Public opinion and political parties are dividing across states, but is public policy following these differences, with conservative publics and majority Republican parties enacting more conservative policies? Christopher Warshaw finds shifts in state public opinion are reflected in policy, but not always through majority parties. Mark Richardson finds that majority state parties still need to win over centrist legislators and governors to pass charter school and abortion policies. Both find signs of life in statehouse democracy.

The Niskanen Center’s Political Research Digest features up-and-coming researchers delivering fresh insights on the big trends driving American politics today. Get beyond punditry to data-driven understanding of today’s Washington with host and political scientist Matt Grossmann. Each 15-minute episode covers two new cutting-edge studies and interviews two researchers.


You can also subscribe to the Political Research Digest on iTunes here.


Grossmann: This week on Political Research Digest, are red and blue states enacting red and blue policies? For the Niskanen Center, I’m Matt Grossmann.

Public opinion and political parties are diving across states, but is public policy following these differences with conservative public’s and majority Republican parties enacting more conservative policies? New research finds that shifts in state public opinion are reflected in policy, but not always through party majorities.

I talked to Christopher Warshaw, of George Washington University, about his comprehensive new study with Devin Caughey, published in the American Political Science Review, called “Policy Preferences and Policy Change: Dynamic Responsiveness in the American States, 1936-2014.”

But majority state parties may need to win over centrist legislators and governors to pass new policy. I also talked to Mark Richardson, of James Madison University, about his new research with Joshua Clinton, published in the Journal of Public Policy, “Law Making in American Legislatures: An Empirical Investigation,” showing their power.

Americans often think politicians ignore their opinions, but Christoper Warshaw says public policy does respond to public preferences.

Warshaw: There is democratic responsiveness in the American state. You might think that’s an obvious finding, but in previous findings on representation in the American state have all been based on a small set of policies or a single snapshot in time. What our study tried to do, we tried to develop an overall index of how liberal state policy is, in a couple of specific substantive domains, as well as how liberal public opinion is. This enables us to use panel evidence to say, does state policy respond to changes in public option? At the end of the day I think evaluating responsiveness is one of the crucial questions for establishing the character of democracy both in the United States as well as elsewhere in the world.

Grossmann: Warshaw and Caughey find two routes to public influence, through party change and through incumbents shifting their views.

Warshaw: What we focused on our study is two general ways that public opinion could affect policy. One is where when you get more liberal you tend to elect, you’re more likely to elect, Democrats. That’s sort of the way representation happens. The other is within party, even if the party that controls government doesn’t change, politicians or elected officials are responsive of margin of public opinion.

Of course you could have both of those things happening at once. In fact, what we find is about half of the effect of public opinion does come from partisan selection, but about half seems to happen through incumbent elected officials changing their position.

Grossmann: But public opinion shifts don’t always change the party in power.

Warshaw: The partisan control of government is almost uncorrelated with changes in public option in the states. We look at Massachusetts, Maryland, very liberal states, have Republican governors. What that shows, I think, is that in large part the parties are actually, unlike the federal level where they’re sort of constrained by the national platform, at the state level parties are able to reposition themselves around the voter. As a state gets more liberal, like Massachusetts we see somebody like Charlie Baker, a very very liberal Republican.

Grossmann: They use big composite measure of public opinion and public policy across issues and time.

Warshaw: We’ve aggregated about 150 state policies over the past 80 years or so, to develop an index of how liberal each state is in each year over the past 80 years. We’ve also aggregated all of the public opinion surveys, roughly speaking, that are publicly available. We aggregated all the issue questions in those surveys to estimate how liberal or conservative the public is on economic and social issues.

This enables us, to at the very least, get the temporal order right. We can say, if the public gets more liberal, in sort of year p, in year p.1 does state policy follow.

Grossmann: They separate opinion and policy into social and economic spheres, but find state policy usually varies along one left right dimension.

Warshaw: We found for policies that state policy is mostly one dimensional. You get only modest gains from breaking policy into different domains. Whereas for public opinion, at least in the mid-20th century, public opinion was strongly two dimensional. The states that were most liberal in economic issues in many cases were Southern states. They were obviously very conservative on social issues and racial issues. Thinking about public opinion as sort of a one dimensional measure didn’t really seem to make a lot of sense.

Grossmann: Warshaw says public opinion has a greater impact on social issues.

Warshaw: We found that actually public opinion seems to have a larger effect on social issues than it does on economic issues. That could be because any number of reasons explain that. Economists might say that, well an economic issue, states are more constrained by pervasive bottom dynamics, or the sense of economic competition. I think the public opinion person might say, well, the public’s views on social issues are more coherent than they are on economic issues.

Grossmann: And they found some evidence that responsiveness is increasing over time.

Warshaw: We find that responsiveness seems to be stronger today than it was in the mid-20th century where perhaps we find little to no evidence of dynamic response super early in the time period.

Grossmann: Sometimes public opinion is thermostatic, with liberal policies leading to more conservative views. Warshaw says the big state differences arise only with sustained left or right movement.

Warshaw: In terms of looking at responsiveness, even though there are thermostatic shifts, like if Democrats take over government, the public’s probably going to get a little bit more conservative in the next couple years. You still have these long-term changes, where Vermont is getting much more liberal over time. Would it have gotten even more liberal if there’d been lots of Republican governors in the last 20 years? More than there have been. Probably. They still do have these long-term shifts in public opinion.

Grossmann: They find that states liberalized economic policy through the 1970s, but have since continued to liberalize social issue policy alongside their public’s.

Warshaw: We do show that our economic policy gets more liberal until about the 1970s, when it basically stops. Until the early 70’s, around 1970, economic policy pretty steeply liberalizes in the states, and then it levels off. Our measure does actually capture many continuous policies, it’s not just dichotomous policies. We have welfare benefits depending on a variety of different things. We try to those, we try to normalize, we didn’t want to include spending that just reflects how wealthy the state is. We tried to make it comparable across states, such as dichotomous policies. Whereas on social policies we do feel like state policies continued to get more liberal all the way through the time period.

Grossmann: Mark Richardson looks more specifically at how legislators and governors produce policy change, and says centrist legislators are still important.

Richardson: If you hope to achieve policy change, whether it be sponsoring a bill or getting it passed to become a law or trying to target specific elites, whether it be a governor or a member of the legislature, to achieve policy change that way. You have to think about all institutions and the separation of power in them. You can focus on more centrist legislators who are going to be on the edge of the coaliltion towards the middle of preferences in ideological distribution of legislators, and also thinking about getting support of the governors. You need to think about what laws those institutions are all making together and getting that broader coalition that you need.

Grossmann: Richardson and Clinton used measures of legislators ideologies in state’s policy outcomes.

Richardson: Measuring policy change, that’s been a real challenge that we I would say made progress on, but we don’t fully get policy preferences or member’s policy preferences on the same scale as policy change. That limits the strength of the claims we can make. What we did do is say we’ve got a measure of policy change, and thanks to Shor and McCarty, a couple people behind [inaudible 00:08:12] that measures the ideology of legislators across states. We can use those measures and look at relationships.

Grossmann: Along with the policy liberalism measure that Warshaw and Caughey created, Richardson and Clinton added two more specific policy outcomes on abortion and charter schools.

Richardson: The benefit in looking at the more fine grained polices, is that we can get more policy change. Looking at the general measure policy, liberalism still equates on 148 policies over eight decades and produces an aggregate measure of liberalism. That tends to be very stable across time, as you can imagine, because you’re looking at multiple policies. Looking and specific policies allows us to look at change in a specific segment of policies across time, and then actually look at say, how did we see the change in chamber median leading to changes in abortion policy specifically. We clearly didn’t think about how that can affect people’s specific access to say, contraception or access to certain medical procedures, or people’s access to charter schools in the other case.

That’s why we like looking at charter schools and abortion as two salient, important policy areas. It’s nice that abortion is more mature area to clearly map it to liberal conservative dimension that orients American politics generally, where the charter school is a newer policy that’s maybe not fully politicized so we can kind of see how that policy evolves over time.

Grossmann: They were trying to evaluate whether majority parties get to set the agenda and move policy toward their median member, as some say happens in Congress.

Richardson: If you’re looking at a bill going to becoming law, looking at the preference of the governor, we can’t ignore the governor and maybe her veto. If you’re gonna think about that process. Now, that’s not to say that you might find if you just looked a single chamber, more influence from institutional structures, like the Hastert Rule, that would give the majority party more control.

Maybe I would say majority party support in a single chamber, which that would suggest the Hastert Rule versus the majority members in the House, they have to support a bill to bring it to the floor, is not going to be sufficient for bill to actually become law. You have to get through the Senate and the Governor.

Grossmann: They first simulated what outcomes would like from each view of policy making.

Richardson: How to adjudicate between theories like this is the use of simulation. Both the theory with the negative agenda power for a majority party, also predicts an important role for the median in determining policy outcomes. Similarly, the models based more on filibuster pivots, veto pivots, and executive also have an important role for the median. Without the simulation we wouldn’t be able to say much, because the correlation between policy outcomes and the preference of the median members could be consistent with both of those models.

Grossmann: Richardson says median legislative preferences are more important than party control.

Richardson: It might be you don’t have to focus on getting change in party control, if you can get a move in the median member or the most centrist member. Also, you want to focus on obviously the governor too. If you’re looking to enact policy say, focused on, can you get the median member, the most centrist member, of the policy issue on your side? The center member on your side? Whether it be through electoral change or persuading that member that’s already in office, then you also want to, say maybe that center member is already on your side in the legislature, but you’ve got a governor that’s opposed and you want to try to focus on getting change in the governor though electoral change, or again through persuasion. Kind of focusing on those two people specially is going to be what you want to do.

Grossmann: He thinks the insights would work for congress, and for the states.

Richardson: The model was of course originally designed looking at Congress, not the states, but they have pretty similar institutions over all of them. A key difference between say Congress and the states is most states don’t have a filibuster, so in our data we only tracked five states as having some kind of filibuster or procedure, so that’s an important difference. Some states don’t have, say a veto override provision that only a simple majority is required, I should say a super majority, to over ride a veto, only a simple majority is required. So I think overall the model should translate pretty well.

Grossmann: The results also help explain why there’s so much stability in policy in the face of party change.

Richardson: A key insight from our findings would be that what matters is kind of that median member, that centrist member of the legislature, and the governor. You could, presumably, and certainly very probably in the south, without having a change in party control, but if the preferences of the median member are not changing much, given the change in party control, that we see through replacement you’re getting very small changes in the median. Even though you have a change in party control at the legislature. We wouldn’t expect there to be very much policy change at all.

Grossmann: Richardson found the most change in charter schools policy.

Richardson: If we were ranking stability, charter school policy in our data is the least stable. Abortion policy more stable. Then the [inaudible 00:13:35], as you mentioned, is the most stable. I think if we look at, say for instance, charter schools, what we see interestingly in that case, and this goes back to some of what I mentioned in terms of policy change being important at the beginning, and then how it changes over time.

Richardson: California adopts a more permissive charter school policy on initial adoption, and we would say that now is a more conservative position in that it allows charter schools to have less oversight from say the state, be less bound by state law, things like that. That would be a more conservative position in our accounting. So California adopts a more liberal position than Wyoming, excuse me, adopts a more conservative position than Wyoming, initially, and then over time as it becomes politicized, we see California’s measure becoming more liberal, i.e. more restrictive over charter schools in Wyoming becoming more conservative. But at the end of our data, our data begins in 97′, ends in 2014, we still see the charter schools, more permissive of charter schools, i.e. more conservative charter school policy than Wyoming.

Grossmann: Warshaw says there’s still a lot of room for policy differences across states, but that public opinion is getting increasingly nationalized and sorted by party.

Warshaw: I think that on the public opinion side, certainly politics is nationalizing. I think that increasingly the main difference in public opinion across states is simply the percentage of Democrats and Republicans in each state. If you try to model the liberalism of U.S. state public now it’s really all about this percentage that are Democrats and Republicans. It certainly wasn’t true 50 years ago, right, where you have huge within party differences across states on different issues. Now, within party, basically Democrats hold roughly the same beliefs whether you live in Wyoming, West Virginia, or Massachusets.

In terms of the policy and remaining differences, I think what we found in our study is that there are large and really durable differences across states in terms of policies. In particular if you point out the southern states, they’re just much much more conservative when the rest of the country, it’s definitely relative to public opinion in that state.

Grossmann: Richardson says, as a result we’re likely to see polarized policy changes ahead.

Richardson: The effective increase in polarization suggests that we should see policy moving away from these centrist outcomes. If you look over time at the McCarty data you’ll see that over time the chamber median are moving farther away from the minority party median and closer to the majority party median. Polarization is causing party’s to be more similar within state legislatures, and also causing that median member to be more conservative. That suggests we’re going to see more extreme policy.

Grossmann: So far, Warshaw has not found much effect of institutional differences across states, but still wants to find out whether any reforms can make a difference in matching opinion with policy.

Warshaw: To me, the most interesting question to hear affect our institutions. The thing that I think about when I’m walking my dog, what are the institutions that might encourage responsiveness, how do we measure the effective institution, in order to figure out whether we should … how we can improve the character of American democracy or democracy worldwide. We have to have a sense that some institutions are better worth their reputation than others, and I think that in that sense it’s still very early days in the literature.

Grossmann: Both Richardson and Warshaw are mostly optimistic about their results.

Richardson: Normatively, we want to see public policy mapping back into the public’s preferences, and ideally kind of the average member of the public would get their policy views, would have the policy views reflected in policy. But I think certainly the model for having policy be more responsive to the median should result in policy that is presumably more centrist.

Warshaw: A basic metric for democracy is, do the view so of the public affect the policies the governor produces. Certainly if the answer to that is no, it calls into question whether it’s a democracy I think, or whether it’s majoritarian democracy. But if the answer to that is yes, that doesn’t necessarily prove that it’s a perfect democracy, but it at least provides some evidence that the public’s voice matters for governmental decisions.

Grossmann:  There’s a lot more to learn. Political Research Digest is available biweekly from the Niskanen Center, and on iTunes. I’m your host Matt Grossmann. Thanks to Christoper Warshaw and Mark Richardson for joining me. Join us next time to find out whether whites vote on the basis of racial identity and anti-black stereotypes.