Chapter 3 Theory
The major hypotheses I investigate is whether, which, and to what degree partisanship, ideology, and interests outweigh and/or condition the effect of constituency opinion and needs when explaining legislator behavior. The primary mechanism assumed to be driving responsiveness is the electoral connection. When that electoral connection is tighter–for example, when issues become more salient–legislators should be more responsiveness. When the electoral connection is looser–for example, when knowledge of statehouse activity by constituents is absent–legislators should be less responsive to their constituents, and more affected by interests or their own characteristics like ideology. Heterogeneity in these effects will also be considered, as detailed below.
3.1 Heterogeneity by Legislator Party
Do Republican and Democratic legislatures represent their constituents in symmetric fashion? If so, the image of the parties as equally extreme relative to a moderate public and trading off ``leapfrog representation’’ (Bafumi and Herron 2010) may not be correct. Most scholars using national evidence argue that the Republican party is not a mirror image of the Democratic party, and is much more prone to extremism (Hacker and Pierson 2015, Mann:2016). The quantitative evidence for this conclusion comes from studies of elites and public partisans. For example, McCarty, Poole, and Rosenthal (2013) cite NOMINATE scores to show that Congressional Republicans appear to be polarizing faster than Democrats, and Grossmann and Hopkins (2016) critically examine the internal activist coalitions that make Republicans more ideologically oriented than Democrats. Ellis and Stimson (2009), Ellis:2012 show that Republicans have an advantage in representing the public’s symbolic conservatism which nevertheless conflicts with their weakness in representing their more substantive issue-based liberalism. Lelkes and Sniderman (2016) cites the greater ideological awareness by Republican partisans as a source of the party’s ability to resist positions broadly popular with the public as a whole.
But this evidence is overwhelmingly at the national level. Partisan asymmetry in representation is very much an open empirical question. Shor (2020) complicates the picture by showing heterogeneity of asymmetry geographically for state legislatures, with Democrats polarizing faster in the South and West, and Republicans in the Midwest and Northeast. J. R. Lax and Phillips (2012) find that states controlled by both parties overshoot the moderate public in roughly symmetric fashion.
At the legislator level, Broockman and Skovron (2017) shows evidence that state legislators systematically misperceive the opinion of their districts (measured with district-level MRP). A party asymmetry exists in that Republicans misperceive district conservatism more than Democrats, a difference partly driven by the greater political activism of district Republicans in contacting and lobbying the legislators. If true, we should expect to see lower responsiveness and congruence by Republicans than by Democrats, something that I will test.
3.2 Heterogeneity by Constituent Knowledge
Do the heterogenous presence of informational resources (like state capitol reporters) do so as well?
3.3 Heterogeneity by Salience
Does policy salience amplify constituency influence? J. R. Lax and Phillips (2009), Lax:2012 show that state policy on more salient issues like abortion and same-sex marriage is more responsive than those on less salient issues. Similarly, legislators should feel more pressured to follow district opinion on policies where there is more information.
Health policy, especially components of ACA implementation like Medicaid expansion, is a similarly fairly salient policy in general (J. Pacheco 2011, Pacheco:2016). Nevertheless, within health policy, there are big differences in information that could be relevant. Medicaid expansion or work requirements is likely to be highly salient, and our expectations would be for higher responsiveness and congruence in that policy area. Medicaid payment reforms, provider network requirements, or electronic medical records implementation are far more technical and hidden, and likely to activate particularized interests far more than public opinion. The disaggregation of the policy topics will be crucial to test this causal pathway (see below).
3.4 Heterogeneity by Electoral Environment
In a second paper, I aim to see the degree to which legislator votes on especially salient issues affect the electoral fortunes of state legislative candidates, as they did to some degree in Congress in 2010 (Nyhan et al. 2012). No such study exists at the state legislative levels. If evidence of such electoral effects exists, then some predictions can be made as to the outer constraints on legislator votes undermining the ACA at the state level.
The constituency to whom legislators are responsive has long been a topic of interest to scholars (Fenno 1978). Is it the primary election constituency (Brady, Han, and Pope 2007; A. B. Hall and Snyder Jr 2013; A. B. Hall 2015) or the general election constituency (Canes-Wrone, Brady, and Cogan 2002; Ansolabehere, Snyder, and Stewart 2001)? All extant research focuses on Congress, and I aim to extend that to the state legislative context. I will conduct two surveys and ask questions around the time of the primary as well as the general election to try to tease out the information that constituents might have regarding their candidates.
3.5 Heterogeneity by State Variation in Issue Ownership
Scholars have found that health care is an issue owned by Democrats [Petrocik (1996); Petrocik:2003; Egan:2013], which is to say that voters identify Democrats as the party that overwhelmingly prioritizes health care. Yet the evidence for this finding comes exclusively from national survey data. I will investigate the degree to which issue ownership varies by state, since in many states single party domination might amplify the reputation of the ruling party and the undermine or muddy the reputation of the dominated party. If true, this should condition the relationship between issue ownership and responsiveness and strongly qualify the Egan (2013) finding that congressional responsiveness is moderated by issue ownership.
3.6 Heterogeneity by Institutional Difference
J. R. Lax and Phillips (2009); J. R. Lax and Phillips (2012) and Lewis and Jacobsmeier (2017) show that institutional differences across states, specifically varying degrees of professionalization and the presence of direct democracy, affects the representational relationship between state opinion and state policy. I can see whether these findings hold at the microlevel. One such difference are campaign finance rules. In 2010, the Citizens United Supreme Court decision struck down restrictions on independent expenditures during election campaigns for both federal and state elections. In some states, these restrictions did not exist, while they did in others. The difference-in-differences estimate for this change can be interpreted to explicate how much less representative are legislator decisions when unrestricted campaign expenditures are allowed when they used to be restricted. Another such decision is Janus v. AFSCME (June 2018) which will make it much harder for public sector unions in many states to engage in political activity (Feigenbaum, Hertel-Fernandez, and Williamson 2018). But since a number of states already restricted these unions in similar ways (for example, Wisconsin), we can take advantage of the natural experiment afforded by the exogenous shock of the decision to compare the differences in the trends of the treated'' states (those whose unions were legally unobstructed) to those of thecontrol’’ states (those whose unions were legally bound).
3.7 Heterogeneity by Interest Types
Hertel-Fernandez, Skocpol, and Lynch (2016) discuss interest group conflict as the source of the heterogeneity in state pickup of Medicaid expansion. One way to test this at the district level is to disaggregate bills by lobbying interest, assumed and measured. For example, bills that attract professional lobbies (hospital, insurance, and physician associations) are likely to be those that affect the material interests of the participants. Alternatively, physician and specialty density can be used as a measure of presumed material interests. On the other hand, physician preferences are likely to be heterogeneous on the more ideologically salient bills, where liberal and conservative physicians divide in their mobilization and lobbying of their local representatives. Some bills attract exclusively economic interests in lobbying, while others attract ideological forces.