Luis Schiumerini on Incumbency Bias

Decreasing incumbent capacity and affective polarization are making incumbency salient, but also more of a curse than a blessing.

Luis Schiumerini

In this episode of The Democracy Paradox, host Justin Kempf speaks with Notre Dame political scientist Luis Schiumerini about his new book Incumbency Bias: Why Political Office is a Blessing and a Curse in Latin America. Schiumerini challenges common assumptions about incumbents, demonstrating that holding office can create systematic advantages – or disadvantages – depending on the institutional context. Before the conversation begins, Kellogg Doctoral Student Affiliate Tomás Gianibelli joins Kempf to share his experience working with Schiumerini and to explain why this research reshapes how scholars think about democracy.

 The Democracy Paradox is made in partnership with the Kellogg Institute of the Keough School of Global Affairs at the University of Notre Dame. 

Read the full transcript here.

Key Highlights

• Introduction - 0:20
• What is Incumbency Bias - 6:06
• Incumbency Disadvantage on the Rise - 23:58
• Term Limits - 31:39
• Implications for Democracy - 36:52

Links:

Learn more about Luis Schiumerini

Learn more about his book Incumbency Bias: Why Political Office is a Blessing and a Curse in Latin America

Learn more about Tomás Gianibelli

Learn more about the Kellogg Institute.

Chapters

• 0:20 Introduction
• 6:06 What is Incumbency Bias
• 23:58 Incumbency Disadvantage on the Rise
• 31:39 Term Limits
• 36:52 Implications for Democracy

Transcript

Introduction

jmk: Today's guest is Luis Schiumerini. He is an assistant professor of Political Science at the University of Notre Dame, a faculty fellow at the Kellogg Institute for International Studies, and the author of the book, Incumbency Bias: Why Political Office is a Blessing and a Curse in Latin America. But to help me introduce today's episode is Tomás Gianibelli. Tomás is a PhD student here at Kellogg and works directly with Luis. Tomás, how are you doing today?

Tomás Gianibelli: Hi Justin. Thank you for having me. I'm great.

jmk: So, Tomás, I have always felt that incumbents really pose a challenge for democratic theory. I mean, it creates an unlevel playing field especially if incumbents really do have an advantage. But what caught me by surprise when I was reading Luis’ book and during this conversation that I had with him is it really challenged some of my assumptions because he finds incumbency to be a disadvantage in some situations. So, when Luis first introduced these ideas to you, were you surprised by his findings when you learned about this research?

Tomás Gianibelli: Well, I think the idea of an incumbency advantage is a very US-centric idea. But when you go outside the US, you start finding these situations where being an incumbent creates a disadvantage. For example, in Argentina, the place where I'm from and Luis is from, at the subnational level, a lot of incumbents have these disadvantages. And I think that Luis does an excellent job in trying to understand why.

jmk: So, tell us a little bit about something that you've learned from Luis that helps us open up as to who he is and what we might learn from this episode.

Tomás Gianibelli: Well, I first met Luis when I was applying to the PhD program and I quickly saw and realized the depth and dedication Luis poses to academic work. Plus Luis has this strange capacity to really dissect theoretical problems and also to understand how some theories can be tested empirically.

jmk: I just find that he's so passionate about the subject matter. He is a very passionate person who cares very deeply about political science and in particular about democracy. So, it's such a refreshing opportunity to have a conversation with people like that. And so, I just really feel that that comes out in this conversation with Luis.

Tomás Gianibelli: I think that this offers a very refreshing take on how to interpret incumbency advantage and disadvantage. I think that's a novel idea that the listeners of your podcast are really going to find interesting and exciting.

jmk: Yeah. What really surprised me was the way that he talks about incumbency both as an advantage and a disadvantage, because I do think that that shakes a lot of the assumptions that we have about democracy and we get into that within the conversation about how that might challenge the way that we think about democracy, but also democratic backsliding as well.

Tomás Gianibelli: I think it touches upon a, a central question that we have nowadays that is do democracies deliver? And how is that related towards perceptions that people have on how democracies deliver or not.

jmk: So, Tomás, it's a huge opportunity here at the University of Notre Dame to work with scholars like Luis at the Kellogg Institute. Why don't you just take a moment to talk about one of the projects that you've had the opportunity to work on since you've come here.

Tomás Gianibelli: Well, I'm mostly interested in the relationship between politics and economics. So for example, I'm interested in how informality in Latin America shapes preferences towards redistribution and also affects democracy. I'm currently trying to test that with some empirical data. Also, I've been working on another project for a class that is basically trying to understand the positions of political parties after the 1980s and how that relates to support or backsliding of democracy in Latin America.

jmk: Well, the 1980s is a fascinating time period because it's right at the moment when so many countries were democratizing and you really have the birth of the Workers' Party in Brazil. You've got the birth of parties within Argentina and you have transformations within countries like Venezuela that eventually lead to either consolidation within democracies or setting themselves up for backsliding like we see within Venezuela. So, that sounds like a fascinating project. Are you planning to attend the upcoming conference on democracy here at the Kellogg Institute at Notre Dame in May, the Global Democracy Conference?

Tomás Gianibelli: Yes, for sure. I'm, I'm really excited to be part of it.

jmk: Definitely. Make sure that you guys are looking for it on May 19th and 20th. It's going to be a fabulous event. Make sure that you mark it on your calendar. Of course, the Democracy Paradox is made in partnership with the Kellogg Institute, part of the Keough School of Global Affairs here at the University of Notre Dame. Thank you so much, Tomás, for taking the time to be able to help me introduce Luis.

Tomás Gianibelli: I hope your listeners enjoy Luis

jmk: So, for now, here is my conversation with Luis Schiumerini.

Interview

jmk: Luis Schiumerini, welcome to the Democracy Paradox.

Luis Schiumerini: Pleasure to be here, Justin.

jmk: Well, Luis, I was very impressed with your recent book, Incumbency Bias: Why Political Office is a Blessing and a Curse in Latin America. It feels like a very timely book because there's a lot of questions about term limits and incumbents holding onto office. Your book doesn't touch on some of those subjects, but it does ask some important questions about the role of incumbency, not just in terms of the top executive of a country like a president, but also many other offices within a country.

So, I think this is both very timely book and a very important book, plus it challenges some of our assumptions, which is always something that I like to do on this podcast - challenge some of our preconceived notions about some of these concepts. Why don't we start here? You argue that incumbency can be a disadvantage at times. Can you explain to us why incumbency would ever be a disadvantage for somebody who's holding political office?

Luis Schiumerini: So, there are good reasons to think that incumbency may not be a blessing necessarily, because incumbency, at its core, exposes politicians. It's ultimately about whether what's being exposed is something that people like or don't like. So, why is it that sometimes it's an advantage and sometimes a disadvantage? As far as your specific question of why it might be a disadvantage, I argue that this is because sometimes institutions make it very hard for incumbents to please voters. Specifically, institutions create high expectations. While institutions have high policy scope, that means that incumbents have massive responsibilities. Voters expect a lot of them. Think of a president. Think of a governor. Think of a mayor.

In Brazil, which is a country that I study, mayors are basically in charge of health and education, but those expectations often are not matched by the fiscal resources that is the incumbent’s capacity to deliver. Incumbency disadvantage emerges when citizens have high expectations because incumbents have high policy scope and incumbents have relatively low capacity to meet those expectations.

jmk: So, Luis, it sounds like incumbency advantage or disadvantage happens because of a mismatch between the capacities of those in office and the expectations of those governed, the voters. If the voters think an incumbent, an office holder, has way more ability to impact their lives. They're going to be extremely disappointed if they don't see that office holder, that governor, that mayor impacting their lives in the way that they expect. On the flip side, if they think that a mayor or a governor or president doesn't have much influence over their daily lives, over policies and governance, they're going to be pleasantly surprised if things actually go in a direction that they want it to move. So is it really just that simple? Is it just a mismatch between expectations and capacities?

Luis Schiumerini: It is. That’s the argument I make. But I think you’re implicitly getting at a question of what voters know and how voters learn, because I think implicit in your question is if voters are accustomed to this, why don’t they adjust their expectations like you would do with the assessment of an employee or your team when you watch football. The thing is that for voters, it's hard to adjust their expectations because they don't know exactly what the capacity is that incumbents have to meet their expectations fully is given volatility. Voters have relatively fixed expectations. They know the mayor is in charge of roads, the governor is in charge of the economy, the president also oversees inflation or something like that. That's fixed and voters have a good sense of that.

What the voters don't know is that times change in terms of the ability of the incumbent to deliver on those things, because that has to do with changing circumstances such as the international economy or fiscal shocks in the case of some national politicians who have very limited ability to actually produce resources. So that's why even though voters could learn, they have difficulty because the prior mayor could have had good times and the present mayor could have had bad times. So that impedes their learning and adjustment of expectations. That's why it seems simple, but it is complex on the part of voters, given the difficulty that is intrinsic to evaluating policy makers in contexts that are changing.

jmk: So do you find that incumbency advantage or incumbency disadvantage stay stable for long periods of time or do you find there's some volatility between those two? Like sometimes an incumbent for a position, like, let's say, president has an incumbency advantage for multiple cycles, but then events change, voters learn, and suddenly incumbency becomes a disadvantage. Does that happen or is incumbency generally stable over long periods of time?

Luis Schiumerini: Because the features that I argue drive incumbency bias, advantage or disadvantage, are fairly structural, there is a lot of stability in the direction of the incumbency bias. For example, in the US Congress or Argentina whenever there are elections almost invariably there is an advantage. The same goes for mayors in Brazil. They are typically at a disadvantage. However, there is volatility over time. On average, you'll see that if you look at the whole there is an advantage in the US Congress and disadvantage for Brazilian mayors. Now, elections can change from year to year and that where the volatility comes from, because the capacity is volatile whereas's expectations are fixed.

Let's say Argentine Governors have an average incumbency advantage of 40 percentage points, but they have cycles in which it is 80 and cycles in which it is 10. Similarly, we have seen a decline in the incumbency advantage of US Congress people, even though on average they are still advantaged. So, I would say there is stability in the direction, positive or negative, and there is volatility in the magnitude.

jmk: So you just mentioned that Brazilian mayors have a real incumbency disadvantage, if I remember right. Tell me more about that. Why is it that a Brazilian mayor would have an incumbency disadvantage, because I usually think of mayors as having a lot of control over what happens within their cities?

Luis Schiumerini: I was also surprised by this finding. Brazil, especially at the local level typically elects strong men that are really hard to beat and yet we find an incumbency disadvantage in Brazil at the local level. So that's difficult to reconcile. In the book, I argue that Brazilian mayors are a case of too big to succeed. Brazil has a very peculiar federal system in which localities are heavily empowered, not mid-level governors, but localities. This is unusual. It is a federal system where most of the policy action is at the municipal level. So much so that mayors are fundamentally in charge of health, education, and transportation. There is some shared responsibility with governors, but it is mayors who do most of it.

In the book I show that there is plenty of survey evidence showing that voters overwhelmingly think that the mayor is the most relevant actor for their wellbeing at the city level. That creates very high expectations where people are laser focused on what the mayor delivers on. At the same time, most Brazilian municipalities say almost 90% fully depend on transfer from the federal level and the state level. So, while voters expect a lot from them, they have little ability to produce their resources, and on top of it, Brazil has one of the most stringent fiscal responsibility laws on the planet. That makes it all but impossible for mayors to direct spending in ways that are not earmarked or to incur debt.

Since the fiscal responsibility laws made in Brazil in 2000, you see almost no debt being purchased by mayors. That's a perfect storm. You have massive expectations, very limited resources to produce. That's why on average Brazilian mayors are disadvantaged. However, we do see exceptions. Around the commodity boom, the second half of the 2000s, we see an incumbency advantage. It is the only time period in which we see an incumbency advantage for Brazilian mayors which is no accident. We see massive transfers going to the local level and those transfers were effectively allocated into public goods spending. I show in the book the mechanism that drives the connection between transfers and incumbency bias is fundamentally spending on public goods.

jmk: So even though they're responsible for the administration of so many of these policies, they don't really have control over the amount of resources that they have for those policies. So voters are expecting that they have a lot more influence over their daily lives because they see how mayors oversee so many aspects of governance and yet the actual decisions behind them are really made at the federal or the state level.

Luis Schiumerini: Yes. But a key piece of the story is the fiscal responsibility law that limits discretion. So, in my book, I say there are two components to incumbent capacity - the resources, but also the discretion. Many countries in the world have some national governments that do not produce resources. This is the case for Argentine governors. However, Argentine governors have absolute discretion. They can do whatever they want with the money because the rules are lax, but also because there is no enforcement mechanism.

In that case, even though the money comes from the federal government, they can do whatever they want with it, so that allows them a lot of capacity. That's why even though Argentine governors have many responsibilities in terms of the policy authority they have - and voters expect a lot from them, they still have all the money to meet those demands, whereas Brazilian mayors are in a difficult position. Both are fiscally independent, but one has discretion and therefore capacity, that is Argentine Governors. Brazilian mayors are also limited but lack the discretion. For context, the fiscal responsibility law was the grounds in which Dilma Rousseff was impeached. You may debate the degree to which that was fair, but that illustrates the strength of that fiscal responsibility law.

jmk: So why don't voters recognize how much capacity Argentine governors have to influence their lives? Why is it that voters don't see the real power and authority that a governor has within Argentina and take that into account when they vote in elections?

Luis Schiumerini: I think that there are two parts to that. I think voters see something that is very tangible and it's natural that they respond to it, which is what governors or incumbents do with that money. Voters, and I show this in the book for all the cases that I study which are Brazil, Argentina, and Chile, respond to public goods spending. When that money is spent on things that are of general interest and tied to the welfare of citizens, they respond to that. So, in that sense, voters are incentivizing good behavior. It's not clear to me that they should be behaving differently. I mean, we could speculate with a counterfactual.

Voters perhaps could incentivize even stronger behavior if they imagined that money could be spent better or if they knew that the capacity was there. But I think that that's debatable that voters could be offered a counterfactual. So I think that many models of retrospective voting find, and there has been a lot written on this, even political philosophers advocate for this behavior, it makes a lot of sense that voters would reward that as long as there is observable evidence of money being invested in things that voters care about.

jmk: So is incumbency disadvantage the exception and incumbency advantage the norm?

Luis Schiumerini: I'd say no. I think there is no norm. The book argues that the norm would be incumbency bias in the sense that it's intrinsic to electoral accountability. That's one of the central lessons of the book. This is nothing mysterious. We tend to think that if we see incumbency advantage, there is something amiss. Maybe politicians are stealing elections or repressing the opposition or if there is a disadvantage it is because voters are so alienated that they just throw the bums out. What I'm saying is that most of the evidence that we have about voting behavior, not only in developing countries, but also advanced democracies, is that voters reward performance.

To the degree that voters actually reward performance, incumbency bias is a natural result because the only way in which you do not observe incumbency bias is if expectations and capacity are perfectly matched and that will be rare. So, incumbency bias is the norm, whether it's an advantage or disadvantage for the incumbent really is a matter of institutions. Still, I find this is difficult to measure for every country in the world. I haven't done it, but I think from what we know, it's almost a 50-50 split at the subnational level.

jmk: So is the rise of personalism within politics then increasing the amount of incumbency bias throughout the world?

Luis Schiumerini: I think so, yes. I think so. I'd say this is a bold claim and you are generalizing to many offices, but I think that even if you think about presidential elections in the US and in large settings, I think it becomes very clear that when people compete as people and personal characteristics come to the fore, incumbency is a very salient cue. It's going to be about what I did as the incumbent and about what you didn't do as the incumbent. That will be the argument of the challenger. So, personalization and incumbency bias, I think, certainly go together.

jmk: But incumbency bias is not necessarily an advantage or a disadvantage. It could be one or the other. So does rising personalism within politics, does it lean more towards advantage in incumbents or disadvantage, or does it just depend?

Luis Schiumerini: I think it really depends on whether incumbents are running in conditions that make them favorable to them. We could talk about this as a segue into an important question that I've been thinking hard after finishing the book because it's something that wasn't the focus, which is the recent trend of incumbents losing elections across the world, which is quite surprising. So even though my book is not about national elections, I think it carries some lessons that would help us understand what's going on. And what you have right now are perhaps two features that my book emphasizes are helpful. One is that there is increasingly less capacity of incumbents across the world to meet citizens demands.

At the baseline, we tend to think correctly that presidents have a lot of resources, a lot of tools to meet citizens' expectations. At the presidential level, if you don't have the resources, you can get debt. It's almost unlimited what you can do. You can overheat the economy in the run up to the election and create an illusion of prosperity that will be paid later, but will allow you to get reelected. This is a story that is well known. It's been documented in the US. Javier Milei just won a legislative election in Argentina through the help of the Secretary of Treasury in the US. However, over time I think this capacity has declined because the policy tools of incumbents at the national level are less able to meet citizens' expectations.

The affordability crisis, for example, illustrates that the policies they use to make the economy grow do not trickle down to people's pocketbooks. The last presidential elections of the US illustrated this very clearly. Look at the macro figures. The economy was doing well in the US. Even if there was inflation, it was outpaced by salaries and by growth. Then you ask people and they were unhappy with what's going on. I think it was not just that people were looking at the wrong media or being lied to. I think that there is an increasing gap between the policy tools that incumbents use and people's material welfare, the ability to buy a house, the ability to buy groceries.

Furthermore, I think with the transformations in the economy, people are having different types of jobs that are not tied to traditional wage bargaining mechanisms or traditional industries. It makes them less sensitive to what the government is doing. So, I think there is a decrease in incumbent capacity. Secondly, and this ties to the idea of personalization, I think what we're seeing is that affective polarization is hard on incumbents, because it's easier to mount a negative coalition against someone than to mount a positive coalition in favor of someone. There is a negativity bias that comes with affective polarization that is stacked against the incumbent.

In the US before 2016, the sitting incumbent candidate had won two thirds of presidential elections in the US. That's a one third incumbent advantage. That's a very large incumbency advantage. But we just saw a sequence of two losses. This is quite unprecedented. The defeat of Trump and then the defeat of Harris, who was from an incumbent party. In those cases what you observed was that there was a sense in which a large part of the coalition of people who were unattached, disaffected, rallied in support of the challenger. So I would say decreasing incumbent capacity and affective polarization are the two forces that are making incumbency salient, but also more of a curse than a blessing.

jmk: Well, the United States is an interesting case, particularly for the presidency because not only have we seen turnover in multiple elections recently but even dating back to 2012 when Barack Obama was running for reelection, he saw a significant decline in his vote share. In fact, a lot of people thought Barack Obama might lose the 2012 election, so you really have to go back all the way to 2004, more than 20 years ago now to find an incumbent that actually increased their vote share in reelection. That's a surprising outcome for an office that people still think implicitly has an incumbency advantage. It sounds like the presidency in the United States clearly suffers from an incumbency disadvantage at this point. Do you think that that's the case, or do you think that we're just looking at the numbers wrong?

Luis Schiumerini: So, I think that in general, it's hard to identify what is purely incumbency bias and what's rooted in differences between incumbents. That's why my book is on subnational elections because I can draw on a lot of observations to use methods that allow us to distill what's purely an incumbency effect and purely something else. Incumbency bias, as I define it, is the systematic difference due to holding office and nothing else. So, when we see that an incumbent loses or wins an elections in presidential races in the US, it's hard to know. But going back to your point, I think that the higher rate for reelection of incumbents in the US compared to challengers is not purely incumbent bias, but just the fact that incumbents are of better quality.

Going to your example of 2012, did Obama beat Romney because he was the incumbent or despite the fact that he was incumbent, because he's a better quality challenge? I'm pretty sure that Obama was a better quality candidate than Romney. So structurally, I think going to your point, we overstate incumbency advantage from the mere fact that incumbents tend to win in presidential elections in the US. I think that it's often more of a liability or, at least, it comes with some baggage.

The fact that incumbents overcome it is not so much the effect of holding office, but the fact that sometimes there are exceptional candidates, as most incumbents in the US that have been reelected are. If you think about it for the most part, they have been high quality candidates, not because I judge them high quality, but objectively in terms of the things that you might attribute value to from the perspective of voters.

jmk: Obviously, to win an election, there must be something about you that allows you to win that election and the more significant that office is, the more competition there is for that. So, while you might not be the best in office, you clearly demonstrated something about yourself in the way that you campaign and the way that you're able to communicate and speak to voters. It sounds like there's really intangible advantages that somebody who currently holds office will have, because of the fact that they're the type of person who's already proven that they're capable of winning an election that gives them a quote unquote incumbency advantage that isn't really about holding office, but really just about the fact that by being an incumbent, it means you've already proven your ability to be a strong campaigner.

Luis Schiumerini: This is a hugely important point that you're making, Justin, and this is not just an empirical curiosity. It really goes to the core question. The reason why we see incumbents systematically winning or losing elections is something about the candidate as opposed to a challenger, not about incumbency per se. That's actually what we might expect of democracy. Elections are selecting winners based on characteristics of the candidates. So, what my book tries to do… And by the way, this was something that Bob Erikson in the 1971 piece that started the literature on incumbency advantage already identified then is that when we talk about incumbency and what I call incumbency bias proper is the effect of holding office per se. That's what we need to do.

To the extent that we find that holding office per se has an effect separate from quality then we need to investigate why, because it could be rooted in unfair practices that we would deem undemocratic and that would indicate something profoundly wrong with how institutions work. If, however, what we find is that the mechanisms through which office holding matters systematically are compatible with electoral accountability, then incumbency bias is a much less concerning phenomenon.

jmk: Let's circle back to Argentina because I find that's a very interesting country, particularly at the presidential level. You had stated that the presidency gives you a lot of resources that should make the office advantageous for incumbents, but Argentina's been a country with cycles of turnover for the last few cycles where the Peronists win power, someone who's more conservative wins power, the Peronists took power again, and now we have Javier Milei, who's far more conservative, even more conservative than Macri, who had held office the cycle before the cycle beforehand. Is Argentina a case where incumbents suffer significant disadvantages because of the economic conditions within the country or is it because of the structural nature of the institutions themselves?

Luis Schiumerini: That's a great question. So, Argentina I think fits neatly what we were just talking about in the US and really the broader trend of incumbents having a hard time getting reelected at the national level across the world. Before Macri’s defeat, two elections ago, all Argentine presidents running for office had been reelected and in Latin American over 90% have been reelected. You recently had Javier Corrales on the podcast and he has a very interesting article with Michael Penfold in the Journal of Democracy, where they talk about this. So, presidential elections in Latin America in general are known to favor, at least as far as rates of reelection go, the incumbent. However, recently we saw two sequential incumbent defeats in Argentina in presidential elections.

But I think that what we're observing, as you rightly pointed out is really how external conditions affect performance. The prior incumbents that had won elections in Argentina had overwhelmingly done well in terms of the economy, which is the metric that voters used to evaluate the president - economic growth and employment or inflation, which is a highly salient issue. All of them did well objectively, and also if you look subjectively at voter evaluations when they got reelected, Macri did not do well. Alberto Fernandez did not do well. They were ousted as a result.

So, Argentina's presidential elections illustrate what the core point of the book is, which is electoral accountability, meaning very plainly, voters electing or ousting comment on the basis of performance in office can create incumbency biases. But because conditions shape what incumbents can do, that has been an issue for the Argentine presidents that lost the last set of elections in addition to their own incompetence because they made mistakes.

jmk: So, has this research changed your mind at all about term limits either for the need for them within certain offices or maybe reasons that they're not as necessary for others?

Luis Schiumerini: Great question. So according to some research term limits are bad for democracy. There are all these papers often written by politician economists and economists showing that when an incumbent faces a term limit, they underperform. There is a lot of work showing that for US governors, but also for Brazilian mayors. That's the article of faith. Term limits are bad for democracy because they discourage good performance. I think that my book suggests that perhaps that should be qualified or at least consider the context. So, the Argentine case, I think, is worth evaluating. And what we find in the Argentine case is that term limits are not only absent, but some provinces have indefinite reelection. Not only the inability to run for three terms consecutively, but the ability to stay in office forever.

What my book finds is that term limits help break the incumbency advantage of governors in Argentina by reducing the ability of the person to run for office. Because a significant degree of the incumbency advantage of governors is personal, term limits break that personal incumbency advantage and therefore can bring some oxygen into democracy. So, I would summarize by saying it is true that term limits can discourage performance and in settings where there are no concerns with democracy that may be a net negative to bring term limits in. However, in settings in which there is a risk of building a hegemony, which is the case in many Argentine provinces, we may want to consider term limits as a way of breaking those hegemonies.

jmk: I'm surprised to hear you say that scholars argue that term limits are bad for democracy because so much of the literature on democratic backsliding is referring to the removal of term limits as a form of executive aggrandizement and a real threat to democracy particularly when we're talking about the presidential level. It surprises me that it would be argued as good for democracy for incumbents to run for office for an indefinite period of time.

Luis Schiumerini: That's a great question. I had not thought of that tension, but I think you touch on something which is core to this debate in that I think we should think of different facets of democracy. I think that people who have argued that term limits are bad are not focusing on democracy as a regime and its survival, but about electoral accountability and the degree to which elections encourage good performance. Now what you are posing is a bigger risk to democracy. Settings in which the abolition of term limits is paving the way for democratic backsliding, ultimately the disappearance of competitive elections.

So, I think then you have a tension between electoral accountability and the regime. In those settings, term limits are certainly called for. This is partly a reflection of the different settings that are being studied. People will talk about term limits as something bad when they look at the performance of US legislators or governors or mayors and they narrowly look at public goods dispersion whereas the big question of presidents breaking democracy really calls for thinking of term limits as a way of salvaging democracy.

jmk: Yeah, especially when we think about autocratic governments oftentimes we find that the longer somebody stays in power, the more poorly they govern. The first few years are usually a little bit better than later on within their tenure. So again, it's just a surprising statement, surprising finding for anybody to say that we should be removing term limits, because maybe that incentive for a third term might make them perform better, since it could also be that some of those incumbents just aren't going to do that well in the second term. Maybe the first term is when they're at their peak performance and the second term happens to be worse performance. So if we allow them to do a third term, it could be just more of the same of the from the second term.

Luis Schiumerini: I think that question has not been studied with that nuance. What people typically find is that in the last term they do worse than in the prior term and it is inferred that that's because of the lack of reelection. Therefore let's allow reelection, but nobody had thought of the question of whether there is a time cycle where even if you allow reelection further down the line, because of decay or corruption or just becoming older, they underperform. I don't think we have research that allows us to make those nuances and distinctions out of the life cycle and how that would allow us to design the optimal term limit. It's just thought in a much more general way as presence or absence.

jmk: So, the big question that's on my mind is that as we think about incumbency now, and we're thinking about it in terms of term limits, we're thinking about it in terms of performance and this entire conversation. How has the study of incumbency really affected how you think about democracy?

What does incumbency bias say about democracy itself?

Luis Schiumerini: Well, that's a great question. I think that it allowed me to think about democracy in a much more nuanced way, but also to clarify some common misunderstandings. I think for the most part, there is an assumption that incumbency bias, and by that I mean incumbency advantage is bad for democracy. So, this Journal of Democracy piece by Javier Corrales and Michael Penfold makes this point. They argue that the presence of incumbent advantage questions the presence of electoral uncertainty and attacks a core tenet of democracy.

Now, my own work questions this interpretation. I think that there are some issues with considering that incumbency advantage because it affects electoral uncertainty is bad for democracy. I think at the most intuitive level thinking that incumbency advantage is bad for democracy would lead you to think that the systematic reelection of Congress in the US or even the very long spell of social democratic governments in Sweden after the Second World War is indicative of a lack of democracy. Nobody would argue that. This gets at a deeper point. The reason why we don't find it intuitive to think that the US Congress or Swedish elections are undemocratic is that we don't care so much about the electoral outcomes when we judge what regimes are democratic. We care more about the process.

Most scholars of democracy, and I include myself there, think that democracy is about procedures, having competitive elections, having basic freedoms that allow us to compete and vote, and also having limits on power. So you can have procedures that are democratic and elections that still favor the incumbent because the incumbent does well. So, what we ought to look at more closely when we worry about incumbency bias and democracy therefore are the processes that produce incumbency bias. A lot of the literature on incumbency bias in that sense gives us cause for concern. Most of the work on incumbency advantage in the developing world is that this is about clientelism and the widespread abuse of the machinery of the state in ways that makes it all but impossible to affect income.

The reason that that happens is not because voters evaluate income and wealth, but because the incumbent won through reasons that have nothing to with voters' preferences. This is similar to Sue Stokes’ argument about reverse accountability. It is incumbents who are controlling voters rather than the other way around. That's an undemocratic reason for incumbency advantage, massive clientelism. But conversely, most work on incumbency advantage also suggests that democracy is broken because corruption is widespread, so voters just throw the bums out without even looking. That also is a recipe for democratic failure. Voters have essentially withdrawn from the elector process. So, the processes that produce incumbency bias are such that accountability doesn't work here because voters cannot control politicians, because it is the other way around or because politicians are such crooks that voters cannot even pick correctly. That's almost an absence of democracy.

But what I would suggest contrary to that is the processes, the mechanisms, that allow for incumbency bias are perfectly compatible with definitions of democracy. It is about voters evaluating incumbents for their performance in context in which voters do not fully observe the capacity of incumbents and incumbents are constrained by the expectations that institutions make voters place on them. What I learned through this process is that incumbency bias is perfectly compatible with electoral accountability and perhaps the natural result of it.

jmk: But in terms of democratic backsliding, though, we've seen a large number of executives in office seizing greater powers and expanding their capacity that they have to govern while in office. Does that give them a greater advantage as an incumbent because they're going to be able to deliver more for voters and voters are going to see that as a reason to reelect them, so we're going to see incumbents with an incentive to erode democracy or does it merely expose them because voters expect more from those incumbents when they're seizing powers and seizing authorities and trying to deliver more? Does it expose those incumbents if they're not able to deliver the impressive results that voters are going to be expecting?

Luis Schiumerini: That's a great question. I think that you're pointing to something that is tricky and it's ultimately hard to distinguish, which is sometimes you acquire more capacity, but that also creates more authority, which increases expectations. I think that I would distinguish two ways in which incumbents can increase capacity. The institutional expansion of democratic roles of an incumbent is not a foolproof way of increasing advantage, because it also increases expectations. On the other hand, what incumbents are doing is not increasing their authority in terms of policy scope, but what they're doing is increasing their ability to actually weaken democracy by going after the opposition, controlling the media, or even meddling with elections.

That would certainly be advantageous, but that is then placing us in a setting in which we are talking about an incumbency bias that is fundamentally undemocratic. This hyper-incumbency advantage that has been talked about by Ken Greene in the context of the PRI in Mexico and Levitsky and Way in their book on Competitive Authoritarianism talk about this. There is a hyper-incumbency advantage that exists outside democracy. A new kind of regime where incumbents have the ability to tilt the playing field through undemocratic means.

My book shows that the playing field can be tilted through democratic means. There could be perfectly democratic reasons why it's harder for an incumbent to get reelected or it's hard for a challenger to beat the incumbent that have to do with performance evaluations. However, these other tools that are undemocratic like repression and coercion create an incumbency advantage that becomes a competitive autocracy.

jmk: Well, Luis Schiumerini, thank you so much for joining me today. The book, once again, is called Incumbency Bias: Why Political Office is a Blessing and a Curse in Latin America. Thank you so much for joining me today. Thank you so much for writing the book.

Luis Schiumerini: Thanks for having me Justin. This was a pleasure.