Do Major Policy Enactments Affect Public Evaluations of Congress?
The Case of Health Care Reform
David R. Jones
Professor of Political Science
Baruch College, City University of New York
Abstract
This study analyzes changes in individual evaluations of Congress immediately before and after enactment of national health care reform in 2010. It tests three alternative hypotheses: that enactment increased the likelihood of approval by demonstrating congressional competence; that it decreased the likelihood of approval by calling attention to partisan processes; or that it differentially affected citizens’ evaluations depending on their individual policy preferences. The results show enactment polarized citizens’ evaluations of Congress, with supporters of the bill increasing their approval of Congress and opponents decreasing. These findings represent the first concrete evidence that enactments can affect evaluations of Congress.
[Forthcoming in Legislative Studies Quarterly, May 2013]
Enacting laws is Congress’s main job. Yet we still do not know whether enacting landmark legislation affects citizens’ evaluations of Congress, or if so how. Understanding how enactments affect evaluations of Congress is important because congressional evaluations have significant effects of our political system. Attitudes towards Congress’s job performance cause changes in citizens’ ideological leanings (Chanley, Rudolph, and Rahn 2000). They also affect votes and outcomes in congressional elections (e.g., Jones and McDermott 2009). And they lead to changes in the degree of trust in government (e.g., Citrin and Luks 2001; Gershtenson, Ladewig, and Plane 2006). From a normative perspective, it would be troubling if the public’s evaluations of Congress do not respond to major enactments, since this would appear to diminish the incentives for elected officials to act in accordance with citizen preferences.
Most studies of the effect of major enactments on congressional approval have taken place at the aggregate level, and have found little evidence of any overall impact. However, literature in a closely related field—the impact of Supreme Court decisions on public support for the Court—shows that the additional precision allowed by a shift from aggregate to individual level analysis can sometimes lead to major new findings (e.g., Grosskopf and Mondak 1998; Hoekstra 2000, 2003; Kritzer 2001). To date, this promising approach has yet to be fully applied to the question of how enactments affect views of Congress. This is unfortunate given that congressional evaluations have been shown to have at least as wide a range of consequences for public opinion and behavior (e.g., effects on ideology, on trust in government, and on voting) as do Supreme Court evaluations.
In an effort to advance understanding of how major enactments affect public evaluations of Congress, this study adopts the individual-level, quantitative, case study approach that has proven so informative in studies of the Supreme Court (e.g., Grosskopf and Mondak 1998; Hoekstra 2000, 2003; Kritzer 2001). Specifically, I analyze how evaluations of Congress were affected by enactment of national health care reform in 2010. In the remainder of this study, I begin by reviewing the existing literature on how legislative actions affect public evaluations of Congress. I then discuss the choice of health care reform as a case study, and lay out the specific hypotheses to be tested. Next, I describe the data and method used in the study and present the analysis. I conclude with a discussion of the theoretical and practical implications of the findings, as well as avenues for future research.
Legislative Actions and Public Evaluations of Congress
The literature suggests three general theories of how congressional actions might affect public evaluations of Congress.[1] In this section, I present these general theories and then discuss what we currently know and do not know based on existing empirical evidence.
At a basic level, when the Congress acts or fails to act this may send signals to the public about its technical competence or effectiveness—its ability to get its job done. In politics, competence is widely considered to be a “valence issue”, meaning that it is favored by practically every citizen (Stokes 1963). In this case, evidence of a productive Congress as opposed to one mired in gridlock should improve evaluations of Congress across almost all citizens. Consistent with this conventional wisdom, Binder (2003, 117) posits that aggregate evaluations of Congress will be affected, at least in part, by “what Congress accomplishes.” In support of this general notion, empirical research finds that citizens’ perceptions of government effectiveness increase approval of Congress (Mondak et al 2007), and a higher percentage of legislative failures over a two-year period lowers average approval of Congress (Binder 2003).[2]
Another possibility is that a congressional action may increase the public’s focus on process concerns. The public generally dislikes congressional processes, particularly partisan conflict (Durr, Gilmour, and Wolbrecht 1997; Hibbing and Theiss-Morse 1995; 2002; Ramirez 2009). When major legislation is being debated in Congress, this debate—and the partisan conflict that it often engenders—can be more salient than usual. If a major legislative action results in increased attention to processes that are overly partisan for the public’s taste, this may lower public approval of Congress. Consistent with this view, research finds that citizens’ perceptions of a discrepancy between preferred and actual governmental processes decreases approval of Congress (Hibbing and Theiss-Morse 1995; 2002). Other research shows that quarterly approval of Congress decreases during periods when there are more veto-override attempts, which are inherently conflictual (Durr, Gilmour, and Wolbrecht 1997), and when there are more votes divided along partisan lines (Ramirez 2009).
Finally, citizens may instead focus on a legislative action’s policy content. Classic democratic theory argues that citizens prefer governments whose policy preferences are similar to their own. According to Easton (1965), when people receive their desired policy outputs from government, their support for government increases. In general, this outlook suggests that a congressional action will boost evaluations among citizens who support the specific policy, while lowering evaluations among citizens who oppose it. Consistent with this view, research finds that individual citizens’ perceptions of overall policy discrepancy decrease their support for Congress (Hibbing and Theiss-Morse 2002; Jones and McDermott 2009; Wlezien and Carman 2001).
While the literature reveals quite a bit about how evaluations of Congress are affected by certain kinds of actions, by its inaction, and by perceptions of its actions, we still know very little about how congressional evaluations respond to one of the most consequential actions of all: the passage of major new legislation. The main approach to this question has been to analyze the effect of the total number of major enactments in a quarter on aggregate approval ratings during that quarter (Durr, Gilmour, and Wolbrecht 1997; see also Jones and McDermott 2009; Ramirez 2009; Rudolph 2002). Thus far, these efforts have not found any statistically significant effects of enactments.[3] The collective impression left by these studies is that the public simply does not care about major enactments.
However, it is not possible to confidently infer how major enactments affect public evaluations of Congress based on these existing data alone. Doing so would be subject to at least two different types of inference problems. First, if certain kinds of enactments produce a negative reaction and other kinds a neutral or positive reaction, an analysis that aggregates them all together may find net effects to be insubstantial even if effects for specific types of enactments are substantial. For example, the process-based theory implies that only enactments characterized by high partisanship will have a negative effect, but aggregated measures of major enactments likely include many bipartisan items, thus making inference problematic. Second, if a given enactment is expected to produce a positive reaction among certain types of people and a negative reaction among others, analysis of the aggregate reaction to that enactment may find no net change even if most citizens’ opinions do change. For example, the policy-based theory implies that citizens who favor a bill will have the opposite reaction to its enactment than those who oppose it, but in an aggregated measure of public evaluations these effects may wash out, again making inference problematic. Recognizing some of these potential limitations, Ramirez (2009) concludes his aggregate study noting that “given the great deal of heterogeneity at the individual level, future research may uncover more complex relationships among subaggregates of the public.”
Indeed, the closely-related literature on public reactions to Supreme Court decisions shows that where aggregate studies produced null findings, studies of specific decisions that analyzed individual level opinions found significant effects of these decisions on public evaluations of the Court (e.g., Grosskopf and Mondak 1998; Hoekstra 2000, 2003; Kritzer 2001).
To date, this disaggregated approach has generally not been applied to the question of how enactment of laws affects evaluations of Congress. One exception (of which I am aware) to this rule takes the form of a telephone survey experiment.[4] In this survey experiment, conservatives who were told that Congress had recently passed a liberal enactment—a supplemental appropriations bill increasing federal Medicare spending—were less likely to approve of Congress than were conservatives who were not told, while very liberal respondents who were told about the same enactment were more likely to approve of Congress than were very liberal respondents who were not told (Jones and McDermott 2009, pp. 22-31). While this study offers some support for the policy-based theory, the artificial aspects of its experimental design (a verbal stimulus as opposed to real world stimulus) limit its external validity. In order to provide a more appropriate test of existing theories and obtain a more complete understanding of the effect of enactments on evaluations of Congress, what is needed is analysis of an actual policy enactment.
Enactment of Health Care Reform
In this section, I provide brief background on congressional passage of health care reform, discuss its appropriateness as a case study for this research, and derive three specific hypotheses to be tested.
The particular piece of legislation used in this analysis is the enactment of national health care reform in 2010. Congress gave final approval to the Patient Protection and Affordable Care Act on March 21, 2010. The law overhauls the nation’s health care system, extending access to more than 30 million people by expanding Medicaid and providing federal subsidies to help lower and middle-income Americans buy private coverage. It requires most Americans to have health insurance coverage, and regulates private insurers more closely than they were previously.
There are several reasons this particular bill is useful for analyzing the theories at issue in this study. First, all of the theories discussed in the literature section are arguments about how citizens use information they receive about an enactment. However, the public is often unaware of what legislators are doing (e.g., Stokes and Miller 1962) and thus ordinary enactments would not permit testing of these theories. Meaningful testing requires an enactment on which public attention was high. With three quarters of the public reporting following the story closely during the week it passed, enactment of health care reform clearly meets this standard.[5]
Second, a fair test of the competency-based theory—that the public will respond positively when Congress demonstrates that is can accomplish something of consequence—requires selection of an enactment that is widely seen to be important. Enactment of health care reform clear fits this description. Universal health care coverage had been a longstanding issue in American national politics since at least as far back as 1932 (Igel 2008). Not surprisingly, then, news outlets were nearly unanimous in declaring the passage of health care reform to be a momentous event in the history of congressional lawmaking (e.g., Hitt and Adamy 2010).
Third, a fair test of the process-based theory requires a bill whose enactment process was highly partisan and conflictual. The vote on final passage of health care reform received no support from Republican members, passing only with votes of Democrats (D: 219-34; R: 0-178). During the final passage debate, a series of Republican members angrily denounced the bill on the floor of the House—with Republican Minority Leader John Boehner at one point shouting, “Hell no!” during his closing speech. In the news media, Republican members complained that Democrats “rammed health care down our throats” and denounced legislative procedures used by the Democrats, such as reconciliation, that limited the possibility of Republican input. This partisan rancor was widely reported in the press (e.g., Garber 2010).
Fourth, to help distinguish the predicted conditional effects of the policy-based theory from the predicted uniform effects of the other theories it is useful that public opinion on the selected bill not be completely one-sided, but instead contain a reasonable degree of variation across citizens. A poll taken shortly before enactment demonstrates that this was the case for health care reform: 48% opposed the bill, 37% supported it, and the remainder had no opinion.[6]
Finally, in order to rule out alternative explanations for any changes in public evaluations of Congress that are observed after enactment, it is important to select an enactment that took place when virtually no other significant events were occurring in the national government. According to content analysis of news coverage conducted by the Pew Research Center’s Project for Excellence in Journalism, “health care reform dominated the news” during the period included in this analysis (Holcomb et al 2010).[7] Overall, given the nature of the theories to be tested, enactment of health care reform is a case study that is particularly well-suited to the task at hand.
In addition to providing leverage on the theoretical issues at hand, the case study also sheds new light on the passage of health care reform as an important historical event in its own right. For example, as of this writing, significant efforts are being made to repeal the health care law. Part of the rationale for this repeal is the notion that people were unhappy with Congress for passing it in the first place (Halloran 2010). By systematically analyzing actual changes in evaluations of Congress surrounding enactment of health care reform, this study provides some additional perspective to these repeal efforts.
Applying the general theories from the previous section to the case of the 2010 health care reform bill’s enactment, I propose three specific hypotheses. If citizens were primarily concerned with the legislation’s policy content, then its enactment should increase support for Congress among those who favor the bill, decrease support among those who oppose the bill, and have no predicted effect for those without a policy preference. If citizens viewed the enactment simply as evidence of the Congress’s competence, then enactment should increase support for Congress, regardless of a citizen’s policy views. Finally, if citizens were primarily concerned with the level of partisanship in the process of this enactment—and if enactment increased perceptions of congressional partisanship—then enactment should decrease support for Congress, regardless of a citizen’s policy views.[8]