weaver 1986; Arnold 1992).
The second source of welfare state resilience is what has come to be called entrenchment—the ways in which various individuals and groups become invested in particular programs and thus gain increased incentive to defend them (and often increased capacity as well, due to the resources that policies bestow on them). The concept of entrenchment is grounded in the idea of policy feedback (Weir and Skocpol 1985), the observation that policies and programs, once enacted, can reshape politics in fundamental ways. Some policy provisions give rise to strong support constituencies and make cutbacks highly visible, while others do not. These constituencies usually include program beneficiaries, of course. But they can also include third-party providers (e.g., doctors and hospitals) and others aided indirectly by social programs. In dealing with these stakeholders, strategies of blame avoidance are crucial, though they are difficult to engineer, given the sophistication and monitoring capacity of many of these actors. In short, the barriers to retrenchment, though always daunting, are variable, not fixed.
Table 1 presents a basic inventory of retrenchment activities, arrayed roughly according to their likelihood of success. The table describes the background conditions that make each form of retrenchment more likely and provides some examples. As the table suggests, the goal that Republicans set for themselves was virtually unprecedented: we have no modern US example of cutbacks of current benefits of this scale.2 Rolling back a program that reaches tens of millions of Americans directly and pumps hundreds of billions of dollars per year into the health care sector should not be easy.
Perhaps, however, the ACA was uniquely vulnerable with regard to the first source of resilience: the general popularity of social programs. Maybe the ACA was simply unloved, and Republicans were doing the popular thing by trying to dismantle it. Although social policies are usually well liked, not all are. In the run-up to the welfare reform law of 1996, between 60 and 70 percent of the public supported the major elements of welfare reform: time limits, mandatory work requirements, and the replacement of cash payments with in-kind benefits like child care (Shaw and Shapiro 2002). While public views about welfare were not monolithic (Weaver 2000), a strong case can be made that welfare reform was aided rather than hindered by public opinion.
Was the same true of the ACA? Certainly it never enjoyed the overwhelming popularity associated with Social Security and Medicare. But neither does it look like welfare. In general evaluations of the program, opinion has split along partisan lines, with unfavorable views slightly outweighing favorable ones—at least until 2017, when positive evaluations began to trend upward. Still, polls have shown consistent support for all the central elements of the law except the individual mandate. Moreover, opponents of the ACA include a nontrivial portion who think it does not go far enough, and an even larger share who believe that, despite their reservations about the law, it should not be repealed and replaced. A fair reading of the evidence is that the negative reactions to the ACA constitute expressions of partisan grievance more than assessments of concrete policies and that the concrete policies in the law are actually pretty popular, albeit not nearly as popular as older pillars of the American welfare state (Jacobs and Mettler 2016).3
In any case, it is not plausible to argue that Republicans' repeal and replace bills provoked public support, because they did not. To the contrary, they polled terribly—whether viewed on their own or in comparison with the ACA. In one survey taken on the eve of the House vote, just one in six Americans (17 percent) said they supported the Republicans' current bill (Bump 2017). Even among Republicans, the bill had just over 40 percent support.
These bottom-feeding numbers may well be unprecedented. Figure 1 presents the results of a recent analysis of fifteen major bills that were considered in Congress from 1990 on (and on which multiple surveys exist). This analysis suggests that the Republican repeal drive may be the most unpopular piece of major legislation with a real chance of passage in the past quarter century (Warshaw 2017). (The next most unpopular, revealingly, is the tax bill that Republicans ultimately enacted at the close of 2017.) By way of comparison, polls taken on the eve of ACA's passage—after a year of intense attacks on the law as a government takeover of medicine—showed 35 percent (Fox) to 47 percent (ABC-Washington Post) of voters in support (Blendon and Benson 2010).
In short, our mystery will not be solved by positing that Republicans were taking a popular route. If there was any ambiguity about that before their drive began, there was none once it launched. In fact, by pushing extremely unpopular alternatives to the ACA and thus making the ACA's popular features more salient, Republicans did more to improve the law's public standing than Democrats had in seven years: between October 2016 and October 2017, according to a Kaiser Health Tracking Poll (Kaiser Family Foundation 2017), support for the ACA increased from 44 percent to 51 percent.
Past work on retrenchment suggested that this abysmal polling should have been a flashing red light. Bills so unpopular should die; indeed; they shouldn't even be born. So why did Republicans press forward nonetheless? To answer that question, we need to turn from the sources of welfare state resilience to the roots of GOP radicalization.
The ACA Meets Asymmetric Polarization
A revealing irony of the eight-year battle to repeal the ACA is that the law itself grew out of Republican ideas. Many of its core elements reflected the bipartisan reform passed in Massachusetts in 2006, which had been supported by Republican Governor Mitt Romney (much to his inconvenience when he ran against Barack Obama in 2012). The idea of the individual mandate, in particular, traced back to proposals put together by right-of-center health care experts in the late 1980s (Butler 1989) and early 1990s (Pauly et al. 1991). Democrats added more progressive elements to this mix, including a major expansion of Medicaid, the joint federal-state program for low-income Americans. But the ACA was hardly a wild-eyed liberal project, and indeed, to squeeze the law through the Senate, Democrats jettisoned some of the more progressive elements of the plan that President Obama campaigned on in 2008—most notably, the “public option” allowing those without coverage through their employers to buy into a Medicare-like public plan.
Yet during the battle over the ACA, not a single Republican voted for the final legislation in either house of Congress. And once the ACA passed, Republicans moved into a stance of unceasing opposition that has no real precedent in the history of federal social policy. Not only did they vow to repeal the law in every election from 2010 on; most Republican officials did everything they could to block and derail the implementation of the law at both the state and federal levels.
In her magisterial history of Social Security, Martha Derthick notes that major social policy breakthroughs constitute “boundary” disputes—a fierce struggle over whether government should enlarge its sphere—but that “after a boundary is broken, politics can be expected to subside” (Derthick 1979: 375). This, indeed, was the pattern with many major social policies of the twentieth century: initial conservative opposition gave way to acquiescence and, eventually, to at least partial buy-in. Needless to say, it has not proved to be the pattern with the ACA. Politics did not “subside” after the ACA passed. If anything, Republicans became even more fervently opposed to the law.
Many Republicans have ascribed their scorched-earth stance to the partisan nature of the ACA itself. On this view, GOP hatred of the law is so intense and persistent because it passed without a single GOP vote. Some political science research does suggest that laws enacted with slim margins in Congress are more vulnerable to later revision (Maltzman and Shipan 2008; see also Berry, Burden, and Howell 2010), though it is not clear whether slim margins have an independent effect beyond the weaker support for a law they signal. A recent event-history analysis of the repeal of landmark laws between 1951 and 2006 indicates that repeal is initially less likely when laws are passed under unified government but that such laws are more vulnerable in the medium term—presumably because of the weaker buy-in they receive from the opposition party (Ragusa 2010). (The analysis also finds that the chance of repeal then declines more or less monotonically, suggesting that entrenchment makes repeal progressively less likely with the passage of time.)
In light of these analyses, a case can be made that the ACA would have been less vulnerable had it attracted significant Republican support. But this only pushes the question back: Why did it prove so difficult to gain such support? Why were House and Senate leaders—and, in particular, Senate Minority Leader Mitch McConnell, who staked out a strategy of unwavering obstruction even before Obama's inauguration—so successful in keeping Republicans unified in opposition? It is hard to argue that this was mostly a function of Democratic strategies or the structure of the ACA. For starters, President Obama and his allies initially hoped they could attract support from at least a few moderate Republicans in the Senate (Brill 2015). More important, while the ACA was significantly to the left of Republican goals on health care, this was more a function of the rightward shift of Republicans than any movement among Democrats. (Indeed, Barack Obama's health proposal was arguably to the right of Bill Clinton's.)4
In short, the key to understanding the unified, ongoing GOP opposition is not what happened on Capitol Hill in 2009 and 2010 but what has been occurring within the Republican Party since at least 1990. As political scientists have extensively documented (Hacker and Pierson 2005, 2015, 2016; Mann and Ornstein 2013; Bonica et al. 2013; Skocpol and Jacobs 2011), the two parties have not moved away from each other at equal speeds. Instead, polarization has been driven by the sharp and ongoing rightward movement of the GOP. This can be seen in roll-call votes in Congress, in the positions of presidents and vice presidents, in ideological divisions on the Supreme Court, and, to a lesser extent, in voting patterns in state legislatures (Hacker and Pierson 2015). Even more important, it can be seen not just on measures of relative positions, such as voting divisions, but also on measures of absolute positions. On core economic issues that divide the parties—taxation, federal spending, and, yes, health care—GOP rhetoric and proposals have moved sharply right, without a corresponding leftward shift on the other side. In short, polarization has been asymmetric.
Political scientists have been slow to examine asymmetric polarization because it is not what canonical models of American politics would predict. In a two-party system in which candidates must ultimately face voters of both parties, there should be strong pressures for ideological convergence. Candidates might veer to the right or left to win over the more extreme voters who dominate party primaries, but they should be highly constrained in their capacity to do so by the imperatives of the general election, when they will have to attract centrists as well. If polarization occurs in these models (and, in general, highly polarized parties should not be a stable equilibrium), that polarization should be more or less symmetric. Otherwise, the party that didn't stray from the center would be hugely advantaged. Thus, the disproportionate shift of Republicans away from the center raises two difficult questions: why have Republicans moved so sharply right, and why have they not paid an evident electoral price for this movement? We do not attempt to answer these questions fully here (for a more complete answer, see Hacker and Pierson 2015, 2016: chap. 8). Instead, we want to emphasize those aspects of our answer that seem most germane to the GOP repeal push.
With regard to the first question, why Republicans have moved so sharply right, we emphasize three factors. First, Republicans have a distinct electoral base: their most committed voters see politics in stark us-versus-them terms, resist compromise, and rely on highly partisan news sources in a way not mirrored on the other side of the spectrum (Grossmann and Hopkins 2015). Among these voters, moreover, Obama's presidency and its legacies, including the ACA, have evoked deep—and undeniably racialized—hostility (Tesler 2012). Republicans have both stoked and profited from these rising “anti-system” attitudes, notwithstanding the tensions they sometimes create within the party.
Second, activist groups and donors associated with the GOP are more extreme and more influential than those associated with Democrats. Moreover, these groups tend to reinforce Republicans' conservative stances. By contrast, Democratic backers often represent centrist positions or focus on noneconomic issues. Thus, Republicans are emboldened by their activist backers, while Democrats are often cross-pressured. In addition, conservative groups have become expert at recruiting, certifying, and training candidates, as well as staffing their offices once they are elected. For example, the network of contributors and organizations associated with Charles and David Koch is now roughly equal in financial heft to the Republican Party itself (Mayer 2016; Skocpol and Hertel-Fernandez 2016). And these groups have not simply increased Republicans' desire to pass inegalitarian legislation. They have also increased their ability to do so, since these actors play a fundamental role in backing GOP campaigns (as well as providing comfortable post-career perches for those who retire from or lose office, further blunting electoral incentives for moderation).
Third, these forces have had the greatest effect on the Republican Party's positions and aims with regard to tax policy. Over the past generation, the GOP has transformed from a generally antitax party into a party that places tax cuts for the affluent above all other priorities, including the traditional conservative goal of deficit reduction. In this respect, the most salient benefit of rolling back all or part of the ACA is that it directly and indirectly enables big tax cuts—directly because the ACA contains several highly progressive revenue sources that Republicans were determined to reverse, and indirectly because the savings from these cutbacks free up fiscal space for cuts in corporate and estate taxes and other changes to the tax code that will disproportionately benefit high-income taxpayers.
Tax cuts for the rich, however, are viewed skeptically even among Republican voters (Filer 2017). Thus, not only did the Republican repeal drive threaten to impose big losses on beneficiaries and third-party providers; it did so in pursuit of a goal that was itself distinctly unpopular. How did Republicans expect to get away with this one-two punch? Here we come to the second question raised by asymmetric polarization: why Republicans have had the means, as well as the motive, to pursue extreme policies.
This question is best explored in the context of the actual strategies Republicans chose in 2017, which we take up shortly. But to foreshadow that examination, Republicans believed they were insulated from backlash by at least three factors. First, and probably most important, they anticipated that they enjoyed a substantial electoral cushion. To hold Congress in 2018, for instance, it is generally estimated they can afford to lose the national popular vote by six or seven points (Silver 2017). Pundits tend to attribute this advantage to GOP gerrymandering, which certainly bolsters Republicans' edge in the House. Yet Republicans are roughly as advantaged in Senate elections, and, unlike district boundaries, state lines are fixed.
The deeper source of the GOP edge is that America's electoral system rewards parties for holding territory in addition to garnering votes. In recent decades, Republicans have grown increasingly dominant in the most sparsely populated parts of the nation, while Democrats have gained in urban areas (Hopkins 2017). This geographic edge has obvious benefits in the Senate, where every state (however sparsely populated) gets two seats. But it is also of great value in the House, because it gives Republicans large but not overwhelming margins in large swaths of the country—an advantage reinforced, yes, by Republican gerrymandering. The more “efficient” distribution of GOP voters means that the pivotal seats that determine control of Congress are substantially more Republican leaning than the country as a whole. Although a big Democratic surge could overcome these advantages, GOP leaders are betting they can weather a more typical electoral storm.
Which brings us to the second source of Republican confidence: they believe that Republican voters will stick with them even though they don't like major GOP policies. The foundation of this expectation is what political scientists call “negative partisanship” (Abramowitz and Webster 2016). Voters haven't just become increasingly loyal to their party; their political preferences are increasingly driven by hatred or fear of the other party. Moreover, such tribalism appears far stronger on the GOP side. In 2016, tens of millions of Republican voters cast their ballot for a presidential candidate they acknowledged was unqualified for the job, mainly because they could not bring themselves to vote for his opponent. For Republicans, negative partisanship provides another layer of protection against backlash driven by unpopular policies: if voters can be mobilized by animus toward the other side, elected officials don't need to attend to their actual preferences.
Finally, within the context of their extreme policy ends, Republicans designed their health care bills in ways that aimed to minimize the degree to which voters might mobilize against them. Particularly important here was their reliance on the devolution of policy responsibility to the states, which GOP leaders hoped would reduce the degree to which national politicians were held accountable for unpopular cuts. Along with the other factors just discussed, strategic policy engineering was meant to provide “backlash insurance” (Hacker and Pierson 2005) that GOP elites believed would insulate the party from significant electoral fallout.
Because Republicans failed to enact their bills in 2017, we cannot know for certain how accurate these expectations were. What we can know is that Republicans acted in ways that suggested any concerns about electoral risks placed only the loosest of constraints on their actions. Given their narrow margins and the enormous unpopularity of their bills, this itself is telling. The scholarship on retrenchment has emphasized the barriers to social policy reversal created by “the electoral imperatives of modern democracy” (Pierson 1996, 179). But during the 2017 struggle, these imperatives did not seem so imperative. Instead of electoral risks, the main external constraints that Republicans faced related to the uneven entrenchment of the ACA, our next topic.
The Limits of Entrenchment
The resilience of the welfare state is a result not just of its popularity among voters but also of its entrenchment in social and economic life. The line between popularity and entrenchment is fuzzy—the latter obviously increases the chance of the former. In general, however, entrenchment refers to those features of the program that (a) activate particular groups or individuals in its defense or (b) affect the degree of social and economic disruption that reversing prior initiatives entails. Entrenchment can make major change to a program so disruptive that policy makers will be unwilling or unable to challenge it even when that program is not popular or lacks salience among ordinary voters.
Social scientists have intensively studied when and how policies (and other institutions) become entrenched. The formula for success turns out to have two main ingredients: strong support coalitions, generally based on the receipt of clear and tangible benefits, and specific investments. Strong support coalitions are most likely when laws deliver concentrated benefits that can be traced back to government, the beneficiaries are either already organized or can be relatively easily organized, and these beneficiaries have sufficient resources, organization, or sympathy to enjoy political influence. Specific investments occur when people premise their lives and livelihood around the expectations created by social policies, making them (and public officials) more sensitive to the losses and dislocations that will occur if these policies are changed. As a broad rule, support coalitions and specific investments are likely to become a bigger factor over time, as groups and individuals make high-stake commitments that increase their incentive and capacity to defend concentrated benefits.
Judged according to this formula, the ACA entered 2017 relatively underentrenched (Oberlander and Weaver 2015). After all, many of the law's provisions did not take effect until 2014—by which time Republicans had already seized both houses of Congress. The deeper problem was that the ACA was designed with enactment, not entrenchment, in mind. The strategy of the law's architects was to expand and improve coverage in ways that (a) would minimize obvious disruptions to existing coverage and (b) could attract the support or at least acquiescence of powerful stakeholders, such as health insurers and providers. Ironically, this strategy reflected the entrenchment of prior health care arrangements. Despite its many flaws, the US health financing system provides reasonable protection to the majority of Americans, while obscuring the true incidence of its extraordinarily high costs. These high costs, in turn, support a formidable assortment of health industry players who profit handsomely from the system's many inefficiencies. The ACA was designed to bypass (read: buy off) these potential sources of opposition. That was the way to get more people covered and to create a secure source of fallback coverage for everyone. Unfortunately—and, for the most part, unavoidably—it was not the way to build a program that would create strong support coalitions or extensive specific investments. Put another way, some of the ACA's central political liabilities in 2017 were its central political advantages back in 2010.
What were those liabilities? Most obvious was the very weak development of the ACA's regulated marketplaces. The CBO had projected they would enroll roughly 23 million Americans; their actual enrollment in 2016 was closer to 10 million (Jackson 2017). This low number was a product of ongoing interference with the ACA's implementation by Republicans, as well as weaknesses in the law itself. Predictably, it was accompanied by greater than expected costs per patient (i.e., adverse selection). In turn, the small size and high cost of this population deterred commercial insurers from participating in the marketplaces, limiting the degree to which they became a strong element of the ACA's support coalition.
Importantly, many of these features of the ACA that weakened entrenchment were forced on, rather than pursued by, the designers of the law. For example, the version of the legislation that passed through the House in late 2009 contained provisions that integrated the law much more closely with employment-based health insurance. It also envisioned a single national regulated marketplace rather than state-by-state implementation. And it contained the so-called public option that would have made a Medicare-like public plan available through the ACA's individual marketplaces. Both these marketplace-related provisions would probably have increased the size and allegiance of the population receiving regulated private coverage. But like the employer-related provision, both were stripped out of the law in the Senate because of the resistance of more conservative Democrats. The caution and leverage of these Democrats, in turn, reflected both the considerable clout of provider groups and the fierce attacks and unified opposition of Republicans.
The biggest example of the ACA's compromised entrenchment, however, was the uneven reach of the law's Medicaid expansion. During the 2017 debate, Republican governors from states that expanded Medicaid proved to be crucial allies of the law—ones that, unlike Democratic governors, congressional Republicans could not wholly ignore. But the strength of these partisan cross-pressures was much less than it would have been had all states expanded their Medicaid programs. Thus, the ACA's entrenchment was greatly limited by the Supreme Court's 2012 decision allowing states to turn down the expansion (an opportunity that eighteen states, all Republican controlled, continue to exploit, although four are currently considering expansion). Here again, Republican resistance proved crucial in shaping entrenchment—on the Supreme Court and in the states alike.
The limits of the ACA's entrenchment suggest that more attention should be paid to how opponents of the welfare state, not just its designers, affect the support coalitions and specific investments that new social programs generate. In 2009 and 2010, Democrats understandably focused more on enactment than entrenchment. Yet several of the biggest weaknesses of the ACA reflected not deliberate design choices but the unified and ongoing opposition the law faced. The upshot was that in 2017 the ACA was significantly more vulnerable to Republican resistance than it might have been, because of prior Republican resistance. How Republicans tried to capitalize on this vulnerability and why they failed are the next questions.
The GOP Strategy and Its Discontents
After seven years of unrelenting hostility to the ACA, Republicans assumed unified control of Washington in 2017 for only the second time in sixty years (the first, including a brief interruption, ran from 2001 to 2006). Yet the roots of their strategy for repealing and replacing the 2010 law ran deeper. Since the 1990s, the asymmetric polarization of the parties has energized a loose but increasingly formidable GOP coalition with ambitions to fundamentally roll back the American welfare state. From the moment of President Obama's election in 2008, GOP congressional leaders argued that if Republicans were given unified control of government, they would forge a new governing contract. These priorities were most concretely embodied in a series of budget blueprints developed by Paul Ryan, the Wisconsin congressman who would rise from chair of the House Budget Committee to chair of the House Ways and Means Committee to Speaker of the House in 2015 (and announce his retirement in 2018).
The Ryan budgets all had three unifying features: they massively cut taxes, especially on the wealthy; they massively cut spending, particularly on programs for the poor but also on those that benefited the middle class; and they hid the deficits they were certain to produce by using budget-scoring sleight of hand (Hacker and Pierson 2012). After 2010, these budgets rolled the goal of repealing the ACA into the package. But the overarching objective remained high-end tax cuts. For example, while Republicans demonized the ACA's cost-control measures for “robbing” Medicare (in fact, the law actually expanded Medicare's benefits while improving the program's long-term finances), they carried over all the Medicare cost savings in their post-2010 budgets while cutting the ACA's progressive tax increases and coverage-enhancing provisions.
In short, the Republicans' repeal drive was not a stand-alone goal. It was part of an integrated strategy that centered on freeing up fiscal space for large tax cuts. With their sweep of 2016's election, Republicans finally had their chance to deliver. A week after the election, Ryan boasted to reporters, “Welcome to the dawn of a new unified Republican government. This will be a government focused on turning President-elect Trump's victory into real progress for the American people” (DeBonis 2016). Ryan's choice of the phrase “Trump's victory” rather than “Trump's agenda” was telling: Ryan was not promising to implement all of the president-elect's policy proposals. Trump presented himself as a very different kind of Republican than Ryan, professing a muscular populism alongside traditional GOP priorities. Still, “repeal and replace” had been among Trump's top promises, and both he and congressional Republicans vowed to move quickly.
Republicans cycled through three main legislative strategies for dismantling the ACA. The first, which might be called “shock and awe,” was to simply pass a repeal bill that completely eliminated the ACA at a fixed point in the future. The idea was that Democrats would then be forced to work with Republicans to come up with an alternative before the law's new expiration date. The second strategy—call it “Trojan horse”—revolved around bills that preserved some of the elements of the ACA but set up policy dynamics that would essentially reverse the law's main provisions over time. The third and most audacious strategy deserves to be called “mystery meat.” At several points, Republican leaders in the Senate told rank-and-file members to vote on a bill that was just a placeholder for negotiations with the House, which had already passed legislation. Observers of the Senate debate were greeted with the spectacle of Majority Leader McConnell assuring his fellow Republicans that he was committed to making sure that the mystery meat bill on which they were voting did not become law.
All of these strategies revolved around the budget reconciliation process, which has become the one fast track for passing legislation in the Senate. Under this process, budgets can be passed with a simple majority, bypassing the sixty-vote threshold required to end Senate filibusters. When Senate Republicans had been in the minority, they had led the way in transforming the filibuster from a rare occurrence into a routine practice of the minority party. Now, they needed to bypass the Democratic Senate minority, and with only fifty-two Republicans in the Senate, that meant using budget reconciliation. It also meant they could lose only two Republican votes (in a fifty-fifty situation, the vice president casts the tie-breaking vote).
Yet the budget process constrained as well as enabled Republicans. Under its rules, legislation had to be germane to the budget and could not create deficits beyond the ten-year budget window. The deficit rule was not an issue, since all the GOP plans were designed to produce budget savings that could be used for tax cuts. But Republicans discovered that many of the ACA's regulations could not easily be undone without breaching the rules. (In practice, the Senate parliamentarian—a nonpartisan officer—polices these rules; Senate Republicans could overrule the parliamentarian but that would essentially mean ending the filibuster, a radical step that pivotal members have been unwilling to take.)
The shock-and-awe strategy was a direct response to these constraints, since it would leave the construction of an alternative to the normal legislative process. Impending repeal would hang like the sword of Damocles over Congress. But it quickly became clear that this approach was too sweeping and risky to attract the support of the least extreme Republicans. In effect, Republicans discovered that they did not have large enough margins to pursue the most ambitious type of retrenchment outlined in Table 1, namely, completely eliminating an extant social program that already had beneficiaries.
Nonetheless, their fallback strategy—more cosmetically pleasing legislation that was a Trojan horse for doing away with much of the ACA—still fell far outside the bounds of feasibility suggested by previous scholarship on retrenchment. As noted, Republicans sought to cut Medicaid by up to $1 trillion over ten years and transform the other parts of the ACA in ways that were certain to increase insurance and out-of-pocket costs by thousands of dollars a year for those buying insurance through the individual marketplaces (CBO 2017a, 2017b, 2017c, 2017d). Because they could not repeal the main regulations in the ACA directly in the reconciliation measure, they resorted to various means of defanging them, such as defunding their enforcement, creating new escape routes and safe harbors, and giving states “flexibility” they could use to escape them over time. But what was inside the Trojan horse was a radical retrenchment of the ACA.
This strategy was not simply a response to the budget rules. Republicans also believed it could limit future electoral fallout for national Republicans. The key here was the long-standing GOP strategy of devolution, embodied in waivers for states to cut back popular insurance regulations and in the transformation of Medicaid into a federal block grant that would be allowed to grow only at a constrained rate. Though justified with calls for state experimentation and control, these devolutionary steps were designed to shift onus for those cutbacks away from Washington and onto state governments. It would be a mistake, therefore, to argue that Republicans were unconstrained by the imperative of blame avoidance. Still, what is striking is how bold GOP leaders' aims remained.
Nowhere was this more evident than with regard to Medicaid. By 2017, the program had expanded to cover some 75 million Americans (including the Children Health Insurance Program). Moreover, it has become a lifeline for many of the working-class white communities that voted overwhelmingly for Donald Trump in 2016. Thus, Republicans' aggressive efforts to retrench Medicaid provided a clear signal of how unconstrained by public opinion they were—even opinion within their own party—and how confident they appeared to be that they could manage the resulting electoral risks. Although GOP leaders railed against the ACA, the biggest effects of their bills would have been on Medicaid, and these effects would have gone well beyond reversing the 2010 expansion of the program. “Repeal and replace Obamacare,” it turned out, mostly meant “repeal” Medicaid as it had been known for decades and “replace” it with a program that was designed to shrink dramatically over time.
In one sense, the GOP focus on Medicaid is unsurprising: Republicans have long wanted to limit the program's funding. (During the 2017 debate, Speaker Ryan blurted out to a reporter for the National Review: “We've been dreaming of this since we were drinking from kegs” [Yglesias 2017].) Nor have contemporary Republicans been known for their affection for programs that target low-income Americans. But both the party's electorate and Medicaid have changed substantially in recent years. As the party's base has shifted toward rural and working-class voters and Medicaid has grown, the program has become a vital source of protection for many of the party's core constituents.
And yet Republicans in Washington appeared mostly unconcerned about the potential for backlash. Part of the reason, as discussed, was that they believed devolution would limit accountability. Even more important, however, was their confidence that GOP voters—with their deep animosity toward Democrats and heavy reliance on partisan sources of information—would back them regardless. The rise of negative partisanship, with voting increasingly dictated by tribal loyalties rather than policy stances, relaxes the constraints of public opinion on legislators (Abramowitz and Webster 2016). Because such tribalism appears strongest on the Republican side of the electorate, and because it is Republican elites who are most hostile to the welfare state, negative partisanship provides another mechanism by which asymmetric polarization has lowered the barriers to retrenchment: when members of the retrenchment-seeking party are highly unlikely to defect, the scope for changes they might dislike becomes far greater.
In sum, the GOP assumed unified control of Washington seeking fundamental cutbacks in relatively popular policies and programs. And while Republican leaders backed away from the most ambitious strategy to achieve that goal—shock and awe—the Trojan horse bills they embraced still departed starkly from the careful blame-avoidance strategies anticipated by the scholarship on retrenchment. These bills relied on devolution to diffuse blame, but they still threatened huge losses, losses that would be particularly large for core GOP voters.
Of course, Republicans came up just shy of assembling the fifty votes they needed. They remained short even when they tried to convince Republican legislators to embrace mystery meat that could at least keep the repeal drive moving forward. But they failed for reasons that depart from as well as reinforce the conventional perspective on retrenchment.
Revenge of the Right
Why did Republicans fall short? One potential answer is that defeat was foreordained: retrenchment of this scale is simply impossible. But this argument is unconvincing—Republicans came within a few votes of victory. Had their majority been anything more than razor thin, they likely would have succeeded. As we have argued, moreover, the very fact they tried to pass such ambitious legislation requires explanation. If the barriers to retrenchment had remained as formidable as previous research on retrenchment would lead one to expect, it is hard to believe that Republicans would have set their sights so high.
For all the vulnerabilities of the ACA, Republicans faced significant mobilized opposition from the day Donald Trump took the oath of office. Not only did Democrats in Congress universally oppose the Republican effort—which, like the ACA, received not a single vote from the other side of the aisle—but outside groups on the left also staged a spirited campaign to publicize the most extreme elements of the legislation. Most striking of all, almost every major organization with a stake in health care, from the American Medical Association to the AARP, declared its opposition. The insurance industry was closer to the fence, but it too made clear it was not enthusiastic about Republicans' plans.
The consequences of entrenchment were also on display in the opposition of several Republican governors who had supported the expansion of Medicaid. As noted, the incomplete implementation of the Medicaid expansion meant that the number of such in-party critics was relatively small. But they were another source of cross-pressure on national Republicans as they pursued large cutbacks in Medicaid.
Still, most Republicans in Congress appeared immune to these demands. For example, every one of the Senate Republicans who represented states where Republican governors spoke out made clear they would vote in favor of the repeal bills. (The ambiguous case is Senator Lisa Murkowski of Alaska, who, along with Senator Susan Collins of Maine, consistently expressed doubts about the legislative effort; the governor of Alaska also pushed back, but he is an independent rather than a Republican.) Republicans did worry about potential electoral fallout; they did pursue strategies that were designed to minimize the visibility and traceability of their efforts; and they did face cross-pressures due to the entrenchment of the law. Still, these forces were far less constraining than prior work implies they should have been. At the same time, several factors not highlighted by this scholarship proved at least as crucial in undoing Republicans' plans.
Let us stipulate the obvious: the central problem Republicans faced was that they had such narrow margins. Using the reconciliation process meant it was possible to pass a bill with only their own party's support in the Senate (though it limited what it could contain), but it meant that even the slightest within-party opposition could doom the effort. Again, though, the fact that Republicans could afford so few defections makes the ambition of their effort notable.
Equally notable—and more surprising—is where the risk of defections arose. In standard models of legislative politics, centrist members of Congress are the pivotal actors. By contrast, extreme conservatives (or liberals) have little leverage because they cannot credibly claim that they will oppose legislation that moves policy in their direction, even if that legislation does not go as far as they prefer. During the health care fight, however, the right of the party imposed clear constraints on what congressional GOP leaders felt they could do. Whenever those leaders tried to broker compromises to appease more moderate members, conservatives threatened defection.
How can we understand conservative resistance to conservative legislation? Conservatives faced genuine electoral risks—from the right. Because they tended to come from overwhelmingly Republican districts, they knew the greatest threat they faced was a primary challenge financed by outside conservative groups. Moreover, as Ruth Bloch Rubin (2017) has convincingly argued, conservatives within the House—under the banner of the “Freedom Caucus”—are much more organized than the more moderate members who ostensibly represent the pivotal votes in that chamber. This organizational edge allows them to form long-term positions and discipline wayward members in ways that increase their capacity to extract concessions, even when they are a distinct (and extreme) minority. Whatever the explanation for the Hard Right's willingness to defect, it posed a significant—and, to standard models, surprising—challenge for GOP leaders.
On a deeper level, the constraints that GOP elites faced were rooted in acute tensions within their own coalition and between their own policy goals. The scholarship on retrenchment has generally assumed that opponents of the welfare state will choose their strategies so as to minimize opposition. But again and again, Republicans chose their strategies in ways that increased the hurdles they faced. They did so for two interwoven reasons: first, to maximize the scope for achieving their primary aims—tax cuts for the affluent and fundamental retrenchment of the American welfare state—and second, to prevent defections on the right. The GOP legislation was so radical because “repeal and replace” was a means of achieving extremely ambitious policy goals that went well beyond the ACA. It remained so radical because whenever GOP leaders tried to scale back their ambitions, they risked defections from their conservative flank.
Postmortem analyses have stressed GOP incompetence. Whenever Republicans put forth actual bills, rather than mystery meat designed to keep the dream of repeal alive, the CBO came back with unwelcome news: their bills simultaneously raised premiums and out-of-pocket costs, lowered the quality of insurance plans, increased the chances of insurance market death spirals, put new pressure on state budgets, and massively increased the ranks of the uninsured. But these ugly outcomes were not, for the most part, a result of sheer policy ineptitude. Instead, the main reason for the bill's flaws was the overriding priority Republicans placed on delivering huge tax cuts to a narrow set of beneficiaries. Some of the bills rolled these tax cuts into the package; others aimed to produce large amounts of deficit reduction that would allow tax cuts in the next legislative round.
Imagine an alternative scenario, in which Republicans had curbed their appetite for tax cuts and used the funds freed up by that choice to instead grease the wheels for conservative structural reforms in American health care. As the retrenchment literature shows, taking benefits from people is easier if policy makers compensate some of the losers and delay the most negative effects. Republicans could have made the cuts in Medicaid smaller and more gradual, in the process buying greater support from GOP governors and the health care industry. They could have offered bigger subsidies to Americans buying insurance on their own and, in particular, increased the subsidies for older voters and for voters in rural areas (where health care tends to be costlier). The goal would have been not to shore up the ACA but to build a gentler off-ramp that would have transitioned people out of it and into a system that was closer to Republicans' vision. They could, in short, have softened the immediate pain in return for cementing a long-term policy victory.
Without question, this strategy would have had a greater likelihood of attracting the more moderate Republicans the GOP leadership ended up losing. But it would have threatened defections on the right side of the party, which only highlights the tightrope Republicans walked. Indeed, most of the procedural strategies GOP leaders adopted—refusing to hold committee hearings, rolling out legislation at the eleventh hour, insisting on a vote even before the CBO had provided its full analysis—were designed not only to hide what they were doing from Americans but also to finesse divisions within their own party. The goal for party leaders was to turn the repeal vote into a simple yes or no decision within their caucus and then rally Republicans around the imperative of passing something, anything, that would please their big donors and electoral base. “Just say yes” reached its apotheosis when Senator McConnell implored his GOP colleagues to swallow mystery meat they all hated, merely because it would move them closer to that goal.5
The New New Politics of the American Welfare State
For those who believe that policy outcomes in a democracy generally should reflect majority opinion, the scholarship on retrenchment offers a comforting conclusion: welfare states are resilient because they are popular and because politicians must be attentive and responsive to that popularity. Yet, in the battle over the ACA, the views of ordinary voters were not the main constraint on Republican ambitions. Party leaders pursued policies that were manifestly and massively unpopular, yet they came remarkably close to success—the dog almost barked.
To explain why, we have turned to the growing body of work on the asymmetric polarization of the parties. This research has emphasized that Republican radicalization is rooted in distinct electoral and organizational realities that have simultaneously increased the party's desire to move toward economically conservative, anti-system stances and its capacity to do so. Many of these realities were on display in the party's drive to repeal and replace the ACA.
At the same time, we have examined why the entrenchment of the ACA was more limited than its designers and defenders hoped—often because of prior GOP resistance. The processes through which policies develop support coalitions and become integrated into social and economic life are not automatic, nor are they simply a product of legislative design. In today's polarized political climate, entrenchment is itself a contested process, in which opponents seek not only to cripple policies at the outset but also to limit the degree to which they become embedded over time. Here, too, the asymmetric polarization of the parties looms large.
We began with the metaphor of the dog that almost barked. We close with another that highlights the gradual yet profound structural changes occurring in the American polity. What asymmetric polarization is doing to the political environment is similar to what global warming is doing to the natural environment. If we travel to warming wilderness, we see flora and fauna growing where they once could not, displacing but not completely replacing what was there before. On mountains, for example, the tree line is rising, and the mix of trees that can grow at high altitudes is changing.
The wilds of Washington, DC, are undergoing a similar transformation. They still contain familiar inhabitants doing some of what they did before, but it is now possible to see activities that would have been unthinkable a generation ago—like a direct assault on a major part of the welfare state. The political “tree line” for extreme policies has shifted, so that today even slim majorities can translate into radical legislative drives.
The result is a new ecosystem in which efforts at retrenchment that could not have survived in the past now have some prospect of success. In this respect, asymmetric polarization is reshaping our political environment in ways not so dissimilar from how global warming is transforming our natural environment. Like trees growing on mountain slopes that were once barren, extreme policies now have a chance of taking root where they once were destined to die. As the failure of the GOP repeal drive indicates, there is no guarantee that these invasive species will thrive. Nor is it likely that the political climate will remain so hospitable to extremism if Republicans suffer truly catastrophic electoral losses in 2018 and 2020. Still, extreme policies now have a chance of surviving and spreading where once they would never have even been seen.
For comments on and conversations about this article, we gratefully thank David Mayhew, Eric Patashnik, Eric Schickler, Theda Skocpol, and Chris Warshaw, as well as all the participants in the conference on which this special issue is based. At the press, Jennifer Costanza and the entire production team were of great help. We also appreciated the stellar anonymous-reviewer feedback we received—a reminder once again of how much all of us in academia depend on the kindness of (very smart) strangers.
The mid-1990s welfare reform—which involved smaller cuts to far less popular programs—is the next closest US case. Other prominent examples include the repeal of the Sheppard-Towner Act (funding for maternity and childcare) in the 1920s, the ending of the “Model Cities” program in the early 1970s, and repeal of the Medicare Catastrophic Coverage Act not long after its passage in 1989.
The single best source for polling data on the ACA is the Kaiser Family Foundation's health tracking polls, which can be found—along with many revealing reports—at www.kff.org/program/public-opinion-and-survey-research/.
Although President Clinton's proposal placed greater emphasis on market competition, at least rhetorically, it was far more ambitious than Obama's. The Clinton plan included a “play-or-pay” requirement that employers cover their workers (“play”) or “pay” into a pool to finance publicly sponsored coverage. Moreover, it would have required all but the largest employers to cover their workers through regulated health insurance marketplaces. By contrast, the ACA, like the Massachusetts model, focused on those lacking workplace coverage, and it required neither that employers pay for publicly sponsored coverage (instead, it contained a “penalty”) nor that they enroll their workers through the regulated marketplaces.
A final point: Republicans might have managed their delicate balance if they had had a more effective leader in the White House. Trump was certain to sign anything that passed, but his sway over wavering members proved minimal. His one big effort to pressure a GOP holdout—a thinly veiled threat made to Senator Murkowski by his interior secretary—backfired. His statements on the various bills reinforced the impression that he was disengaged and erratic. Trump was a sideshow during the health care debate. If he had been the kind of ringleader that President Bush was during the fight over tax cuts in 2001—or that Barack Obama was during the health care debate in 2010—Republicans might, just might, have come up with a winning performance.