Do rich white Americans always get their way in policymaking? Or does it matter who is in charge? Agustin Markarian finds that different groups see their policy preferences better represented depending on which party is in power. White Americans get what they want more under Republican control—and not only because white voters are mostly Republicans. In the Senate, Republicans with higher Black populations also fail to represent their views on their congressional votes. Whether it comes to policy outcomes or legislative voting, different groups win under different parties.
Transcript
Matt Grossmann: Different groups win policy under Democrats and Republicans, this week on the Science of Politics. For the Niskanen Center, I’m Matt Grossmann. Who gets their way in American policymaking? Do only the rich or only whites win policy that they support? It may not be that simple. It matters which party is in power, with different groups’ opinions represented under Democratic and Republican administrations.
This week, I talked to Agustin Markarian of Loyola Chicago about his new American Political Science Review article with Jacob Hacker, Mackenzie Lockhart, and Zoltan Hajnal, Race, Responsiveness and Representation in US Lawmaking. He finds that white voters get what they want more under Republican control, and not only because they are mostly Republicans. In the Senate, Republicans with higher Black populations also fail to represent their views on their congressional votes. The results challenge some prior findings, that class is the main line of representational failure and that the rich get their way no matter who is in power. But the results also suggest that each party’s electoral group constituency is not ignored, they get their way more when their party wins. I think you’ll enjoy our conversation.
So tell us about the findings and takeaways from your new paper on race, responsiveness and representation in the US.
Agustin Markarian: Sure. So this is a project that we started five years ago, it’s been long in the making. It was a significant data collection effort. We were hoping to be able to do a lot with it. And also, knowing the importance of the topic, us, the reviewers, the editors at APSR, all wanted to get it right. And we see it as one of the most broad and robust studies on race and policy responsiveness and policy representation that has been conducted. We know that it builds and we think that it builds on a really long tradition that has mostly focused on class and income, but race is also extremely important in the United States. Some might say it’s one of the most or the most important descriptive identity that shapes power, that shapes elections. And so, we wanted to know, does policy match what some groups want more than others? There’s obviously important questions here of influence. But also, we thought about this as just who wins on policy, who gets what they want from the US government, and who gets it more?
And so, what we end up finding, I think, to some folks might be pretty surprising. We find that small aggregate gaps in policy responsiveness across racial groups. We actually don’t find that communities of color lose on policy a lot more than white Americans do over the period that we look at, which starts in 2006 and goes up to 2022. But we do find that there’s a lot of variation based on party control, particularly party control of the presidency and the Senate. And racial and ethnic minority groups end up losing a lot more on policy than white Americans when Republicans control the presidency or the Senate, and we see reverse effects to some degree, particularly some advantages for Black Americans when Democrats are in control of those two institutions.
Matt Grossmann: So the first and major analysis is looking at Americans’ responses to survey questions, which are designed to get opinions on stuff that is before Congress, or at least being talked about in policymaking circles. And as you said, you find that overall, racial differences in responsiveness are low, but that white voters do get what they want under Republican control. So how should we interpret those together, given that, as you said, we’re talking about one and a half administrations maybe in terms of the evidence that we have, and maybe the more important finding is the null, that overall, when we get enough data, we just don’t see a whole lot of differences. How would you respond to that?
Agustin Markarian: Yeah, no. What we actually do see is we see really big gaps under Republicans, and we see maybe some small advantages for communities of color, particularly Black Americans, not so much for Latinos and Asian Americans, under the Democratic administrations that we look at, which are the Obama administration, which lasted eight years, and two years of the Biden administration. The gaps are larger under the Obama administration in favor of Black Americans than under the Biden administration. That might say something important. The gaps are pretty consistent when it comes to Republicans under the Bush administration and the Trump administration.
But one of the important things here is that we have more years of Democratic control, of Democratic presidencies, but also of Democratic Senates, during the period that we look at than years of the Republican Senate. And so, what we believe ends up happening is that small advantages for Black Americans and maybe for Latino and Asian Americans vis-a-vis whites under Democratic control of government ends up washing out pretty significant, pretty large disadvantages for these same communities under Republican control. And I think what it really says is that who governs, who controls these institutions really matters. And it’s some of the maybe more cynical takes on party politics, that the two parties are the same, or they’re the same for racial and ethnic minority groups, doesn’t really hold in the data that we’re seeing.
And it ends up washing out maybe that the aggregate analysis is really important if we assume that Democrats are going to be in power about one and a half times as much as Republican governments are going to be. But if that’s not the case, if there’s equal distribution of party control moving forward, or if Republicans govern for longer, then those aggregate really muted gaps might go away. We might actually see larger gaps on aggregate, if that’s what happens.
Matt Grossmann: So we’ve been talking about descriptive relationships. I know that people are also interested in the causal question. But even at the descriptive level, I think people are interested in, is this party more than race, and you’re able to show that the party differences do explain part of these relationships. In other words, Republicans get more of what they want under Republican control and Democrats under Democratic control, but that it doesn’t fully explain the racial effect. But I guess I wanted to just take a step back and say, aren’t those partisan differences the key story, if we’re going to be talking about who gets what they want, the partisan differences are bigger on a lot of questions, and the effects are not just which part of the racial gap that they explain, but also independently of race, what they explain as well?
Agustin Markarian: Yeah. So we find that partisanship explains much, but not all of the racial gaps that we see. When it comes to Democratic control of government, when Democrats control the presidency or the Senate, if we control for respondents’ partisanship, the gaps pretty much disappear, we no longer really see an advantage for Black Americans relative to white Americans under Democratic control. Under Republican control, even when we control for respondents’ partisanship, racial gaps still remain, particularly for Latino and Asian Americans, but for Black Americans as well. And this is a hard thing to interpret a little bit,
Because one might think about partisanship as a confounder. So okay, we’re controlling for this confounding variable that’s accounting for the relationship. But we could also maybe think about partisanship as a mediator. People’s race shapes what party they identify with, and if that is the case, if people’s race is shaping which party they choose to identify with, then by controlling for partisanship, we’re in some ways controlling for the direct effects of race. Race operates partially through partisanship and affecting policy responsiveness, and we can’t really tease these things out with the data that we have. It’s a really complicated causal story here. Is partisanship a confounder? Is it a mediator? Is it both? And so, that is a hard thing to figure out when we’re thinking about what really matters here. Is it really race that matters? Is it really partisanship that matters? Does race matter through partisanship? Those are questions that we still have that we’d love to explore in future research. We don’t think this data is the best equipped to answer those questions.
Matt Grossmann: I guess independent of whether it explains the racial relationship though, I guess it’s an important finding overall that partisans now do differ on issue questions and that Democrats get more what they want when Democrats are in charge. It’s not a surprising one. But how would you weigh that relative to this literature that you’ve been building on race and income and these other effects, given that we have partisan elections, partisan governance, and more polarization on issue questions now, maybe the most important thing that’s happened is the obvious one, which is just that we have partisan representation.
Agustin Markarian: Yeah. I think there’s very clear evidence here that parties are representing their core voters the best, that they are maybe not trying to represent everybody equally or represent the median voter, per se, or appeal to cross-partisans or trying to appeal to their own voters, at least in the policies that they pass. That seems pretty clear. This might be something that is maybe unique to this era of closely contested elections with high levels of polarization, but it is very much so the case, and it maybe contends against some of the more traditional models, whether that’s the model of median voter theory or other models that suggest that parties maybe are trying to appeal to broad coalitions.
Maybe we’ll have a chance to talk about this later, about some of our other work that we’re doing, but it really seems to be the case that the Democratic Party’s a little bit more egalitarian, the gaps are smaller, whether we’re looking at race or other identities. Whereas under Republican control, even when we control for partisanship, even when we account for who Republicans are, their policies are still much more representative of some identity groups that partisanship cannot fully account for.
Matt Grossmann: So you also have another type of analysis in this paper, where you move to explaining individual senators; votes based on their own constituents. And here, you find, again, a party effect, but you also are able to narrow it down to these states with large Black populations and high levels of white racial resentment. So tell us about how that fits into your story and how you interpret it.
Agustin Markarian: Yeah. So the gaps we see in policy responsiveness look pretty similar to the gaps that we see in what we call dyadic representation, whether peoples, members of Congress or senators or the member of the House votes on issues the way that somebody would want them to vote on it. If they support the policy, do they vote yes? If they oppose the policy, do they vote no? And perhaps surprisingly, on aggregate, what we actually find is that people of color win on dyadic representation, they get better dyadic representation than white Americans do on aggregate. We think a lot of this has to do with potentially racial segregation, the concentration of racial groups in particular districts, maybe majority, minority House districts. We see that effect only in the House, we don’t see that effect at all in the Senate.
We do think that dyadic representation maybe is a little less important than getting the policies that you want passed and the policies that you oppose blocked. Great, your member of Congress voted yes on your bill, but the bill didn’t pass. I would rather have the opposite happen to me. But we think that this is really useful to better understand maybe some of the things that are going on, some mechanisms at play, and to also just leverage heterogeneity across legislators’ characteristics and across Senate characteristics, state characteristics, House characteristics.
And this tells us some pretty important things, for example, that Black Americans actually receive worse representation in the Senate when they live in states with more Black Americans. This seems a little counterintuitive to what we think Democratic theory would suggest, that groups have more power the more people they have in a district. We see that that’s not the case for Black Americans particularly, not so much the case for Latinos either. But for Black Americans, it’s actually in the inverse direction that we might expect, and we think that this might be due to racial threat. A lot of these are states in the South, but not exclusively, the pattern holds outside the South as well.
And one of the final analysis that we engage in in the paper is seeing whether the average level of racial resentment in a state or in a House district predicts how big the white/Black gaps are in representation, and we find that it does. We find that it’s particularly predictive in the Senate and among Republicans in the House, that Black Americans that live in places that have higher levels of white racial resentment receive substantially worse representation, and once we account for this, we no longer see that population matters or is predictive of gaps at all. It really is these racial ideas, these feelings of racial resentment that some white Americans within those states harbor, that seems to be driving these patterns. And I just want to shout out some really good work by Kevin Morris on voter suppression that finds that state senators from more racially resentful places tend to support more voting restrictions than state senators from less racially resentful places.
Matt Grossmann: So on the one hand, this paper has hundreds of thousands of cases, if we think about it as does an individual get their preferred policy outcomes. On the other hand, we could think about it as having just over a hundred-some cases, if we’re thinking about it as we have policies and who gets what they want on those policy outcomes. And of course, we have to think about how do we get these cases, and it’s always impossible to figure out what the universe of potential things that the government could be doing that we want to consider.
Here, we have who decided to put them on the survey, regular basis, we have to put some on each time. And we came up with a list, that seems reasonable to me, but is going to be disproportionately things that move in a liberal direction, things that are more supported by Democrats than Republicans overall. And of course, anytime we do analyses that are your main findings of Republican versus Democratic control, then we’re splitting those cases in half and we’re even more reliant on what that agenda looks like. So talk to me about how you think about this agenda and what it means to have responsiveness on this agenda, especially when this or maybe any agenda that we come up with is going to be mostly new things that the government could do, which might appeal to liberals more than conservatives.
Agustin Markarian: Yeah. So using the CES, the Cooperative Election Study, as our data source, that comes with some trade-offs, huge samples, particularly on even numbered years. This means that we can actually analyze MCs’ behavior, members of Congress’ behavior, in a way that we wouldn’t be able to with a normal 2,000 person, let’s say, nationally representative survey. But that means that we have less questions, we have somewhere between five to 10 policy questions a year, which is much fewer than, say, the gold standard in this space, which is like Martin Gilens’ work. So that’s a trade-off there.
The other trade-off is the one that you bring up. We’re really affected by the choices made by the people that design the Cooperative Election Study. They select what policy questions they ask about. They give a lot of details on these questions, they say, “The US government wants to pass a bill that would do X, Y and Z. Do you support or oppose this policy?” But there’s a two-tier agenda-setting problem here. The first tier is the agenda set by Congress, that is not representative of the whole world of bills and policies that people have opinions on. And then, there’s also a second tier of agenda setting, which is the choices made by the people that organize and conduct the Cooperative Election Study. And they might have particular biases and be particularly interested in some questions more than others, those might be more liberal questions that aim to change the status quo. I think this is really important if you’re looking at overall policy responsiveness. There’s really good work out there that looks at is policy responsive to the public at all.
I’m a little less worried about these biases affecting our analysis of gaps. Let’s say we had a question pool that had more conservative policies on it. Well, my suspicion is that we would see similar types of gaps on those questions, across racial groups or across other types of groups, that gaps would not be too affected by these choices. I’m not saying that they’re not. I’m saying I’m a little less worried about it than if we’re analyzing the aggregate effects of public opinion on whether policies pass or not. But still a big question and something we try to address in the paper. It’s an issue that I think all of this work is going to have, to some degree or another.
Matt Grossmann: So one difference from some other universes is that actually about half of these end up succeeding, which is higher than some others, suggests that maybe these were the ones that are most in contention for succeeding. But we also have differences between the parties in particular, also between the public and the outcomes. That is the public usually favors these proposals, even though, in that sense, only half pass, and Democrats are more favorable than Republicans. So how do you think about the responsiveness you are investigating in comparison to just the basic status quo bias issue, that the public wants more to get done than ends up getting done? And should we think about or do you think about my success in not being in favor of something and it not happening as the same or different from me wanting something to happen and it happening?
Agustin Markarian: Yeah. So it’s the theoretical question, and then I want to make sure I’m just not giving the empirical answer. So I think that’s a really good point, there’s a status quo bias that’s likely to favor Republican preferences, more conservative preferences. That’s the bias of the system, particularly a system with so many veto points with super majoritarian thresholds in the Senate. And Republicans are more likely to be white, or white people are more likely to be Republican, however you want to think about it. And so, they might be favored by this status quo bias that we see in the system, that maybe, thinking back to the previous question, maybe is not fully reflected in the agenda set out by the CES, the questions that they ask.
That’s a hard thing to try to fully understand. What we do empirically to try to kind of get at this is we re-run our analysis looking only at questions that passed, and then looking only at questions at policies that failed. Do the gaps still remain when we split our sample this way? And we find that they do. The patterns look pretty consistent whether we’re talking about the aggregate patterns or the partisan control patterns. Of course, as you’ve mentioned, we’re getting to a pretty small n of questions here, but we find that the findings are still at least substantively robust, even if sometimes they fall out of statistical significance.
Matt Grossmann: So as you mentioned, one big other universe of cases that’s been worked on a lot in a similar kind of analysis of opinion to policy outcomes is the Martin Gilens data set, which aggregates a whole lot of different polls and outcomes rather than these, and before many of these were on this Cooperative Election Study. But there are some differences, not just in race as important has been studied less, but also on the basic findings in the Gilens data. That is you don’t seem to find very big income-based differences in opinion, and you don’t find huge differences in responsiveness based on income. So how do you see that in relationship to previous work?
Agustin Markarian: Yeah. So this is starting to preview one of our papers that’s currently under review, that really expands this work that here focuses on race, expands it to look at a whole range of identities, income, religion, education. Do we see gaps based on these other identities, and do we see gaps on these identities that depend on party control of government? And what we find is somewhat similar to what we’re finding on race, that income gaps, we’re not finding them on aggregate, we’re not finding that the rich win more. The rich here means making more than 150k a year. We’re constrained by the data. But we don’t find that those folks get what they want on policy on aggregate much more than lower income people.
This is in contradiction to what Gilens finds. We do find effects based on party control. We do find that the rich get more what they want under Republican governments. But again, still not fully aligning. There’s a lot of potential reasons why we find different results than Gilens. One of them might be, like you said, we’re looking at a different set of data, these CES questions, that are more limited maybe in scope, more specific in their salience, where Gilens looks at a broader set of questions over a maybe longer period of time. There’s some differences in the analysis. We’re looking at individual level, he’s looking at policy question level.
But we’re also looking at a different era. Gilens’ data ends in around 2004, I believe, and ours starts in 2006, there’s actually no overlap. And the parties have changed, the politics has changed. And if we think a little bit, for example, of just party changes, the Democratic Party of Bill Clinton is pretty different to the Democratic Party of Joe Biden. One pushed for major welfare reforms, the other one pushed for some pretty robust industrial policy. And so, maybe there’s just differences in what Democrats are advancing and how well they’re representing lower income workers that is accounting for these differences in our findings. Maybe it’s a funky methodological question, a question of it’s an issue of the questions that we’re looking at. I tend to lean to the former, but I can’t be sure.
Matt Grossmann: So we’ve been careful to talk about responsiveness and descriptive relationships. But of course, people are interested in the extent to which this public opinion or opinion among different groups actually causes differences in policy outcomes. And just as a first cut, I guess, this paper made me more skeptical. That is overall public support doesn’t seem to have a huge relationship with whether we get these changes. Some of that relationship could be reverse causality, that people are in favor of things that are more likely to pass. There could, of course, be things from political party and interest group behavior rather than these policymakers directly responding to the public, as I know you know and are investigating. And even this big finding that you all have, that responsiveness is different based on different parties and government, could be reframed as if we’re looking at what policy passes, the main thing we need to know is which party is in control, rather than how does the public feel about something. So how do you think about responsiveness, given that if we’re really interested in what drives the policy process, this might not be all that important?
Agustin Markarian: Yeah, no, that’s a great question. And it’s very possible that what we’re finding is some sort of spurious relationship, some indirect effect of, let’s say, interest group lobbying, or there’s a completely alternative explanation here for what we find that’s not about parties intentionally trying to represent some of these constituencies. And we use this word responsiveness, which suggests that they are responding to the interests of particular groups, and we can’t be 100% sure that that’s actually what is going on. I think there’s some evidence, like Dan Butler has some really good experimental work showing that if you present legislators with information on constituent preferences, they actually do update their beliefs, and potentially even how they vote on legislation. So I think that there’s some good suggestive evidence out there. We can’t really test that here in our paper.
I do want to say, I think that the why questions, as political scientists, we’re always so interested in the why, trying to understand why things work, how things work. But the what questions are also sometimes really important. Before we can try to understand why things happen, we sometimes need to know what is happening. And so, just to give an example of why I think these results are important irrespective of the mechanism driving them, I want policies that I like to pass. Whether they pass because I like them or whether they pass for some alternative reason, I’m still happy that policies that I like pass and policies that I don’t like fail.
I’m a big fan, I’m a big supporter of universal childcare. If the Trump administration were to somehow miraculously pass universal childcare, I would be happy. They probably wouldn’t pass it because people like me support universal childcare. They might pass it because a coalition of business interests and social conservatives worried about population decline support it. But I still get a policy that I like put into the books and I’m still happy. I’d prefer it if it was because I wanted it, but only marginally. And so, I think the findings that we find about who wins and who loses, I think that’s another way of thinking about it, not who’s policy-responsive too, but who wins and who loses more on policy issues, I still think those findings are extremely important to understand really who’s winning and who’s losing more in American politics.
Matt Grossmann: So as you mentioned, you’re now moving to other kinds of gaps, by income, by religion, by gender, by education, in addition to race, and looking at how those differ based on Democratic and Republican control. And again, I think descriptive findings are interesting, but even interpreting those, one interpretation is something comes on the agenda, we get public opinion about it, and then we see what happens. But another interpretation might just be different groups in society have different relatively stable opinions, and they affiliate with parties on the basis of those opinions, and they’re in the right parties. That is, overall, the groups that are affiliated with each party tend to get more of what they want when those parties rule. How would you interpret each of those?
Agustin Markarian: Yeah. These mechanism questions, they’re so hard to tell, and it’s different roles for different studies. We don’t fully see our studies as being able to or aiming to fully answer these mechanism questions, but I still think you’re asking them because they’re important. Whether results are driven through direct/indirect effects or some form of reverse causality is going to be hard to tell.
I think some issues are more likely to be reverse causal. So let’s say we’re thinking about defense or foreign policy. Maybe groups don’t have or people don’t have very strong opinions on those issues. Some people, but most probably don’t. But other issues, like social and cultural issues, or maybe economic issues, I think that people are more likely to have real beliefs on, real thoughts about, that are independent from what the current set of elites are telling them to do.
A lot of Republicans opposed the One Big Beautiful Bill. A lot of Democrats opposed TARP, the bailout of the banks back in, I believe it was 2009 or 2010, and those bills still passed. The parties still passed bills that were against what their constituents wanted, and their constituents didn’t automatically change their opinion on those bills. So the reverse causal story, I think, is more likely to be the case on some issues or some policy domains than others. And what we actually find, both in this paper and in the one that is currently under review, is that the gaps that we’re seeing exist across policy domains, and they’re actually sometimes bigger on some of the more salient issues, on social and cultural issues, on economic issues. And maybe that’s because we expect parties to be more responsive to these particularly salient issues, to be more attuned with the interests of their constituents.
Then there’s this third possibility that you brought up about party sorting, there are people just joining parties that fit their opinion. I think that’s very likely to be a large part of the story. Is that responsiveness? Is that not? That’s a trickier question there, I can’t really fully answer that. But yeah, I think party sorting is definitely a big part of the story, and I would not want to try to rule that out at all.
Matt Grossmann: Yeah. I don’t think it’s necessarily a mechanism question even, it could just be an interpretation question. You’ve interpreted it through the legislative responsiveness literature. But let’s say you just did an analysis where you just looked at public opinion on 50 broad issues, and then which ones were most associated with the Republican Party’s platform position and the Democratic Party’s platform position. So there, we’re not specifically necessarily talking about legislation before Congress, but you probably would find a similar set of patterns, that some groups’ opinions are more associated with one party or the other, and that tends to be the party that the group has affiliated with. Would that be a different finding or would that be a similar kind of analysis to what you think you’re doing?
Agustin Markarian: I think those would be observationally equivalent, given the data we have. No, I think that’s 100% the case. If that is what’s going on, if this is perfect party sorting or some… Not perfect, but just that is what you describe is what’s happening, we would find the same results. I think maybe there’s some type of analysis that can be done over time to see, and we’re looking to do some of that work. We’re going to do work, for example, on accountability. Do people, do groups that don’t get what they want out of a party, or don’t get what they want out of their particular MC, do they switch parties, do they try to vote them out? Is there an accountability mechanism here for bad representation for people that are not representing their constituents very well? We’re trying to do some of that work, thinking about who can hold their legislators accountable, and I think that might help answer some of these questions.
There can only be responsiveness if there is accountability, to some degree. Otherwise, the whole mechanism breaks down and it just doesn’t work anymore. And so, if we find some evidence of accountability, particularly accountability imposed by certain groups, by certain core constituencies, I think it would say something about legislators responding to public opinion. If we don’t, then it might just be 100% a story of pure party sorting and very little accountability. And we can only get so far with some of this data and answering some of those questions, but for right now, what you hypothesize and what we find and the story we’re telling are very much observationally equivalent stories, I think. Maybe I’ll get back to you on that in a week after I’ve thought about it more.
Matt Grossmann: On the other hand, it could be an important finding that the groups that are associated with each party do tend to go with that party’s actions in government because we have all of these ideas about some sort of false belief that these groups have. In other words, we hear all the time that religious people are not well-represented by the Republican Party, or low education people are not well represented by the Republican Party, or Hispanic Americans shouldn’t be moving toward the Republican Party because they are not associated with the opinions of those groups. So are there any findings in this new paper that support that story, that actually, that groups don’t go with the obvious coalitions that you might expect, or the new coalitions that you’re seeing materialize in electoral politics?
Agustin Markarian: Yeah. So for example, in the new paper that we have that we’re working on right now, we find that Republicans best represent what we oftentimes would think as actually their core constituencies. There’s advantages, as we know from this paper, for white Americans relative to people of color, there’s advantages for men relative to women. There are very, very large advantages for evangelical Christians relative to all other religious groups, even if we’re talking about mainline Protestants. We think that’s a really interesting finding. But particularly large advantages for evangelical Christians relative to Jewish Americans and relative to atheist, agnostic, non-religious folks. Not really big advantages for less educated Americans, but we do see that the Democratic Party over-represents highly educated people relative to less educated people. And we’re doing this in a multi-variate analysis version. So all of these findings are controlling for… Let’s say the education findings, those are controlling for people’s incomes as well. We’re trying to tease out these individual intersecting identities.
And so, we do find traditionally what we think about as core party coalitions are being better represented by their parties. And there’s this interesting thing that happens once we control for party ID, which actually shows, for example, that Latinos are underrepresented compared to whites by Democrats once we control for how much Latinos support the Democratic Party. And that might tell us something about maybe why there was a big shift of Latinos towards Trump in 2024, maybe they just felt underrepresented, given their level of support. And so, this story that we’re discussing here, it’s just making me reflect a little bit on the prior question about controlling for party ID, about is this all party.
And so, if people are choosing to support parties, if they’re choosing to identify with parties that better represent them, then how are we even making sense of this question of controlling for party ID? Then party ID is definitely a mediator. It can no longer even be thought of as a confound, if people are actually selecting to parties, not because their parents were Democrats or their parents were Republicans or some sort of social learning mechanism that happens at a young age, but because of the policies that the parties are passing. And so then, well, we shouldn’t control for party ID at all, because that is very reverse causal of the policy outcomes. And so, these are really interesting questions, really complicated ones. We don’t know if we fully have evidence to explain them, but we do see that the parties are representing what I think we would consider to be the traditional groups that we associate with these parties.
Matt Grossmann: So I’ve mentioned several complications with the data, but I also think some of it is inherent in these questions that we’re trying to ask, or that you’re trying to ask, what is on the agenda, how do we assess people’s views on that, and how do we relate them to what’s happening in Congress? So I know you’re five years and counting on this type of analysis, so where do you see this type of research headed? Are there better things that are different types of data that might help to elucidate these questions of representation and responsiveness? And what would your ideal data look like?
Agustin Markarian: Yeah. With the data question, I think there’s always going to be some limits when you’re doing these type of observational studies where you’re asking survey respondents to give their opinion on a policy in a space of constrained information relative to legislators. Seth Hill has good work on this that we try to address in the paper. The information that you can give to somebody on a survey, even if you’re trying to give them a lot of information, it’s not going to be the same that’s in front of members of Congress that are being given analysis by different budgetary offices and interest groups. They’re existing in different information worlds, and that itself is an additional challenge. Even if we said we’re going to move beyond the Cooperative Election Study, do something like what Gilens did and look at much more years, we talked about some of those trade-offs, the agenda-setting threat’s still there. We’re not automatically getting rid of that. Congress is still setting an agenda that’s not fully representative, and surveys, whether it’s Gallup or Pew or whoever, are also choosing what questions to ask about or not. All of those are big issues.
I think another thing that I’ve been thinking a lot about is how much Congress is actually even doing on policy. All of this work is about issues in front of Congress. Congress, realistically, what, might pass three, four meaningful bills a year on a good year, maybe one or zero on a worse year, on a more lame duck session year. And so, I think some of the interesting work moving forward in this space is to look beyond really what Congress is doing and look at executive actions, rulemaking, some of those other processes within the executive branch, that people still have very strong opinions about, but that isn’t being captured by the questions in the CES, and oftentimes not even by questions being asked by others, that really has concentrated power in the presidency. And as we move to a world where the executive is much stronger, we need to start thinking about the questions that we’re asking on surveys, and the way we’re thinking about policy responsiveness is one that has a lot to do with what the president does, how the executive actions are being executed.
Matt Grossmann: So your paper also includes these individual level representation findings, which I think you’re also continuing, about House and Senate members. You mentioned that this might become less important if we just are going to get party line votes on legislation, then trying to tease out individual representation from your House or Senate member might be less important. On the other hand, you do advance this with some new models with a slightly different indicator of public opinion, because you’re looking at public opinion from specific groups. You do have findings that mirror some findings in this previous literature, like the effects of the size of the racial groups and racial threat. There are also other types of studies that look at public opinion versus local industry or some measure of local interest group organizing. So how do you think about your advance there relative to this category of research on what we call dyadic representation and where that’s headed from here?
Agustin Markarian: Yeah. So this guy, Matt Grossmann, has really good work on interest groups and their effects on how legislators are making their decisions, highly recommend that. I think that is part of it. And it’s thinking about what Gilens and Page do, where they look at interest groups and they look at public opinion at the same time, I think that’s fascinating work, and that is such a interesting way to think about this, because these two things can be correlated, they can diverge, they can work through each other. And I think there’s a lot of space to grab the data that we have here, merge it with some interesting interest group data on the same type of issues, and see if we can tease that out a little bit more.
But moving to the question of how much does this matter in an era of polarization, I don’t know. I don’t know what next year looks like when we move to greater levels of gerrymandering and this whole redistricting war that’s going on and we have increasingly less competitive districts, yeah, public opinion’s going to matter less. These institutional decisions, pretty wild institutional decisions, that we’re making right now are, I think, moving power away from the public and the public’s ability to influence legislators, and moving more to a world where legislators can kind of do whatever they want, because people are not going to have realistic choices. At least in this two-party, highly polarized, close contention, highly gerrymandered world, I think that looks worse. The power of people to influence what their members of Congress do gets worse.
Some hopeful signs that I’m seeing, and I’m really moving away from this data a little bit, thinking about what I’m seeing in the world, is that, for example, we do see some discharge petitions happening in Congress right now that are bipartisan, where Republican members are moving away from what their leadership is telling them to do, and there seems to be some level of individual thinking and individual decision-making. Party coalitions are in a period of transition potentially, they’re reformatting. The Republican Party particularly seems to be under contestation between traditional business interests and some of the new, maybe more populous sides of the MAGA movement. The Democratic Party is probably not in a better position at all. Maybe they’re even in a worse place to try to figure out who their voters are and who they’re trying to represent best. And maybe these things actually improve a little bit how responsive members of Congress are to the public.
All of this stuff, it pushes in opposite directions, and it’s really hard to say how much public opinion matters, particularly moving forward, in shaping legislators’ decisions. But those are some of the things crossing my mind right now and I’m giving some thought to without knowing necessarily how we can apply this data to better understanding that.
Matt Grossmann: So we’ve been talking about responsiveness and representation as good things, and you talk about their decline as a potential bad thing. But of course, the specific story you find here is that members with large Black populations and white people who are resentful of those Black populations are the ones responding to those particular characteristics of the district, so that doesn’t sound like a necessarily great story. And you might be able to tell a story that the kind of aggregate responsiveness you find is better than that, in the sense that it is nationally responsive to the party coalitions and doesn’t necessarily have these particularistic racial threat effects, which might be worse. So should we want this dyadic representation, and is the system that you’ve identified, where we get very different types of representation based on who’s in power, maybe been under-liked?
Agustin Markarian: Yeah. There’s been so much… The Voting Rights Act, I’m not going to talk… It was great for political power, it was great to get Black representation, it was great to make sure that people were not being disenfranchised. But it wasn’t really something that was considering really what ultimately I think should be the goal of a lot of movements to redistribute political power, which is making sure that people are getting what they want out of policy, Not just making sure that they have a member of Congress that can advocate for them or that can represent their interests, but making sure that, at the end of the day, policy is responding to hopefully everyone equally. That’s a really hard thing in democracy when you have high levels of opinion polarization across racial groups. If what Black Americans want is very different than what white Americans want, and white Americans are a significantly larger group in the country, it’s going to be a really hard thing to achieve to make sure that Black Americans have an equal voice and equal influence on policy.
I don’t know how we get there. I’m not going to offer policy solutions to deal with that. And given that that is kind of the world that we live in, that we don’t see very large aggregate gaps in policy responsiveness, I think that’s a good sign. But as we talked about, some of those caveats really are it depends on do we have this sort of distribution of party control that tilts towards the Democratic Party moving forward, and if we don’t, well, then these aggregate patterns all are… What we find is not really what we should predict to find if that doesn’t hold.
But it does say something about the ability of political parties, even within a two-party system, to advocate for minority interests and to advance minority interests in a pretty powerful way, and I think that that is a good thing, that Black Americans have enough influence in some ways that… I don’t want to contradict some of the amazing work by past scholars, but that does not necessarily fit into this idea of, let’s say, group capture, that says that Black Americans don’t have any influence over the Democratic Party because they have opinions that are too far from the median, and therefore both parties can safely ignore them. No, Black Americans seem to get what they want a lot out of the Democratic Party, particularly Obama’s Democratic Party, and I think that’s a good thing. And it seems to be the case for some other smaller groups as well.
Matt Grossmann: Anything else that you wanted to include or want to tout about what you’re working on next?
Agustin Markarian: Oh, always happy to tout more work. There’s so many things. So within this, what we’re doing here with this data, we have another really cool paper, one that I’m really excited about, that looks at differences between House outcomes and Senate outcomes and between House members and Senate members. And what we find is that the House is a better representative institution of minoritized groups. So racial and ethnic minorities, women, younger Americans, lower income Americans, they’re much more likely to get the policies that they want out of the House than out of the Senate. Some of this probably has to do with status quo bias in the Senate.
But what we also find is that it doesn’t really reflect when we look at just dyadic representation. Members of the House are not better representatives of this group than Senators are. Besides race, race is the one exception, but for all the other identities that we look at, we don’t see that members of the House are better representing minoritized groups than Senators are. What we see is that agenda control, particularly the filibuster in the Senate, is making it so that a lot of the issues that some of these more minoritized groups, whether it’s women or younger Americans or religious minorities, some of the policies that they prefer are never being voted on. They’re not passing because they’re just not getting a vote. We have some cool simulations in this paper showing that if we got rid of the filibuster, these groups would get much better representation in the Senate, they would be on par with the House, so really excited about that.
And then, I have some solo authored work that’s related, in many ways, that looks at one particular issue, gun policy, and that argues that we don’t see gun policies or gun control advance because the victims of gun violence are disproportionately people of color, and that only when mass shootings affect white communities, not when mass shootings affect communities of color, do states advance gun control legislation. And so, I have a paper on this, but I’m finishing up my book project on it, that really traces all of these other actors, not just state level policy, but traces how legislators respond in biased ways, how interest groups, including gun control interest groups, how their lobbying efforts are also biased based on who is being affected by gun violence. There’s biases in public responses as well, but not in electoral outcomes, and so I don’t think the public has a lot to do with it, I think it’s these elite level actors that are driving it. And so, hopefully, if things go well, that book will be done next year and under contract. We’ll see how that goes. These things are always slower than I hope.
Matt Grossmann: There’s a lot more to learn. The Science of Politics is available biweekly from the Niskanen Center, and I’m your host, Matt Grossmann. If you liked this discussion, here are the episodes you should check out next, all linked on our website. Does anyone speak for the poor in Congress? Does diversity in Congress translate into representation? How voters judge Congress. Why rising inequality doesn’t stimulate political action. And how racial realignment ignited the culture war. Thanks to Agustin Markarian for joining me. Please check out Race, Responsiveness and Representation in US Lawmaking, and then listen in next time.
