A Discussion on the “Kingmakers” of American Politics

American Legislatures are far more white and male than the population as a whole. Women make up 51% of the country’s population but only 25% of legislators are women. African Americans and Hispanic Americans are 13% and 17% of the population respectively, but only make up 9% and 5% of our legislatures. Moreover, according to a 2015 report from the National Conference of State Legislators, “since 2009, the advances of women and blacks in legislatures have stalled.”

That may have changed slightly in the aftermath of the 2018 midterms, which elected a record number of women and minority congresspersons. However, the fact remains that American legislatures do not reflect the populations which elect them.

Professor Michael G. Miller of Barnard College is working on one possible explanation why that might be the case: Party chairs at the county level don’t think that minority candidates can win elections. Along with two colleagues, Professor Miller is conducting a nationwide study of these party chairs to see how these local party mandarins penalize minority candidates and how much that affects whether people will make their first forays into electoral politics. Our Director of Operations, Clayton Becker, recently sat down with Professor Miller to ask him about his work. What follows is a lightly edited transcript of that interview.

Clayton Becker:

So first off, what is the book about and what brought you to this particular project?

Michael Miller:

The book we’re working on right now, it doesn’t have an actual title yet, but the title that we started with was “Kingmakers,” right? Because we became interested in this idea of local partisans making decisions that rippled out throughout the political system. So, both parties are organized down to the county level and there’s a county chairman in place in almost every county in America. And those people are responsible for going to people in their community and encouraging them to run for that–make that first political campaign. So we recognized that if the people, the chairs, the people who are responsible for encouraging citizens to become candidates, if they were putting their thumb on the scale in a way that even they weren’t in touch with, if they had latent bias, be it, you know, race or gender or something else, class, that we would get candidates that they prefer because those are the people that get encouraged.

So if they’re the only people, if it ends up then that the candidate pool is disproportionately white and male, then you’re going to have candidates who are disproportionately white and male, and victors who are disproportionately real white and male. And that goes all the way up into Congress and the Senate and even the Presidency. And so we set out to really capture the extent to which these chairs did harbor these biases and found that candidates of color are significantly–that the electoral chances of candidates of color are significantly downgraded by chairs of both parties, but we didn’t observe effects for women.

Clayton Becker:

So you’ve mentioned in the past that there aren’t really differentiating factors based on political affiliation, but are there differences in other respects?

Michael Miller:

The effects are completely immune to county context for Republicans. It’s a completely flat line. The effect does diminish for Democrats as a county becomes less white, but the point estimate is negative for the vast majority of American counties, something like 90%. I would have to look at the exact number. But suffice it to say that even for Democrats, this effect lingers regardless of the demographics of the county.

Clayton Becker:

And do you think that that’s just a latent bias effect or do you think it has more to do with concerns of electability as we always hear about?

Michael Miller:

So one of the things that once we got these results back, we started talking about them, and decided to develop this into a book instead of just one paper. So, we’ve been going all over the country and actually showing these results, to the party chairs and asking them to explain, and we mostly get a story about electability. And to the extent that chairs will, we’ll talk about this. And the story, will be something like, well, “I’m not sure if voters would…” It’s a straight up identity politics story. So it’s pretty difficult for them to say something like, “well, our county is overwhelmingly white. I just don’t know how voters would react to a black candidate.” But, we also have observed a dynamic in which we are having typically very frank and thoughtful and candid conversations with Southern chairs. And Northern chairs–Northern Republicans specifically are much more likely to say things to deny the importance of racism, to say things like they don’t see race. And so we have observed, qualitatively, a really interesting dichotomy between Northern and Southern Republicans.

Clayton Becker:

So politicians in the South in particular, are more open about why they’re choosing white candidates?

Michael Miller:

It’s not that, I would also not call county chairs “politicians.”

Clayton Becker:

Sure. Fair enough.

Michael Miller:

It’s that they are more prepared and thoughtful about the role of race in their community, so they’re, they’re aware of it. If you ask, a southern Republican chair, you know, how do you explain this finding that we got, you’re going to get a thoughtful answer that spans 50 years of history in the town and they will tell you that race has been an issue not only in the community but also in the Republican Party and that they are conscious of it and actively trying to fix it. Some of them expressed frustration about their own inability to do better. But in the North, we have been accused of in infusing the research with our own liberal professor agenda, things like that. We have not seen that in the South. And so it seems like it is a dynamic that is not prevalent in the South and actually really seems to be isolated around one particular state. That I can’t name.

Clayton Becker:

Fair enough haha. Demographically speaking, I know you said that it’s flat in terms of Republicans, but on the Democratic side, does it being say a majority-minority county, have any effects…like for Los Angeles County or Prince George’s County, Maryland?

Source: Doherty, David, Conor Dowling, and Michael G. Miller. 2019. “Do Party Chairs Think Women and Minority Candidates Can Win? Evidence from a Conjoint Experiment.” The Journal of Politics. Forthcoming. 

Michael Miller:

So this paper is on my website, the top paper on my website. If you go all the way down to, let’s see page 21, you can see exactly what the effects are for Democratic chairs. So right here is where there is no penalty. So, you’ll see that it has to be 30% black before the estimate for Democratic Chairs crosses zero. Well, there’s not very many counties in America that are 30% black. We have the figure in the paper somewhere; it’s something like 10% of American counties or something like that. But look at this line for Republicans, you could not devise a flatter curve. It’s literally flat.

Clayton Becker:

Absolutely. So, basically the takeaway from that then is the county needs to be overwhelmingly African American before these party chairs will select an African American candidate? Or before they even become slightly more likely to do so?

Michael Miller:

The outcome is who they think can win. So their judgment where they would not downgrade a black candidate–they would stop downgrading a black candidate at that 30% mark. Right.

Clayton Becker:

That’s pretty scary actually.

Michael Miller:

The experiment that we did in this paper is designed to recover their latent feelings. We didn’t ask them about their overt thoughts.

Clayton Becker:

That’s actually incredibly disheartening.

Michael Miller:

No downgrades for women though. So that’s the good news.

Clayton Becker:

What do you think is the defining factor there between minority candidates versus women in the downgrading the latency biases? Is it just like historical context or something else?

Michael Miller:

What our interviews reveal is that particularly in rural areas, what this comes down to is that you’re talking with people who just haven’t had a deep relationship with someone of a different race and that’s what the chairs themselves are telling us. So if you ask them to explain why they don’t think their voters would support a black candidates, they will say something like, “well, we just don’t have African American people here.” Which implies that there is no interpersonal connection, you don’t have that shared lived experience. It’s not true with women. When you observe women run and win you develop different thoughts about their potential candidacies, these chairs are also telling us pretty clearly that they believe that women are actually harder working candidates and so they are making a judgment based on what they’ve observed from women who have run. The point estimate for women is actually positive in our experience. It’s not statistically significant, but it’s positive, which implies that chairs are making a positive judgement about them.

Clayton Becker:

And do you think that’s reflected in the number of women candidates we saw in the last election cycle? Or do you think that’s just an outlier?

Michael Miller:

I’m not sure that people coming out of the woodwork to run for Congress has anything to do with our results because our people are mainly recruiting for a local races, things like Sheriff and County Commission and State Legislature. People who come out and run for Congress I think very often are not being recruited by those local people. Only some of them are.

Clayton Becker:

So when we talked a little bit earlier at the beginning about this idea that the disproportionate number of white male candidates are coming up because of this like local party chair, bias against them. What effect do you think that has, if any, on the national level races?

Michael Miller:

The effect is gradual. It’s a bench effect. So, where do you get your congressional candidates from? Usually state legislatures. You get your Senate candidates a lot of the time from the US House. And so if we’re putting white males on our city councils and our county commissions and our state legislatures, then that trickles up, was our theoretical expectation.

Clayton Becker:

And how has that worked out in the results?

Michael Miller:

The only place can observe any effect is right here in the experiment. And then from the rest of that it’s just kind of extrapolating that recruiting dynamic as a potential cause contributing to the overwhelmingly white male flavor of American legislatures.

Clayton Becker:

So, are lower—I don’t want to say lower status—but state legislatures and county boards, are those even more white and male then Congress or is it about the same?

Michael Miller:

Depends on the state. There’s actually a wide, wide range on that. But I think if you add up all state legislative seats in America then the demographics look pretty much pretty close to Congress. I think women are doing a little bit better these days after the last election, but they’re pretty close.

Clayton Becker:

So the last thing that I wanted to ask is, do you think there is a way to convince these party chairs or to change this dynamic, where we get more minority candidates, at the lower levels and build a larger bench for a legislative branch that actually looks like America does today?

Michael Miller:

I don’t know. That’s one of the things we’ve been asking chairs. Because we thought, well maybe just telling them about this is enough, educating them about what’s going on. What the chairs have told us is that, “look, it sounds great to go and recruit candidates of color. I don’t have any.” And that is a real practical problem in many, many rural American counties, particularly in the northern states. Even if they would make concerted efforts to diversify their candidate populations they just, they can’t find folks, they can’t identify folks in their community to run. And so the chairs, many of the chairs are telling us that they really do have a supply problem.

Clayton Becker:

So you’d say then that the base of the problem there, aside from the biases and latency problems and all of that, it’s just there aren’t enough people who want to run for Congress from those communities?

Michael Miller:

People who want to run for office run for office in general? Yeah. But the chairs go farther and say that there’s just not enough people. The literal count of racial minorities is so low in their counties that it’s hard for them to identify potential candidates. Now, from there we could extrapolate things about socioeconomics or whatever. “Are you only looking for certain kind of profession?” But that’s what they’re telling us.

An Interview with Professor Chiara Superti

Professor Chiara Superti is a Lecturer in Political Science and Director of the M.A. program in the Department of Political Science at Columbia University. Prior to Columbia, Chiara received her PhD in May 2015 from the Department of Government (GSAS) at Harvard University, affiliated with the Institute for Quantitative Social Sciences (IQSS). Her work focuses on a variety of topics such as unconventional voting, political corruption, immigrants’ political attitudes, and electoral politics in Southern Europe, Latin America, and Israel. After completing her doctoral work, Professor Superti was awarded the position of College Fellow in the Department of Government at Harvard University.

Associate Editor Inga Manticas recently spoke with Professor Superti about her research on unconventional voting patterns in Cuba and Spain. The following transcript, compiled by Manticas, has been edited for clarity.

 

Could you tell me a little bit about your most recent work on unconventional voting?

Recently I’ve been collaborating on a project on voting in Cuba. In this work, which I’m conducting with Ángela Fonseca from the Pontificia Universidad Javeriana and Jorge Domínguez from Harvard University, we are using data on “unconventional voting” as a way to measure different kinds of citizens’ attitudes toward the regime.

Many people in Cuba have been using blank voting, null voting, or what we call “selective voting” when they cast their ballots. We interpret these practices as a general sign of opposition, but of different types. In Cuba, for the national election, the number of candidates is the same as the number of seats in the assembly. So everyone gets elected. The ballot contains the option to choose “vote for everybody” (Voto Unido). What the government really wants is for the voters to check that box, so that they would automatically vote for everyone.

And so, for a very long time, it was very important for the government that everyone would follow this kind of voting (voto unido) and the election was a form of referendum of support for the ongoing government. In the past 20 years, however, people have increasingly voted only for some candidates out of the list rather than for everyone. This is what we call “selective voting”. The other types of voting practices that we include in the category of “unconventional voting” are: leaving the ballot sheet completely blank or purposefully nullifying it, for example by drawing a picture on it or writing an expressive statement. Together, all of these methods have reached around 20% of the votes, which is a significant amount.

 

Have there been difficulties in procuring any of this data? 

 Well, the data was public; it was on the official newspaper of the Cuban Communist Party and other public sources. The problem was that there was not a lot of fine-grained data, especially for blank and null voting where there is no data at the municipal or district level but only at the provincial level. This limited significantly the kind of analysis we could perform on that.

In the other study, instead, where we look at the candidates’ vote share, we ran a candidate-level analysis.  There are around 600 candidates…

The problem is whether or not we think that the data reported officially is valid. Some people may have doubts about the quality and credibility of the data. So we did some statistical analysis—there are statistical methods to test if there is any obvious sign of electoral manipulation—but we couldn’t find any clear sign of manipulation of the data.

 

Maybe this is outside of the scope of your research, but do you anticipate any major differences in voting or political behavior in general in Cuba because of the transition in their relationship with the United States?

Yeah, that’s interesting. I am not sure.  Among the researchers on these studies, I’m more of the voting side expert than the Cuba expert. But if the opening continues—and it’s unclear what the impact of President Trump’s election will be, exactly, on this — selective voting might increase a little bit.

There is some recent literature in political science that claims that authoritarian regimes are nowadays willing to accept a little bit more of dissent when that dissent provides some kind of useful information. And so some level of criticism is allowed. For instance, there is very interesting work done on China by some colleagues of mine (King, Pan and Roberts, 2013) that looks at online censorship in China. They found that what mainly tends to be censored online is not so much the criticism of regime but the call for any form of collective action.

In this sense, the selective voting is giving a little bit of information to the regime, about what candidates are doing better. The key element that we are missing in our work, because it is not easy to find and may be too early to seek, is to see whether there are any consequences of receiving fewer votes. For instance, are candidates who are getting less support, lower voting percentages, in the long run, being punished by the regime? Is the party saying, ‘you are not a good candidate, you are not going to run any more’? It’s too early to see if that’s true, but that would be evidence that the regime is letting go of a little bit of control to gather information and then use it.

 

So what do you see as your next steps in your research on voting?

At the moment I’m trying to understand better the determinants of unconventional voting and of protest more generally. What kind of experiences makes individuals more likely to use protest as a form of political expression? For example, I’m looking at the case of Israel, which is another area I work on, and I’m trying to understand whether or not exposure to violence affects political participation, and in particular behaviors like unconventional voting and political protest.

 

Why is it important for academics to be doing this kind of research?

Unconventional voting is understudied and under-theorized, because it’s not considered a widespread or significant phenomenon. Recently, political protest in general has, instead, become a much more studied phenomenon, although I think political scientists still don’t spend as much time on it compared to other social scientists. One important goal is to expand the understanding of this behavior, its determinants and effects. I think there is still a lot of work to do in political science on questions like: what kinds of protest forms have more impact? Are people and politicians paying attention to unconventional voting? Would it be worth it to have a “None of the Above” option on the ballot as an official option? Finally, is there some kind of backlash after political protest?

 

What advice would you give to undergraduates interested in going into a career in academia, especially in political science?

 Political scientists are currently focusing on using more advanced and rigorous research methods in terms of both qualitative and quantitative approaches, perhaps with more emphasis on the latter. That doesn’t mean that the discipline only uses statistics, but that this is something to consider or understand in advance if you are interested in pursuing an academic career. It might be useful to take some quantitative methods courses and try to explore if it’s an approach that you like or not.

In addition, it would be useful to take some advanced seminars—which many of you are already taking—to see some of the most recent research and understand what the field is doing. I believe that sometimes, students who have not been exposed enough to the most recent research, are not aware of what the field is really working on, and may be disappointed by it (or love it!). That’s also important because becoming an academic means joining a community and contributing to the advancement of the discipline. So it’s good to have a better understanding of what the field is doing right now.

A Chat with Dr. Marshall

Associate Editor Emmett Werbel recently sat down with Dr. John Marshall, Assistant Professor of Political Science at Columbia University for an informational interview, recorded in the form of a podcast. The first half focuses on political economy research methods, and Dr. Marshall’s specific theses, while the latter half addresses implications of this research on global politics and media structures. Their conversation can be found here.
To read more about Dr. Marshall, feel free to visit his website: http://polisci.columbia.edu/people/profile/1682.
The papers Werbel and Dr. Marshall touch upon the podcast include:
Publicizing malfeasance: How local media facilitates electoral sanctioning of mayors in Mexico;
Signaling sophistication: How social expectations can increase political information acquisition;
Political information cycles: When do voters sanction incumbent parties for high homicide rates?;
“Priors rule: When do malfeasance revelations help and hurt incumbent parties?

A Chat with Katherine Krimmel

Katherine Krimmel is an Assistant Professor of Political Science at Barnard College of Columbia University. Prior to Barnard, Professor Krimmel received her Ph.D. from Columbia University in 2013, was a National Fellow at the University of Virginia’s Miller Center of Public Affairs from 2012-2013 and an Assistant Professor of political science at Boston University from 2013-2016. Her research and teaching interests include public opinion, political parties and their relationship to interest groups, fiscal politics, and the politics of race and gender. Her current book project examines the organizational foundation of contemporary party polarization.

Associate Editor Rachel Weintraub sat down with Professor Krimmel to talk about her current research project. The following transcript, compiled by Weintraub, has been edited for clarity.

 

What is the focus of your current research project?

My research is focused on the major party organizations [the RNC and DNC]. I’m studying how the parties have changed and developed, and specifically at what pace and in which ways they’ve grown over the last century. We have snippets of data over time about their staffs and their budgets, but I’m trying to fill in that history to figure out when and how they became the major professional organizations–or corporations– that we see today.

I argue that the foundation for modern party polarization was built in the New Deal era, long before it started skyrocketing. Polarization is hard for parties- developing alternative positions on a wide range of issues takes a lot of work. This wasn’t work that national party organizations were in any position to be doing in the early 20th century, because they were largely decentralized. Specifically, The New Deal gave the national government a lot more responsibility, which subsequently gave the parties incentives to shift power from the local to national level. The New Deal also ultimately was hard on local party machines. As patronage became less important in party politics, it had to be replaced by something, so parties started to form stronger positions on issues to compete with one another. It’s causal in a loose sense. I’m not arguing that the New Deal caused polarization, but that the roots of polarization that we see today actually go back further than is commonly assumed. Party polarization is not a new phenomenon.

 

How did you become interested in this area of research? Has this project built on any of your prior research?

In some ways, the book project I’m working on right now was sparked by my work on my dissertation, in which I studied group party linkages. Throughout that process, I found that there was little written in the academic literature about the evolution of the national party organizations. My research is therefore trying to fill this gap in the literature.

Additionally, a report published by the American Political Science Association in the mid 1950s actually criticized the parties for being too programmatically similar and not offering voters clear choices on issues. Part of the reason why the APSA felt this was the case– and that party organizations were failing voters in this way– was because the party organizations were flimsy institutions, and therefore not capable of developing positions on a wide range of issues. In tracing their growth over time, I hope to provide a better understanding of how these organizations grew to be better positioned to take clearer positions on issues.

 

What has been the most challenging aspect so far, and how are you planning to address this challenge?

Gathering my data has been very logistically difficult and has been the most challenging aspect of the process so far. For my data, I am using a variety of archival documents. I made a list of all the places where these [archives of national party organization] documents are housed- and came up with a list of 28 sites across the country. Luckily, Barnard students are from all over the place, so I’ve hired a few research assistants to help with data collection.

Additionally, there’s no one right way to synthesize all of the information from these documents, and it’s not something that you’re really taught. I did some archival work for my dissertation, and how I did it then was by collecting and reviewing each document individually. I would make notes on them and tag them for being relevant to different aspects of my research. I’m hoping this time to find better software to streamline this process- I think there’s probably some software that exists now to create a better system for review.

 

In today’s political climate, many point to extreme elite polarization as problematic. In general, how will your research improve our understanding of the contemporary polarization that we see today?

It depends on how you think about polarization, and what exactly that means. Does it mean that opinions of Democrats and Republicans have grown further apart? Or does it mean that there are no more moderates, and only extremes? In that sense, the public is definitely not polarized; there are still plenty of moderates. However, there is no disagreement that the elites have polarized. Polarization is not new; there was also high levels of polarization at the turn of the 20th century. It was really the low polarization in the middle of the 20th century that is the exception. While polarization is definitely at historic highs, today, it’s more than just party polarization that people are talking about. For example, there’s been a lot of talk of urban versus rural polarization, especially in the wake of the 2016 election. It’s definitely something that a lot of people are worried about.

However, not everyone thinks polarization is a problem. Back in the mid 20th century, when the parties were really similar, people didn’t like that either. Like I mentioned before, even the American Political Science Association didn’t like that, because they said if the parties are really similar, what kind of choice does that give voters? The parties need to be different in order for voters to have a clear choice to make. Polarization itself is not necessarily a problem; rather, it’s the phenomenon associated with it– like betrayal and lack of compromise– that are unpopular ideas.

 

Lastly, what have you enjoyed about being a professor and getting to engage with this type of research? Do you have any advice for students who aspire to work in academia?

I can’t imagine anything being more professionally satisfying than getting to answer questions that are interesting to you. As a professor, you are able to do this constantly, whether it is through your own research or syllabi and course building. As for advice for students aspiring to go into academia, I think the colloquia and seminars here give you a good opportunity to practice research skills. They introduce you to the process of doing this kind of research and will get you thinking about what kind of questions you think are important. Overall, research-based classes at Barnard and Columbia provide a good opportunity to hone research skills, and get familiar with different topics.

 

A Chat With Professor Philip Oldenburg

Philip K. Oldenburg is Adjunct Associate Professor of Political Science at Columbia University. His research focuses primarily on Indian politics, and in particular local government and elections. The editor or co-editor of ten volumes in The Asia Society’s India Briefing series, and most recently India, Pakistan, and Democracy: Solving the Puzzle of Divergent Paths, his current project is tentatively called The Indian Politician.

Associate Editor Sarah Poff sat down to talk with Oldenburg on his research and teaching. The following transcript, compiled by Poff, has been sparingly edited for clarity.

Could you tell me about your recent research interests and what sparked them?

Since I am basically now retired, I have started a long term project which has a working title of The Indian Politician. My reason behind doing it is that there has not been a book on this topic of Indian politicians, which to me is amazing. It’s an interesting topic—why hasn’t anybody written it? That said, I have begun working on it.

Over the years, my research on other topics has led me to meet with many politicians in India. I have come to realize they differ immensely from the politicians depicted in the cartoons and newspaper clippings, which are generally very unflattering. The politicians that I’ve met have been extraordinarily hardworking: they don’t sleep at night because people knock on their doors at five in the morning until midnight on a day-to-day basis. I was very impressed by the stamina of these guys; I wanted to find out more about them.

Since you are in the early stages of this long term project, could you tell me about some of your objectives for The Indian Politician?

There many full-time politicians on every level in India, many of whom don’t even make a living out of politics. Thinking about all these people, I wanted to create a picture of what the characteristics of the Indian politician looks like. For example, to what extent is the Indian politician a man? A woman? I want to create a more rounded picture to replace the cartoon character that most people think of as the Indian politician. This was the idea.

Could you talk about the methodology you plan to use?

How does one go about studying this Indian politician? I’ve started by looking at where politicians emerge.

I will be spending four to five months in India for the next five years, observing elections and talking to politicians. So far I have been been clipping newspapers and articles and reading autobiographies, which has proven to be very interesting. I am yet to identify the questions that need to be solved for this book.

I am planning on interviewing politicians on every level. I am interested in asking essay questions with some sort of a sampling. The bottom line is that it is impossible to do the whole of India. I am confining myself to Hindi and Urdu speaking areas, which is around three-fourths of the population. I chose these regions because I plan to conduct all of my interviews myself, and I can speak and understand both of these languages. I will be confining myself to regions I am familiar with. Other people I plan to speak with are local journalists and swingers. These people know the politics of their particular neighborhood and should be able to identify who the politicians are.

Have you faced any challenges in your research thus far?

It will be a challenge to find out who the politicians are, and how I will gather a representative sample. Since India is so large and diverse even within regions, how does one sample all these different types of politicians, as well as the informal local representatives who are involved in politics full-time as middle men? How do I find these people will be a challenge.

Do you have any expectations for what the average Indian politician may look like?

It will be interesting to see what the most important characteristics are: for example, age, education, caste, or other work in civil society. I’ve met politicians of every variety so I don’t know. Ultimately, I cannot hope to believe I will have nailed it in the end. But I hope to raise the issues and make a portrait of what the Indian politician looks like. For example, to be able to say something like, “He is male ninety percent of the time, female ten percent of the time.” I suspect that certain characteristics will be prominent throughout as basic qualifications for being a politician, like stamina, intelligence, comparative wealth, gift of the gab. What else might emerge: that’s what I want to find out. Empathy? Ability to work with the opposition? Both may be valued in different places and different ways. Talking to journalists and interviewing politicians, I hope to get some sense of it.

How many of the classes that you teach are based on your past research?

A lot. Right after college I went to India and since then I have been studying the politics, language, and culture in different regions of the country. I wrote articles on elections and got interested in development research on the grassroots level. I’ve been teaching the course Political Systems in South Asia ever since. The other courses I taught were Contemporary Civilization, Political Change in the Third World, and Ethnicity in Politics in South and Southeast Asia. What I teach in my classes is based of of the work I have done in regards to my Indian and Southeast Asian research, and many of the people on our syllabus are people that I know well.