Comparative Politics Guillermo Toral

BBA BIR Assistant Professor of Political Science

Fall 2024 www.guillermotoral.com

The approach and methods of comparative politics

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

Comparative politics is a field of political science preoccupied with the causes and consequences of different aspects
of political systems and other political phenomena. The focus is on domestic or within-country politics, rather than
relationships across states (as in the field of international relations). Comparative politics is a scientific enterprise,
and as such leverages clear theory and credible empirical strategies (using qualitative and quantitative data) to learn
about general political phenomena – be it by describing, explaining, and predicting their variation. Rather than
normative (as in political philosophy) or mathematic (as in game theory) the focus of comparative politics is
empirical, i.e. phenomena as they occur in the world.

Comparative politics is a very diverse field, and it deals with:

- Cases from all around the world, both contemporary and historical.
- National, subnational, and supranational politics.
- Different levels of analysis. From the macro-level (e.g., at the level of the country) to micro-level units (e.g.,
voter, politician) or processes (e.g., decisions on who to vote for, or whether to join a protest).
- Political institutions (i.e., the settings in which politics occur) and political behavior within them.
- Formal as well as informal institutions.
- The behavior of citizens as well as state agents (e.g., elected politicians, bureaucrats, judges, etc.).

There are many different approaches or traditions to the scientific study of politics. A useful typology of approaches
to the political world (the “five i’s”):
- Institutions. Institutionalist approaches seek to understand politics by focusing on how the “rules of the
game”, or the formal structures of the political system, affect different aspects of government performance.
- Interests. Rational choice and other interest-centered approaches seek to understand politics as a function
of the strategic calculation of individuals (often understood as utility maximizers), and the aggregation of
those individual choices. This approach is more aligned with economic theory and mathematical model, but
often uses more informal phrasing too.
- Ideas. An alternative tradition in comparative politics approaches the political world through a lens that
emphasizes the role of ideas and culture in shaping political behaviors and outcomes. These could include
mass ideologies (e.g., communism, fascism, neoliberalism) but also smaller-range ideas such as
understandings of justice, development, poverty, or the state.
- Individuals. Yet another approach emphasizes the role of specific individuals and in particular of political
elites. In this paradigm, politics is a function of the backgrounds, choices, and even personalities of specific
individuals in positions of power.
- International environment. While comparative politics deals with domestic politics rather than inter-state
relationships, it is obvious that the international arena shapes domestic politics too. A fifth approach to
comparative politics emphasizes the role of the global on the domestic, be it through analyses of trade and
economic relationships, political or military alliances, or cross-national phenomena like migration or the
environment.
Often, to explain a particular phenomenon (e.g., the rise in the support of the far right, democratic backsliding, or
the survival of an autocrat) explanations can be advanced from each of these perspectives or paradigms. By training
our ability to build, understand, and assess arguments from different perspectives, we can gain a richer
understanding of the political world, and a more critical eye towards arguments and their implications.

1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Regardless of the paradigm or perspective, arguments in political science may be of four main types:
- Normative: These are arguments about what should be, and have to do with values.
- Descriptive: These are arguments about trends and patterns, about what is. They have implications for
organizing our view of the world.
- Predictive. These are arguments about what will happen in the future, or among units we have not observed
yet, without necessarily saying much about why that will happen. They have implications for forecasting.
- Causal. These are arguments about relationships of cause and effect, or the levers that drive political, social
or policy change for example. These arguments, which are central to (political) science, have implications
for our ability to change the world.

Social science, including political science and comparative politics, is science because it follows the scientific method.
This method involves the systematic studying of social or political phenomena following a set of procedures that
make the process of learning transparent, and the conclusions we draw subject to review by others. The goal of the
scientific method, in the social sciences like in other fields of science, is to draw generalizable conclusions.

Phases or moments of the scientific method. While the scientific process is iterative and often not linear, it involves
the following phases or moments:1
- Theory. Either deriving from observations of the world (inductively) or from other theories (deductively)
scientists build a theory to explain a particular phenomenon. A theory is a set of logically consistent
statements that explain a feature of the world.
- Hypotheses. Based on a theory about the world, we can then deduce implications even going beyond the
phenomena that we initially observed. It is important these implications or hypotheses be falsifiable, i.e.
testable through empirical research. The more and more specific the hypotheses that we can derive from
the theory, the better for our ability to test it and therefore for our ability to learn from it.
- Data. We can then collect data to see if the hypotheses or implications of our theory are consistent with
what we observe. When collecting data, the scientific method seeks not to “prove” the theory but rather
to look for the data that would best allow us to distinguish if the world behaves according to our theory or
not. Data can be quantitative
- Analysis. Once we collect the data, we analyze it to evaluate our hypotheses. There are different methods
and approaches for analyzing scientific data (e.g., experiments, observational regressions, in-depth
interviews, machine learning algorithms, etc.). In general, however, we typically seek to examine the extent
to which our observations are consistent with the theories, and to get a sense of how certain we can be
that they are. In any case, in science we never definitively “prove” a theory – instead we tentatively reject
alternatives or say that “the evidence is more consistent with X than Y.”

Features of good science. Good science (be it political, social, natural) has a number of features. We should assess
scientific outputs (reports, articles, etc.) based on whether they abide to these standards. These standards also allow
us to think more systematically about how we may go about testing ideas or hypotheses we may have about politics.

1. Clear theory.

A theory is a proposition or set of propositions about the relationship between variables, providing an internally
consistent logic to explain (political) phenomena, and which can be subjected to empirical testing through
observation or experimentation. Good theories:
- have some generality – in science we always strive to learn in general terms, abstracting from cases to learn
about the world.
- are phrased independently of the empirical cases used to build or test them
- specify any relevant mechanisms (or channels through which the key variables of interest are connected)
- specify any relevant scope conditions (or where / when it should apply)

1This approach to science is called positivist. By contrast, constructivist approaches challenge the idea that social phenomena
can be studied as natural phenomena, and emphasize instead subjective meaning and how it is socially constructed.

2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

2. Testable hypotheses

A hypothesis is a specific and testable claim based on a theory. It allows us to systematically evaluate the validity of
the theory. Good hypotheses:
- Are logically derived or deducted from the theory / argument
- Are precise
- Can be subject to empirical testing -- aka are falsifiable
- Are testable with different observations from those used to build the theory
- Are non-trivial or are surprising in some way

3. Rigorous methodology

A research design is a strategy for answering a research question using empirical material. We broadly refer to that
strategy as the study's methodology or methods. Good research designs:
- Produce internally valid conclusions (i.e., results we observe are in fact driven by the variables we claim
drive them)
- Give us a sense of the uncertainty around those conclusions (e.g., a confidence interval around a
quantitative estimate)
- Have some degree of external validity (i.e., allow us to draw conclusions that can be generalized to samples
or settings other than the one we studied)
- They specify their key elements or ingredients: a set of cases; a comparison (e.g., treatment group versus
control group; Trump supporters versus Harris supporters); a strategy for making inferences (causal /
descriptive / predictive); a strategy for measuring uncertainty around such inferences; and a rationale (and
supporting evidence) why the methods we use get at the key quantity of interest

Why making causal inferences is hard. If we want to learn about the effect of X (treatment) on Y (outcome), typically
looking at the level of Y among units with high vs low levels of X will not be a valid strategy. For example, does
democracy (X) cause economic growth (Y)? This is hard because social and political variables are typically correlated
with many other variables. Particularly challenging for credible inferences are variables correlated with both the
outcome and the treatment, aka confounders. Moreover, it may be that the outcome we are interested in is also a
cause of the treatment of cause we care about -- this is called reverse causality. We typically want to argue that a
relationship between X and Y holds ceteris paribus, i.e. " or "all other things being equal". But confounders and
reverse causality hinder our ability to do so. How can we say that that a relationship holds “ceteris paribus”, in a
credible way?

Methods that can be used to get to “ceteris paribus” include: process tracing; controlled case comparisons; statistical
models with controls; quasi-experiments; and experiments

Types of data that can be collected and analyzed in political science include ethnography; interviews; case studies;
surveys; administrative data; and satellite imagery.

Experiments: The gold standard to make causal inference. Experiments are the best way to learn about causal
effects. They are not always feasible or ethical but when they are they are a simple yet powerful method to test
causal arguments. Even when we cannot do an experiment, thinking through the experimental ideal can help us
think through other, non-experimental research designs. Experiments randomly assign an intervention (aka
treatment) whose impact on one or several outcomes of interest we want to evaluate. In the experiment, individuals
(e.g., people, households, firms) assigned to different groups are comparable because they were formed by lottery.
Therefore, any difference we observe between the treatment and control groups after the intervention must be
caused by it. It is the fact that the groups were formed exclusively by lottery that ensures that we do not have
confounders. And because the intervention is implemented before we measure the outcomes, we ensure there is
no reverse causality (since the future cannot affect the past).

3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

4. Transparency

Transparently reporting on the methodology, from data collection to analysis, is a central component of good
science. Transparency kicks in throughout the research process, but is particularly important at the time of
publication, and includes transparently reporting on the procedures used and their justification, and sharing the data
and programming code used for analyses. Transparency matters because it allows others to assess the validity of the
methods and to identify biases or limitations of the research, it others to reproduce the study (with the same data)
or to try and replicate it; and it allows the scientific community (and other stakeholders such as civil society
organizations, journalists, or activists) to reuse that data for other purposes.

Science (including political science) needs to be subject to high ethical standards -- otherwise significant harm to
individuals / communities can ensue. There are key ethical risks in comparative politics, because of two main
reasons. First, variables of interest in comparative politics are often very sensitive and/or consequential (e.g.,
identity, racism, order, democracy). Second, with regards to experiments, variables of interest can often not be
experimented on ethically (e.g., partisanship, electoral system). There are ethical risks even without
experimentation.

The key principles of ethical research with human subjects (in political science, medicine, etc.) are: informed consent,
privacy and confidentiality, minimization of harm, justice and fairness, respect for autonomy, prior review and
approval.

6. Social relevance

Science should contribute back to society, directly or indirectly. Comparative politics can be useful for a variety of
tasks, including designing institutions, improving policy outcomes, making forecasts, and promoting or resisting
changes. In that sense, there are many audiences beyond academic researchers for whom comparative politics may
be useful, including government leaders, political consultants, political commentators, policy makers,
businesspeople and firms, and civil society organizations.

Key skill: Dissecting an academic article

Academic articles are quite particular writing products, and they should be read / approached with an approach
different from the one you may use to read a textbook or a newspaper article. There are a few things to note about
these articles. First, while may seem obscure pieces of academic knowledge, they sometimes shape policy debates
and decisions. Second, they present cutting-edge research on a very specific question, sometimes leveraging
advanced research methods. And finally, they are written as a defense of a claim (theoretical and/or empirical);
authors try to close the door to critiques. There are three key implications of these three facts:
- As a reader, it is critical to understand what academic articles claim, how they support those claims, and in
what ways the evidence may be weak
- They often have a lot of technical details, jargon, often formulas
- They follow a common structure and have plenty of signposts to help the reader

Academic articles have a common structure that you should exploit when reading them. They are in a sense written
in layers as an onion, with significant amounts of repetition and varying amounts of detail.
- Abstract – short summary
- Introduction – longer summary, more details
- Theory (including hypotheses, the predictions the authors seek to test)
- Research design or methods (i.e., what data the authors use and how they analyze it to test their
hypotheses)
- Results (typically three sets of results, in order of importance: 1) main results, 2) mechanisms, 3) robustness
checks)

4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

- Conclusion – summary, implications, gaps for future research

Key take-aways for reading academic articles

- Don’t be intimidated or discouraged
- Exploit the core structure to your advantage
- Do not pay equal attention to all parts in the paper. Put more energy in the abstract and introduction.
- Do not read linearly -- it's not a novel or a newspaper article. Read selectively. First the abstract and intro,
then skim the rest of the paper, then go deeper into most important sections if you need to
- If you don’t understand something, go back to the bigger picture in the introduction or at the beginning of
the section you are at.
- Take a look at the main graphs and tables, where a lot of the message is condensed
- Leverage topic sentences (the sentences at the beginning of each paragraph, typically synthesizing what is
contained therein).
- Be comfortable with not understanding 100% of the paper

Key questions to ask of an academic article

- What is the author’s argument or thesis? Who or what is the author arguing against?
- What are the implications of the argument? What is at stake?
- What evidence does the author present in support of their thesis? What is the research design?
- What are the cases? Are the cases relevant?
- Is the evidence credible? Does the author present evidence about the mechanisms behind the main results?
- Could there be other reasons why we see those results, other than the author’s theory being right? How
does the author address those potential alternative explanations?
- Are the author’s conclusions reasonable, given the evidence?

Questions for reflection and discussion

 In what way is comparative politics (or political science more broadly) scientific?
 How can we analyze political phenomena with the scientific method?
 How can the experimental method be used to learn about politics?
 How can we draw valid inferences about politics when doing an experiment is not feasible or ethical?
 When and how can comparative politics be useful to society / citizens / policymakers / politicians?
 How do different paradigms or perspectives onto the political world differ? Are some preferrable to others?
 Is experimentation in political science feasible, ethical, and/or desirable? Under what circumstances?

Required readings

 Caramani. Comparative Politics. Oxford University Press, 2020. Chapter 2

 Clark, Golder, and Golder. Foundations of Comparative Politics. CQ Press, 2019. Chapter 2

Suggested materials for further exploration

 Alexander, Titus. “Can political science save democracy?” The Loop, March 2023.
https://theloop.ecpr.eu/can-political-science-save-democracy/
 Uchiyama, Yu. “Evidence-based policymaking in the US and the UK.” CEPR, March 2024.
https://cepr.org/voxeu/columns/evidence-based-policymaking-us-and-uk
 Phillips, Trisha. "Ethics of field experiments." Annual Review of Political Science 24.1 (2021): 277-300.
https://www.annualreviews.org/doi/pdf/10.1146/annurev-polisci-041719-101956

5
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

The nature, formation, and development of states

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

Defining the state

States are arguably the most important arena of contemporary politics. They shape political, social, and economic
development.

States have been defined in a variety of ways, and working through different conceptualizations of the state can help
us gain a better understanding of what they are, how they emerge, and what makes them successful. Some notable
definitions of the state include:
 Max Weber: a state is “a human community that (successfully) claims the monopoly of the legitimate use
of physical force within a given territory” (Weber 1918).
 Charles Tilly: states are “relatively centralized, differentiated organizations, the officials of which, more or
less, successfully claim control over the chief concentrated means of violence within a population inhabiting
a large contiguous territory” (Tilly 1985).
 Douglas North: a state is “an organization with a comparative advantage in violence, extending over a
geographic area whose boundaries are determined by its power to tax constituents” (North 1981).
While there are meaningful differences across these definitions, they all share a focus on states’ control over a
bounded territory and their reliance on the threat (or exercise) of physical violence. Coercion is at the center of what
states are and how they rule, and of politics more broadly.

Note how states can be defined independently of nations, which are groups who share a common identity (be it
linguistic, ethnic, or religious). States often build their own nations (for instance by homogenizing their territory
under a single language, religion, and/or collective understanding of the political community). In that sense, at least
in the cases of early state building, states created nations, more than the other way around.

The formation and development of states was enormously consequential for political development around the
world. States have enabled unparalleled levels of economic and social development, but also at the cost of erasing
forms of societal and linguistic diversity, displacing other forms of authority, and mobilizing populations for warfare.

Origins of the state

There are two big families of explanations to give account of the emergence of states: contractarian and predatory.
Contractarian views of the origins of states see them as emerging from a form of social contract or agreement among
individuals. Authors in this tradition include Thomas Hobbes, John Locke, and Jean-Jacques Rousseau. Hobbes for
example argued that individuals lived in the so-called state of nature, where life was “solitary, poor, nasty, brutish,
and short” (Hobbes 1651) because they were not constrained on their use of physical violence against each other to
survive and therefore people had no rational incentives to engage in productive activities. In Hobbes’ view, the
creation of the state solves this situation by establishing an all-powerful coercive entity (the Sovereign) through an
implicit social contract between individuals in the state of nature. Individuals would give up their natural rights in
exchange for protection from the sovereign. While contractarian theorists differ in how much they seek to limit the
authority of the state under civil society, all of them see the state as an entity that maintains order by punishing
(even with violence) individuals who deviate from the social contract. It is worth highlighting that Hobbes’ idea of
the state of nature can be compared to more contemporary theories that see the state as solving a coordination

1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

problem. Similar to other coordination problems, in the state of nature individuals’ payoffs or utility are lower than
they could be, but they are “stuck” in this low-level equilibrium for lack of solutions of their coordination problem.
Hobbes’ view of the social contract can be seen as the state acting as a coordination device, which through coercion
raises the costs of free riding and enables individuals to lead a longer and more productive live.

Predatory views of the state, by contrast, explain the emergence of states by problematizing the conflicts between
citizens and the state. In this tradition of thought, rulers use violence and other resources to extract resources from
citizens in order to ensure their existence particularly with regards to conflicts with other (foreign) rulers. The key
author in this tradition is Charles Tilly, who argued states can be seen as particularly successful organized crime, i.e.
as an extortion racket that provides security in exchange for protection from itself. That is, here the state is both a
guarantee of individuals’ security but also a key threat to it. Tilly focuses on medieval and early modern European
states, and analyzes how changes in military technology and conflict drove the organization and enlargement of
states by lords and kings. For Tilly, there is an intimate relationship between war making and state building; he
famously wrote that “war made the state and the state made war.” In his view, it was bargains between rulers and
the elites whose (military and economic) resources were necessary for state building that drove the political
development of states, which involved eliminating or neutralizing local rivals (partly through military organizations
and bureaucracies), developing institutions for security and conflict resolutions (police forces, courts), and building
capacity to extract resources (tax collection, account keeping). Ultimately, then, states in this view emerged not from
a social contract but as a by-product of rulers’ attempts to survive, and the bargains that ensued. States developed,
with time, by establishing institutions of “quasi-voluntary compliance” (Levi 1988) rather than outright predation,
which made it possible to extract more resources.

There are important differences across the contractarian and the predatory views of the states. Predatory views of
the state are more descriptive, as opposed to the more normative approach of contractarian theories. They are more
attuned to how political order is in fact constructed, and do a better job at explaining the emergence of self-
constrained states.

A recent study of rebel groups in Eastern Congo documents how the behavior of bandit groups resembles the
behavior of proto-states under conditions in which that makes it possible for them to extract more resources.1 The
study therefore provides empirical backing to predatory views of the emergence of states. In essence, the study
shows that in locations where informal gold mines exist, and as the price of gold goes up, these rebel groups establish
themselves as stationary bandits in the villages, providing security and other basic services, as a way to extract
resources from the miners and their families. In contrast, in locations where there is no gold but coltan (a mineral
that is much bulkier and therefore not easily concealed by miners, unlike gold), these rebel groups establish
themselves as stationary bandits in the mines, providing no services in the villages. The study also shows that the
establishment of these rebel groups in the villages increases welfare, despite there being no social contract or
agreement and these relationships being based in extraction and coercion.

State capacity and state building

Not all states are equally capable at maintaining order, protecting their borders, and ensuring social and economic
development. While we typically classify some states as “failed states”, in reality states are in a continuum of state
fragility. No state is free of challengers (internal and/or external), and no state exercises complete control within its
territory. States’ capacity to maintain order is so consequential that it has been argued that “the most important
political distinction among countries concerns not their form of government but their degree of government”
(Huntington 1968).

States developed over time in different realms, including their coercive and extractive capacities. State development
took place, somewhat in parallel, on different dimensions, including:

1De La Sierra, Raúl Sánchez. "On the origin of states: Stationary bandits and taxation in Eastern Congo." Journal of
Political Economy 128.1 (2020): 32-74.
2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

- The development of autonomous legal systems – including codified rules, a hierarchy among them, and
institutions to enforce them.
- The development of a bureaucracy – a centralized and specialized organization that could enforce order,
extract resources, and over time provide services (such as education or healthcare).
- The development of formal political institutions to organize conflict, enact policies, and produce legitimacy.
- The development of a more or less unified identity basis for the state’s population and territory, mostly
through nationhood (a shared collective understanding of the political community) and citizenship (a more
or less bounded definition of who belongs in that community). Benedict Anderson (1983) famously defined
nations as “imagined communities”, in that they socially construct a sense of collective belonging, largely
through narratives and often state-funded or state-promoted cultural efforts (through historiography,
schooling, literature, mass media, etc.). Nations are imagined communities also because they supersede
class or other inequalities, even though members of that political community do not know each other.
Therefore, for earlier cases of state formation and development (as in modern Europe), it can be said that
states created nations. In other, later cases of state formation, it is more appropriate to think of states being
built over preexisting national communities.

At least in the early cases of modern Europe, states can be seen as developing in three main phases.
- Consolidation of rule. In this phase, there are a decreasing number of rulers controlling increasingly large
territories with growing levels of homogeneity.
- Rationalization of rule. In this phase, there are changes in how state power is exercised, with growing levels
of centralization (with more power in the hands of state bodies and bureaucrats, to the detriment of local,
religious, or ethnic sources of authority), with more clear hierarchies (in laws, bureaucracies, and agents),
and more internal specialization and differentiation.
- Expansion of rule. Later, states can be seen as expanding the scope of their rule, for instance by intervening
in larger sectors of social and economic activity, providing a broader range of services, or regulating more
aspects of life.

State capacity is best seen as a multi-dimensional concept. A useful typology of different dimensions of state capacity
differentiates between:
- Coercive capacity.
- Extractive capacity.
- Coordination capacity.
- Compliance capacity.

Accordingly, state capacity can be measured through a variety of indicators. Most modern approaches combine
several indicators to build an index (such as the state fragility index). But it is useful to differentiate between output,
input, and legibility based indicators of state capacity. Legibility is the breadth, depth, and standardization of the
state's knowledge about its citizens and their activities. Legibility-based approaches to state capacity emphasize that
coercion, extraction, coordination and compliance all require making local actors and practices legible to the state
(i.e., visible and actionable in a rational or standardized manner). Empirical measures of legibility include the
existence of censuses or cadasters, strength of statistical agencies, and quality of government data

Key skill: Interpreting a regression table

Interpreting regression tables is a central skill in order to be able to engage with scientific articles (in political science
and across the sciences). We will see regression tables most if not all weeks in class, and there will be questions
asking you to interpret regression tables in the exams. That said, the emphasis in this class is not on the math or
stats, nor on the coding, but rather on engaging with and interpreting regression tables.

What is a regression?
Regression is a type of analysis to estimate how one or multiple variables (e.g., education, gender, wealth) relate to
one variable of interest (e.g., vote choice). Regression is different from, but intimately related to correlation.

3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Regression is not causation (but it can get us there in certain circumstances). There are many types of regression,
the most common one being linear regression (aka OLS, for ordinary least squares). Different types of regression
require different assumptions for us to "buy" the estimates we get.

What do regressions look like?

In linear regression, we typically estimate a model of the form Y
= α + βX1 + γX2 + ε. Y is what we want to explain. X1 (and X2,
potentially X3, X4, etc.) are variables whose association to Y we
want to measure. The Greek letters correspond to coefficients
we will be estimating with the regression. β, in particular will give
us the regression’s estimate of how X1 is associated to Y.
Regression tables typically present OLS estimates of the
association between each independent variable (e.g., X1) and the
dependent variable (Y), together with many other useful pieces
of information. On the right is an example of a regression table.

What are the key pieces of information in a regression table and

where do we find them?
• Dependent variable (y): The variable that the researcher is
trying
to explain (typically at the top of the table or column)
• Independent variable(s) (e.g. X1): The variables that the re-
searcher is using to explain the dependent variable (typically row
names)
• Coefficient (e.g. β): The estimate of the association between a
given independent variable and the dependent variable (main
number in the cells of the table)
• Standard error: The estimate of how certain we can be about
the
size of a given coefficient (typically under the coefficient, in brackets). Standard errors can be corrected through
different procedures (typically informed on the bottom of the table), for example
through clustering.
• p-value: The probability of obtaining an estimate so extreme if the effect was actually zero (typically tables use
stars by the coefficients to represent whether the p-value is under a certain threshold, and a legend below, e.g., ⋆p
< 0.05, ⋆ ⋆ p < 0.01, ⋆ ⋆ ⋆p < 0.001). If the p-value is below the threshold (and therefore there are stars), we tend to
believe the estimate and we say it is “statistically significant” (because we conclude the association is so strong that
is unlikely to be due to mere chance)
• N: number of observations that went into the model. Typically toward the bottom of the table

Where do regression coefficients come from?

OLS regression coefficients come from minimizing the sum of squared residuals, where the residual is the distance
between a given observation’s value on X and its predicted value for the outcome. Calculating OLS coefficients comes
down to an optimization problem. There are formulas one can use to estimate them. In practice, we simply run code
in software like R or Stata, where we tell the program what is the dependent variable and what are the independent
variables. You can see and interact with an online visualization of OLS regression here: https://setosa.io/ev/ordinary-
least-squares-regression/.

Are regression coefficients estimates of causal effects?

Generally speaking no, regression coefficients do not necessarily measure causal effects. Most often, they just
measure associations – how variation in one variable is associated to variation in the outcome, holding all other
controls constant. Unless we have an experiment or something approximating an experiment, we need to worry
about confounders.

4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Questions for reflection and discussion

 How do theories about the emergence of states relate to normative ideals? Which theories are more
normatively appealing, and why? Are they empirically accurate?
 What is the relationship between nations and states?
 How do law, compliance, and coercion contribute to the development of states?
 Are states positive forces in economic, social, and human development?
 How do processes of state formation differ across historical moments, regions of the world, and political
contexts?

Required readings

 Caramani. Comparative Politics. Oxford University Press, 2020. Chapter 4.

 Clark, Golder, and Golder. Foundations of Comparative Politics. CQ Press, 2019. Chapter 4.

Suggested materials for further exploration

 Video - Green, John. “War and nation building in Latin America”, CrashCourse YouTube channel (2015).
https://www.youtube.com/watch?v=v6xi8_7Fy6Y
 Short article - Miklaucic, Michael. "State-building 101: Hard lessons from Afghanistan." Armed Forces &
Society 49.4 (2023): 982-988. https://journals.sagepub.com/doi/full/10.1177/0095327X221088873

5
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Authoritarianism, types, and survival

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

Authoritarianism and political development

Oversimplifying and focusing on very high-level variation, a typical path of political development starts with state
building with an authoritarian form, and only later the development of democracy. In that sense, it is not surprising
that some of the most important thinkers of the state also provide insights about authoritarianism.
- Hobbes theory is, with regards to the nature of authoritarian governance, empirically misleading. In his
view, the ruler’s authority is unconstrained, he rules on his own, and there is full security, except in the
international arena where the ruler lives in a situation akin to the state of nature.
- Tilly, by contrast, suggests that state builders (who are initially authoritarian) partly depend on other actors,
and in fact limited government, concessions, and public service delivery emerge out of processes of state-
building. It is however not a social contract but rulers’ self-interest (and in particular their need for
extraction and self-preservation) that drives those processes.
- For Olson, anarchy or the absence of the state can be seen as “roving bandits” doing uncoordinated and
competitive theft. Anarchy destroys the incentive to invest and produce, leaving little for either the
population or the bandits. He sees dictatorship as a “stationary bandit”: a ruler who monopolizes and
rationalizes theft in the form of taxes. It is in the dictator's self-interest to provide peaceful order and public
goods to increase production. Bates sees political order as an equilibrium in which rulers (which we can
think of as "specialists in violence") choose to protect the creation of wealth rather than prey upon it, and
when private citizens choose to set weapons aside and to devote their time to production (and leisure).
State failure can thus be seen as a process in which rulers go from stationary to roving bandits.

In any case, the path from state building to authoritarianism to democracy is far from guaranteed. When we look at
the historical variation in political regimes we see that overall the number of democracies has increased – but not
monotonically. Especially when we weight countries by population, it is easy to see that most people today live
under some form of authoritarianism, and that proportion has been increasing in recent times.1

Defining authoritarianism

We must not define authoritarianism through the existence of elections, or even the existence of competitive
elections. A lot of authoritarian governments today do hold elections, and even competitive elections (at least
partially or at least for some offices). What then would be a more productive way to define authoritarianism? There
are many different definitions, none of them fully satisfactory. Lidstaet (2020) defines authoritarian regimes as those
“that have no turnover in power of the executive" – putting the emphasis on the importance of turnover of political
elites. If we focus on the path of political development, we could place authoritarianism in between anarchy (Olson’s
roving bandit, or Hobbes’ state of nature) and democracy. Olson’s figure of the stationary bandit may be another
way to define authoritarianism. If we do want to use an electoral criterion, we could define authoritarianism as the
absence of free and fair elections in which multiple parties compete for office, and/or the combination of that and
the absence of key freedoms (such as freedom of expression). A definition that is particularly rich because it focuses
on the nature of authoritarian politics is that there is “no independent authority has the power to enforce
agreements among key actors and violence is the ultimate arbiter of conflicts” (Svolik 2012).

1 You can find useful data visualizations about the evolution of political regimes across history, and the
classification of different countries, at https://ourworldindata.org/democracy.
1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Classifying authoritarian regimes

There is a wide variety of classifications of authoritarian regimes. Each classification emphasizes a certain contrast,
while making it harder to see others.
- Totalitarian vs non-totalitarian authoritarian regimes. This distinction was most emphasized after World
War II, when the trauma of Nazi, fascist, and Communist dictatorships was most visible and relevant. Under
totalitarianism, there is a more extensive control of society, a higher degree of social mobilization, and a
more extreme use of ideology than in non-totalitarian regimes.
- Personalist vs single-party vs military vs monarchical regimes. This classification emphasizes ways in which
authoritarian rulers are deposed from office. In a monarchical dictatorship (not to be confused with
monarchies under democracy), the head of government holds office on the basis of family kin, and the role
is typically inherited. In a military dictatorship, by contrast, the executive power is exercised by military
leaders. In civilian dictatorships, where the government is neither military nor monarchical, the regime can
be either based on a single party (or, alternatively, a dominant party), or it can be based more on a cult of
personality around the ruler.
- Closed vs electoral authoritarian regimes. Another approach to classifying regimes is to examine whether
they are closed (where there are less freedoms, less constraints on executive authority, and less
competition) or to some degree electoral, competitive, or so-called hybrid.
- An alternative, modern approach to classifying regimes is to consider them in a continuum from less to
more competitive, and to examine how they move in that dimension. Not all authoritarian regimes that are
electoral or competitive are equally so: while some are hegemonic electoral regimes (where there are
elections but no meaningful contestation), others are competitive authoritarian regimes where competition
is more substantive and the opposition gets to win real fractions of power. From this lens, we can see that
a growing number of authoritarian countries these days are autocratizing, i.e. moving in a direction towards
less competitive authoritarianism.

When we examine the fate of dictators across different types, we see descriptively that most dictators are succeeded
by another dictator of the same type. This is intimately related to the fact that dictators do not rule on their own –
they actually depend on a support coalition, and face internal threats from within their inner circle.
- Once we focus our attention on that challenge for authoritarian governance, it is easier to see why
monarchical dictatorships are particularly resilient – they experience less turnover, more stability, and in
some metrics more growth than other types of dictatorships. This is arguably related to how the
monarchical political culture allows these regimes to solve credible commitment problems with their
support coalitions, for instance through royal family rules and structures that enable monitoring and trust.
- By contrast, military dictatorships often suffer from the threats of factions within the military. That helps
explain why they tend to have shorter duration and are more likely to lead to democratic regimes – in the
face of factionalism and threats, military leaders may prefer to “return to the barracks” to protect the image
of the armed forces and to keep military unity.
- In dominant-party regimes, the political party plays a key role in the selection and discipline of political
elites. The role of single or dominant parties in structuring political competition helps explain why these
regimes are the second most durable after monarchical dictatorships. It also helps explain why they often
engage in electoral fraud even when they would win the election, since the party and its performance helps
keep the regime united and deter and divide the opposition.
- Finally, in personalist regimes where the leader builds a personality cult which may make it harder for
opposition to coordinate, since people are incentivized to lie about their level of support to avoid
repression. These regimes indeed tend to have higher levels of repression, combined with a smaller inner
circle around the leader.

Authoritarian regimes from the lens of selectorate theory

2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Selectorate theory offers an approach to examine political regimes, putting all of them in a continuum space rather
than discrete categories. Selectorate theory starts by assuming that all politicians are office seeking (i.e., motivated
by the prospect of accessing and keeping power, rather than by policy measures), that there is always a challenger
to the current ruler, and that all actors seek to maximize their (economic) utility. The key factors to describe a
political regime in selectorate theory are: 1) the size of the selectorate S (i.e., the people who participate in the
selection of the leader, be it through elections or some other method); 2) the size of the winning coalition W (i.e.,
the subset of the selectorate that the ruler needs to keep happy in order to stay in office); and 3) the ratio between
the size of the winning coalition and the size of the selectorate W/S (i.e., the loyalty norm).

The selectorate is largest in democracies, large in dominant-party regimes, and small in military and monarchical
dictatorships. The size of the winning coalition, which is also largest in democracies and smallest in monarchies and
military dictatorships, matters because it determines whether it is feasible to keep the winning coalition through
private goods alone or not. The ratio of the winning coalition to the selectorate matters because it encapsulates how
likely it is that a member of the selectorate is in the winning coalition, and therefore the calculus that members of
the winning coalition make when considering whether to defect from the current ruler. The larger W/S, the weaker
the loyalty norm to the current leader. When W/S is small and the loyalty norms are very strong, the leader is more
likely to engage in kleptocracy and corruption, and more likely to govern through private goods rather than public
goods. When both S and W/S are large, as in democracy, the ruler cannot afford to pocket that many resources and
they have an incentive to produce public goods (since W is large) and keep the economy performing well so they can
get more taxes to produce those public goods.

From this perspective, therefore, policy performance is best in democracies because both W and W/S are large,
worst in dominant-party and personalist dictatorships because W and W/S are small (and therefore rulers can survive
by distributing private goods to members of their winning coalition, and otherwise steal), and middling in monarchies
and military dictatorships because W is small but W/S is large (and therefore, with a weak loyalty norm, they need
to generate a good economic environment). From the perspective or selectorate theory, then, government
performance follows from institutions – all leaders are to govern well in settings of large W, large W/S to stay in
office, and to govern poorly in small W and small W/S settings.

The problems of authoritarian rule

Counter to the Hobbesian view, we know that dictators do not rule on their own, and that their rule is always to
some extent constrained by the need to keep their winning coalition happy and to prevent them from defecting to
an alternative. Svolik (2012) examines data on nonconstitutional exits from office (i.e., ends of mandate that are not
by death or simple succession) of all authoritarian leaders in the period 1946-2008 and finds that by far the most
common type is a coup. Popular uprisings, transitions, and foreign intervention are a lot less common ways for
authoritarian rulers’ mandates to end. This is consistent with the biggest threat for authoritarian leaders coming
from within their inner circle, and helps explain why authoritarian rulers can be so paranoid.

All authoritarian regimes face some fundamental problems and dilemmas. These are in a sense inescapable, but
leaders can choose how to address them. Certain types of authoritarian leaders or regimes are better equipped to
deal with some of these dilemmas than others.
- The problem of authoritarian control. Dictators face threats from the population over which they rule. The
two key tools at the disposal of the regime are repression and cooptation. Cooptation can take place
through handouts or private goods (such as rents, transfers, or jobs) or through institutions (such as a
legislature, allowing multi-party elections, or federalism). Cooptation is costly for the ruler, especially if it
takes the institutional form which allows for more credible commitment but constraints the regime.
Repression, on the other hand, is a double-edged sword because it empowers the security forces.
- The guardianship dilemma. A military strong enough to protect the regime against mass unrest and foreign
threats is also strong enough to overthrow it through coup. In fact, military leaders can use this leverage to
demand rents and policy concessions. This dilemma can be addressed in a variety of ways. Some autocrats
keep for themselves the role of minister of defense, and/or create parallel security forces for their own

3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

security, precisely because they do not trust the military. When they have ministers of defense, a common
strategy is to change them frequently to try and weaken the power of military challengers. But recent
research by Anne Meng suggests that rebel regimes (those that establish themselves after winning a
rebellion against a previous regime) are better able to address the guardianship dilemma. Thanks to the
history of interactions, relationships, and trust built during the rebellion war, rulers of rebel regimes are
better establish credible commitments and to deal with the guardianship dilemma. We see less coups and
less rotation of ministers of defense in these regimes.
- The power-sharing dilemma. Broadening elite incorporation mitigates prospects for outsider rebellions (by
either elites excluded from power or the masses), but it raises the risk of insider coups. The distribution of
rents (transfers, jobs, etc.) is not enough to address this dilemma, because dictators cannot credibly commit
to sustain this dilemma. Institutions such as political parties and legislatures help authoritarian regimes
solve the power-sharing dilemma, for two key reasons. First, these institutions provide members of the
support coalition information about the regime and enable them to influence decisions. Second, these
institutions act as credible commitments, allowing members of the support coalition to hold the ruler
accountable and allowing the ruler to credibly commit to a given arrangement. On the other hand, these
institutions can also help alleviate the regime’s informational deficit (driven by preference falsification) –
through parties, legislatures, and local governments for example it may be easier for rulers to get a sense
of the strength of the opposition. In any case, power-sharing is a double-edged sword because of the
countervailing commitment and threat-enhancing effects.

Elections in dictatorships

Many authoritarian governments organize elections, and they sometimes have meaningful levels of competition.
The organization of competitive elections is not necessarily a sign of authoritarian weakness, or of a slow transition
to democracy. In fact, elections can help these regimes be more resilient. Why? First, elections help them deter and
minimize challengers from within, sharing the spoils with groups with power and giving formal avenues for the
opposition to work. Second, to weaken challengers – by setting the rules of the game of competing for formal power,
contestation can help divide the opposition, fostering fragmentation. Third, to gather precious information about
the strength of the opposition and its distribution across the territory. And finally, to build legitimacy, both internal
and external. Elections, like other power sharing arrangements, are therefore not just window-dressing choices.

Questions for reflection and discussion

 How should we understand, and measure, countries’ political regimes or level of authoritarianism?
 Can we come up with one classifying criterion to determine whether a country is autocratic or not?
 How are authoritarian institutions of power-sharing different from democratic institutions?
 Has the nature of authoritarian governance changed over time? If so, how? How about variation across
regions of the world, or levels of GDP per capita?

Required readings

 Clark, Golder, and Golder. Foundations of Comparative Politics. CQ Press, 2019. Chapter 8.

Suggested materials for further exploration

 Podcast – Meng, Anne. “The autocrat’s gambit”. https://www.scopeconditionspodcast.com/episodes/the-

autocrats-gambit-with-anne-meng (also available on Spotify, Apple Podcasts, etc.)
 Video – “The rules for rulers.” https://www.youtube.com/watch?v=rStL7niR7gs (this is adapted from the
book by Bruce Bueno de Mesquita and Alastair Smith. The dictator's handbook: why bad behavior is almost
always good politics. Hachette UK, 2011.

4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Democracy

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

Defining democracy

When talking about democracy, we often refer, implicitly, to very different notions of democracy, such as direct
participation, representation, majority rule, fairness, empowerment, rule of law, good governance, or equity. It is
important to realize there are fundamental tensions between some of these ideas we associate to democracy. For
example, representation and direct participation come from very different traditions (one liberal, one democratic).
Similarly, majority rule can clash with rule of law and fundamental rights. We have come to associate all these
meanings to democracy and to assume they make a coherent set of ideas and institutions, but tensions remain.

If democracy is an amalgam of meanings, among which there are fundamental tensions, how do we make progress?
A first step is realizing that democracy is an essentially contested concept. These are concepts “the proper use of
which inevitably involves endless disputes about their proper uses on the part of their users” (Gallie 1956). Disputes
over the meaning of democracy are everywhere in the public debate (e.g., on whether a cabinet, or a policy, or a
leader is democratic or not). Those disputes are meaningful and consequential politically, since democracy is the
most important legitimizing signifier in contemporary politics. In fact, “democracy” is often used as an appraising
category. Critically, the meaning of democracy itself evolves through these political disputes over the boundaries of
the term.

In empirical political science, scholars have long been trying to clarify these debates by focusing on what democracy
actually is, as opposed to and in contrast with our normative and prescriptive ideas of what it should be. In that
effort, there are two large families or approaches to the conceptualization of democracy. First, procedural definitions
which focus on formal or informal processes, how elites access and exercise political power. Second, substantive
definitions which focus more on issues of equity, social rights, fundamental freedoms, etc. – therefore emphasizing
more the output side of democracy. Scholars often debate whether a procedural (or “minimalist”) definition is better
or not, and the kinds of political engineering that we should engage with to promote democracy.

There are three key authors that contributed very influential conceptualizations of democracy: Schumpeter, Dahl,
and Przeworski. All three seek to understand democracy as it actually is, and give us a way to operationalize it to
identify which countries are and are not democratic.
- Joseph Schumpeter, an Austrian economist, argued against the idea of democracy as “government by the
people for the people.” He argues that is a naïve understanding of democracy that does not represent well
the democracies that actually exist. For him, democracy is an institutional arrangement in which individuals
acquire the power to decide by means of a competitive struggle for people’s vote. The key role of voters,
then, is to produce a government. But voters need not be necessarily very involved or very informed. His
concept has often been referred to as elitist, but he is not making a normative argument rather a descriptive
one. A key implication of his argument is that the political boss is a key figure of democratic politics, since
they are the protagonists of that competitive struggle.
- Robert Dahl, an American political scientist, argued we needed to distinguish between the normative
concept of “democracy” and the descriptive concept of the “polyarchy. For him, democracy is “the
continuing complete or almost complete responsiveness of the government to the preferences of its
citizens, considered as political equals.” Under democracy, citizens must have the ability to formulate their
preferences, to signify preferences by individual and collective action, and to have their preferences
weighed equally in conduct of government without discrimination. Democracy is therefore an ideal type,

1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

something we aspire to but cannot credibly reach in modern societies. By contrast, what we have at best
are “polyarchies”, a separate term he introduces to describe actually existing political regimes that combine
high levels of contestation (or competition among different political options) and high levels of inclusiveness
(or enfranchisement of citizens into the political process). A key implication of Dahl’s argument is that
transitioning to democracy often involves movements in two directions – towards more contestation and
towards more inclusiveness. He argued that increasing contestation first and then inclusiveness would be
a more sustainable path to democracy.
- Adam Przeworski, a Polish political scientist, argued that “democracy is a system in which parties lose
elections.” His concept puts political turnover at the center: in democracy the incumbent cannot be certain
they will remain in office. That element of uncertainty is central therefore to democracy – there is
competition, but critically the outcomes of that competition are not known ex ante. A key puzzle he
highlights is, why would an incumbent party accept defeat and go to the opposition, given that they control
the bureaucracy and the security forces? For him, the answer has to do with political parties and with time
horizons. In democracy, he argues, political actors have longer time horizons: conflicts are never resolved
definitively, there is always a prospect of getting back to office in the near future, and political parties can
constrain candidates to protect their long-term chances of exercising power. With this concept of
democracy, therefore, political parties, and in particular their coercive nature, are key for democratic
governance, since they sustain this key moment of democratic politics – turnover.

Measuring democracy

Measuring democracy matters for a number of reasons. First, whether a country is democratic or not, or how it is
moving in a scale of democracy, sometimes influences policy decisions by governments, international organizations,
and non-state actors like firms. Second, monitoring the quality of democracy across space and time can aid
democracy promotion efforts. Last but no tleast, efforts at measuring democracy, or at ellucidating how different
measures work, can be helpful to clarify our own concepts of democracy.

Yet measuring democracy is hard. For one, democracy is an essentially contested concept. Besides, even if we agree
on a concept of democracy, it is necessarily multi-dimensional -- this makes it hard to operationalize it and to get a
score for it. To make things worse, important dimensions like political competition, fundamental rights, or rule of
law are themselves necessarily hard to measure. Finally, we lack clear aggregation rules for collapsing different
indicators into a single measure of democracy (e.g., what weight should each component have?).

To address these challenges, different measurement strategies used by different institutes adopt different
strategies. These often reflect different concepts of democracy, rely on different data sources, have different
coverages (of countries and/or years), and use different aggregating methods, and even different goals of the
measurement exercise (e.g., detecting small changes versus classifying countries in broad categories). In any case,
serious efforts at measuring democracy are transparent in their conceptualization, measurement, and aggregation
strategy.

The most important measures include Varieties of democracy (V-Dem), Freedom House (FH), Polity (Polity),
Economist Intelligence Unit Democracy Index (EIU), Regimes of the world, and the Bertelsmann Stiftung’s
Transformation Index. These measures differ in their conceptualization, operationalization, data sources,
aggregation methods, and data coverage. Eventually, they produce measures that are positively correlated but with
meaningful differences across them.

Transitions to democracy

There are multiple approaches to theorizing the origins of democracy. Modernization theories, which were popular
in the cold war period, see democratization as emerging from robust economic conditions. For civic culture theories,
citizen beliefs and attitudes drive democratization. Dissent theories, by contrast, see democracy as emerging from
elites’ ability to express and process dissent, to build and unbuild consensus. In economic bargaining theories,

2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

bargaining across (and violence among) organized economic interests; or certain economic endowments drive
democratization. Finally, group-based theories emphasize the role of ethnic and religious groups in the dynamics of
democratization. Each theoretical approach has with it important policy implications for how to promote
democratization.

In any case, explaining the origins of democracy is hard, for a variety of reasons. First, drivers of emergence may not
coincide with drivers of survival. A key implication of this is that we may need different theories for emergence and
survival. Second, there is a problem of confounders: any potential cause is likely correlated with variables that also
correlate with democracy. A key implication of this is that empirical tests of theories of democratization need to
address confounders, for instance through experiments, quasi-experiments, or at the very least by controlling
statistically for confounders. There is also an important issue of reverse causality, which can also dampen our ability
to learn about causal effects. That is because democracy can have a direct effect on many of its potential drivers
(e.g., the economy, ideas). Last but not least, different factors may matter differently at different places / moments
(i.e., different routes from democracy to dictatorship). A key implication of this is that theories may need scope
conditions or bounds on their application.

An influential theory of transitions to democracy is that of Acemoglu and Robinson. Like all good theories in the
social science, they simplify matters to highlight important dynamics and to derive empirically testable hypotheses.
They start by theorizing societies as split between elites (with economic endowments, and contrary to redistribution)
and citizens (poor and in favor of redistribution). This is therefore an economic, structuralist theory of transitions to
democracy. They argue that democratization is above all a reform that enfranchises citizens and therefore,
mechanically, increases redistribution. The key puzzle then is, why would elites ever favor democratization, since it
leads to more redistribution at their expense? Acemoglu and Robinson’s theory is that elites favor redistribution
when there is a threat of revolution from below, which would lead to higher costs than democracy. Democracy is a
way for citizens to lock in their power through formal institutions – with a revolutionary threat, elites cannot credibly
commit simply through transfers or rents, and democracy emerges. From this perspective, then, whether there is
democratization (and whether it is stable) depends on: civil society (costs of collective action and repression);
economic shocks and crises (affecting economic power, coups); wealth inequality (affects probability of revolution,
and the elite's fear of democracy); types of wealth (e.g., land is not mobile).

Democratic backsliding

Democratic backsliding is “the state-led debilitation or elimination of any of the political institutions that sustain an
existing democracy” (Bermeo 2016). Backsliding is at the core of contemporary debates about democracy, because
we see many cases of backsliding around the globe, including in wealthy countries that used to be considered
consolidated democracies. When examining threats to democracy, it is important to realize that these have changed
over the past few decades. In contrast to the cold war era, nowadays we see less military coups and less executive
coups (so-called "auto-golpes") and less election-day fraud. Yet we see more strategic harassment and
disqualification of opposition leaders, and more acts of what is called executive self-aggrandizement.

Therefore, the nature of the threats to the survival of democracy have fundamentally changed. Nowadays, episodes
of backsliding may be less dramatic, violent, or sudden -- but they can still lead to meaningful changes in a country's
political regime and fundamental freedoms. A key implication of this is that identifying backsliding (and resisting it)
is harder when it is more gradual and happens in a more limited set of institutions. Critically, wanna-be authoritarian
leaders likely know that, and adjust their strategies accordingly.

In an influential book, How democracies die, Levitsky and Ziblatt argue that autocratic leaders can emerge in ways
that make them appear manageable – e.g., a populist outsider. Democracy, they argue, is often undone
progressively, with a gradual escalation of actions and reactions. The key to not letting autocratic leaders erode
democracy is in elite gatekeepers- Parties for example can purge, incorporate, or militate against autocratic
challengers and their supporters. Law and culture are key tools in protecting democracy, and in weakening
autocracy.

3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Studying democratic backsliding is challenging because 1) measuring gradual changes in levels of democracy is
challenging, especially without using expert surveys which may have bias, and 2) there are serious problems of
confounders and reverse causality. A recent study of democratic backsliding in US states, published in the APSR in
2023, makes progress on both fronts. 1 The author, Jacob Grumbach, starts by building an electoral democracy index
for all US states in every year in the period between 2000 and 2018, based on objective data about barriers to and
enhancers of electoral democracy, such as gerrymandering, provisions for who gets to vote and how, and voter
registration. Using that data, the author documents wide variation across states and across time. Next, he uses
regressions controlling for both state and year fixed effects (which account for any state-specific or time-specific
confounders), to examine how a state’s levels of competition, polarization, and Republican control relate to its
electoral democracy score. He finds a robust association between a state being controlled by Republicans (a
Republican governor with a Republican majority in the legislature) and a decline in the state’s democracy score.
Exploiting the variation across time and space, he uses two quasi-experimental designs and consistently finds that
Republican control of government causes a decline in the democracy score. This is meaningful because it suggests
that political parties can be key drivers of democratic backsliding, and puts the focus on how parties can resist or
enhance these dynamics.

Questions for reflection and discussion

 How do you understand democracy? How do street-level understandings of democracy differ from those
of empirical political science?
 What are the main pros and cons of different measures of democracy? Is it possible to have one single,
universally valid measure of democracy?
 What explains whether a country transitions to democracy, and whether that transition is sustained or it
reverts back to authoritarianism? Do successful explanations differ for different regions or historical
periods?
 What are the main threats to democratic backsliding, or the erosion of democracy? Does our answer to that
question depend on how we conceptualize democracy?

Required readings

 Caramani. Comparative Politics. Oxford University Press, 2020. Chapter 5.

 Clark, Golder, and Golder. Foundations of Comparative Politics. CQ Press, 2019. Chapter 7.

Suggested materials for further exploration

 Web visualization – “Democracy data explorer.” https://ourworldindata.org/explorers/democracy

 Web post – “Democracy data: How sources differ and when to use which one.”
https://ourworldindata.org/democracies-measurement
 Policy report – Laebens, Melis G. "Beyond Democratic Backsliding: Executive Aggrandizement and its
Outcomes." (2023). https://www.v-dem.net/media/publications/UWP_54.pdf

1Grumbach, Jacob M. "Laboratories of democratic backsliding." American Political Science Review 117.3 (2023):
967-984. https://doi.org/10.1017/S0003055422000934
4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Parliamentarism and presidentialism

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

Why forms of government differ

If competitiveness, inclusion, and turnover are central to democratic governance (as per the conceptualizations of
Schumpeter, Dahl, and Przeworski), why do we observe fundamentally different ways of organizing government
under democracy? There are in essence three main different schemas for organizing democracy, which we call forms
of government: parliamentarism, presidentialism, and semi-presidentialism. There are nuances and varieties within
each of these, but they constitute 3 fundamentally different approaches to how elites get to executive office, how
they get removed from office, and how voters participate in the process.

Forms of government matter because they shape competition and contestation, access to executive power, and
political turnover in fundamentally different ways. Democratic politics therefore take different dynamics in different
forms of government. And research suggests that these different forms of government also have important
implications for economic, social, and human development. In any case, they involve different choices and there are
trade-offs among them. That is perhaps why we do not observe one of these forms of government gaining
precedence over time – if anything, we see a roughly equal split nowadays. While parliamentarism was the original
form of government, emerging out of processes of state building described by Tilly, presidentialism emerged with
the American revolution and rapidly spread across newly independent countries. Semi-presidentialism, in turn,
emerged in the inter-war period in Europe and then rapidly took off with adoptions in Europe and in sub-Saharan
Africa.

Clarifying terms and dimensions

When thinking about forms of government, it is essential not to confuse the political regime (and subtypes of
dictatorships), whether the state is a monarchy or not, and whether the legislature is bicameral or not, with the
state’s form of government, that is whether it is a parliamentary, presidential, or semi-presidential system. It is also
important not to be misled by official titles, which sometimes misname the nature of a political office or a form of
government. For example, heads of government in Spain and Italy are called "presidents", even though these are
parliamentary systems.

There are key terms that is important to define and to be comfortable with:
- Head of state: highest representative of a state, who may be referred to by a variety of titles (president,
king or queen, emperor or empress, etc.)
- Head of government: leader of the administration of the state (may or may not coincide with the head of
state), who may also be referred to by a variety of titles (e.g., prime minister, president, chancellor, etc.).
- Legislative responsibility: a situation in which a legislative majority has the constitutional power to remove
a government from office without cause.
- A formateur is the person designated to form the government in a parliamentary democracy, and is often
the head of government designate. This is the person who seeks the support of parliament to become head
of government. In a presidential system, the figure of the formateur is irrelevant – the president elect is de
facto the formateur, and gets to form their own government.
- An investiture vote is the formal vote in which many parliamentary democracies designate a head of
government after the legislative elections.

1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

- A vote of no confidence is initiated by the legislature in parliamentary or semi-presidential systems – the

government must resign if it fails to obtain a legislative majority
- A constructive vote of no confidence must indicate who will replace the government if the vote succeeds.
Whether the vote of no confidence must be constructive or not makes a big difference, since it will generally
be easier for a majority of legislators to agree on whether to remove the current head of government than
to agree on who should replace them.
- A vote of confidence is initiated by the government in a parliamentary system – the government must resign
if it fails to obtain a legislative majority.
- An impeachment is a constitutional process through which a head of state or government can be charged
with wrongdoing and potentially removed from office in a presidential or semi-presidential system by a
legislature acting similarly to a court.

Classifying forms of government

The key to identify what form of government a country has lies in responsibility to the legislature and, if relevant, to
the president. A democracy where the government is not responsible to the legislature, as defined above, is
presidential. Where the government is responsible to the legislature it is either parliamentary or semi-presidential.
For a country to be semi-presidential, it needs to have an independently elected president to which the government
is also responsible. Note that parliamentary systems can also have a president (if they are a republic), an even a
popularly elected president.

Fundamental differences

At the basic level, parliamentarism is a system of “fusion of powers” or “mutual dependence”, because there is no
separate democratic legitimacy for the legislative and executive powers. In parliamentary democracies, voters elect
legislators who in turn elect the head of government (explicitly where there is an investiture vote, implicitly when
there is not). In any case, the resulting government depends on legislative support for its survival, since a majority
of legislators can end the mandate of the head of government (and with her, her cabinet) with a simple vote of no
confidence. Critical to the politics of parliamentary systems is that the head of government can call for a snap
election at any time, which ends the mandate of the government as well as the legislature’s. Therefore, while there
are deadlines on the latest date by which an election needs to be called (typically, 4 years after the previous one),
the electoral schedule is far from fixed and heads of government use their discretion strategically.

By contrast, presidentialism is at the basic level a system of “separation of powers” because voters elect separately
a head of government (in all cases except the US, they do so directly) and legislators. Legislators cannot end the
mandate of the president through a vote of no confidence precisely because the president is popularly elected in a
separate election. They can however end the president’s mandate through an impeachment. Impeachments are
fundamentally different from votes of no confidence: they require making appeal to legal wrongdoing (in fact they
often follow procedures akin to a trial), and if successful they remove the officeholder but not the government,
typically if the president is impeached in a presidential system their vice-president becomes president. In
presidential democracies, then, elections are for a fixed term. Similarly, presidents cannot end the mandate of the
legislators – they do not have the ability to call for a snap election.

Beyond fundamentals: Dependencies in practice

The basic view of forms of government oversimplifies and misleads, largely because they ignore the role of parties
and of party systems.

For one, the basic view portrays presidents as too autonomous from the legislature. In practice, presidents in
presidentialist systems still depend on the legislature to pass legislation and to avoid being impeached. That is why
we often observe coalition governments also in presidentialist democracies, especially when they use PR electoral
systems and have a fragmented party system.

2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Second, the basic view of forms of government portrays legislatures or parliaments as sovereign in parliamentary
democracies. Yet legislatures often depend on the government. For once, the head of government’s ability to call
for a snap election is an extremely powerful tool they have at their disposal to manage the timing of the political
cycle and opposition in the legislature. Second, the head of government is typically the leader of one of the most
important parties. Where parties are strong (especially in closed list PR systems), legislators of the party of the head
of government are likely to follow their policy line since their job depends on their obedience.

More generally, it is important to realize that heads of government in both parliamentarism and presidentialism
need to build and sustain legislative coalitions for their survival (or at least their legislative success). Depending on
the party system, which in turn depends on the electoral system, we may see coalition governments --and coalitional
dynamics-- in both parliamentarism and presidentialism. The dynamics of coalitions differ systematically however
because heads of government in parliamentary systems depend more directly on having a legislative majority. The
strength of parties as organizations, and internal competition within the parties can reshape the dynamics of both
parliamentarism and presidentialism.

Cabinet formation and cabinet types

In presidential democracies, the president can form their own cabinet without explicit approval of the legislature.
However, when the legislature is fragmented and the president does not command a majority of seats, and/or does
not have strong command over the party, we are more likely to see the formation of coalition governments. In
parliamentary systems the head of government needs the support of a legislative coalition to get the job. It is
important to note however that the coalition in the legislature (the one supporting that leader becoming head of
government) need not coincide with the coalition in the cabinet. It is not uncommon for parties that support the
government in the legislature to not be represented in the cabinet.

We often observe cabinet types that go counter the single-party majority government that one would expect in a
simple democratic system. All these cabinet types are more easily explained when we consider that politicians are
both office- and policy-seeking, and that they compete in an ideological space where not all coalitions are equally
likely.
- A single-party majority government comprises a single party that controls a majority of the legislative seats.
- A minimal winning coalition (MWC) is one in which there are no parties that are not required to control a
legislative majority.
- A single-party minority government comprises a single party that does not command a majority of the
legislative seats.
- A minority coalition government comprises multiple governmental parties that do not together command
a majority of the legislative seats.
- A surplus majority government comprises more parties than are strictly necessary to control a majority of
the legislative seats.

The government formation can get long in parliamentary systems – it depends largely on the fragmentation of the
legislature, inter- and intra-party competition, and bargaining among party leaders. Although note that it often takes
time for heads of government to take office under presidentialism, too -- for different reasons. Presidentialist
systems often have a fixed schedule for the elected president to start their mandate, and that period is often long.
However contrary to parliamentary systems, in presidential systems who the head of government will be is clear as
soon as the votes are counted. In parliamentary systems that is often not clear after election night because who the
head of government will be depends on bargaining across parties and their leaders.

Which form of government is better?

We can use different criteria to assess forms of government, or other aspects of a political system, including
government accountability (clarity of responsibility, electoral accountability, certainty in the political process),

3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

political stability (cabinet stability, democratic backsliding, democratic survival), economic performance (economic
growth, inflation), human development, or corruption. As usual with institutional choices, there are important trade-
offs to consider*. If presidentialism generates more clarity of responsibility and more direct link between voters’
expressed preferences in elections and who gets to exercise power, it appears to come at the cost of worse economic
performance and a higher risk of reversal to authoritarianism.

To understand why, it is useful to consider Juan Linz’s famous 1990 essay , “The perils of presidentialism.” In it, Linz
gives us several theoretical reasons why presidentialism as a form of government is dangerous for the survival of
democracy. In presidentialism, he argues, the head of government has a "strong claim to democratic, even
plebiscitarian, legitimacy", and there may be clashes in claims of democratic legitimacy between the executive and
the legislative powers. Parliamentarism allows to deal with conflicts by going back to the polls (votes of no
confidence, snap elections), which may serve as an escape valve for tensions in the political system. By having a fixed
electoral schedule, presidentalist systems do not have that escape valve and may see political conflicts across
branches of powers escalate more easily. By using fixed terms, presidentalism enhances a view of politics as
demarcated by discrete periods, "leaving no room for the continuous readjustments that events may demand."
Presidentalism operates in the logic of winner-takes-all politics, and thus tends to make politics a zero-sum-game.
Finally, in presidentalism, the head of government is more likely to feel they are the representative of "the people".
This may lead to authoritarian temptations when clashes with the legislature emerge. More recent versions of this
argument emphasize that the danger emerges when presidentialism is combined with a fragmented party system,
which provokes deadlocks and clashes of legitimacy that presidentialism cannot easily solve.

Those same reasons (winner takes all politics, clashes of legitimacy, fixed political calendar without flexibility…) may
explain why presidentialist democracies appear to perform worse in indicators of economic and human
development. At least in observational regressions, including when we control for some confounders like region or
level of GDP, we see statistically significant and negative associations between whether a country is presidential and
its performance in indicators of economic, human, and social development. These findings are however not
definitive evidence of the perils of presidentialism because we cannot control for all possible confounders. Since it
is rare for countries to switch their form of government, and reforms are far from exogenous or randomly assigned,
we cannot know for sure whether the observed correlations are causal.

Despite the advantages, there are also problems with parliamentary democracies. First, they introduce distance
between voters and government, because voters do not directly elect the head of government and vote only for
legislators. Second, they diffuse clarity of responsibility – at least in situations of a fragmented party system and
coalition governments, it is easier for parties to engage in blame shifting and blame avoidance. Because of that,
coalition politics in parliamentary systems can undermine people’s interest in politics and even the legitimacy of the
system. Unlike presidential democracies, parliamentary democracies experience many government turnovers that
are not directly caused by elections, and rather driven by bargains across parties and inter- and intra-party
competition after the elections are held. Finally, government transitions can get very long and very uncertain in
parliamentary systems when the party system is very fragmented.

Questions for reflection and discussion

 What form of government is better? What are the most important trade-offs when choosing whether a
country should be parliamentary, presidential, or semi-presidential?
 How do the dynamics of democratic competition and turnover, under parliamentarism and semi-
presidentialism, depend on the fragmentation of the party system, and on the strength of political parties?
 Does semi-presidentialism combine “the best of both worlds”, or does it create new problems?

Required readings

 Clark, Golder, and Golder. Foundations of Comparative Politics. CQ Press, 2019. Chapter 10.

4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Suggested materials for further exploration

 Podcast – “Parliamentarism vs presidentialism.” https://www.youtube.com/watch?v=gHtsbcGW4Wo

(also available on Spotify and Apple Podcasts)
 Essay – Linz, Juan J. "The perils of presidentialism." Journal of democracy 1.1 (1990): 51-69.
https://muse.jhu.edu/pub/1/article/225694

5
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Elections and electoral systems

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

The relevance of elections

Elections play a key role in transforming the expression of citizen preferences into political offices, that is, the legal
and political authority to rule. In democracies, elections play both a practical role and a symbolic one – they help
decide who takes office, while at the same time legitimizing the system. At the same time, it is important to realize
that a large (and growing) number of authoritarian governments also use elections, although elections can serve
different purposes in these settings, including the cooptation of elites and social groups and the collection of
information about the strength of the opposition. Only a handful of authoritarian governments today do not hold
elections.

Defining and understanding electoral systems

An electoral system is the set of rules that structure how votes are cast at election and how they are then converted
into the allocation of offices. Legislative, executive, and sometimes even judiciary and bureaucratic offices can be
allocated through elections. It is important to realize that a given country typically uses a variety of electoral systems,
since different offices (e.g., head of state, head of government, legislator, mayor, governor, etc.) may have different
electoral systems.

Electoral systems are best thought of as a combined choice from multiple choice sets, or along different dimensions.
The most important ones are explained below, but there are other dimensions of electoral systems that may be
important, including whether and how absentee voting is allowed, or whether voting is done with paper ballots or
with electronic machines.

Direct vs indirect voting

While, given our strong democratic norms, we typically think of elections as direct, many important political offices
are allocated through indirect elections. These include the head of government in parliamentary and semi-
presidential systems, the head of state in countries like India, Germany, Pakistan, and Estonia, the president of the
United States, and members of the upper house in countries such as France, India, Morocco, South Africa. Some
legislative offices are not even elected, rather they are designated – e.g. members of the upper house in Germany,
and some members of the upper house in Spain. There are several reasons why indirect elections persist nowadays,
but they often follow from a logic of parliamentarism and/or federalism.

Electoral formula

The term electoral formula refers to how votes are converted into offices. There are two big families of electoral
formulas: majority or majoritarian electoral systems, on the one hand, and proportional representation systems
(aka, PR systems), on the other. A third family or type of electoral systems, called mixed or hybrid, combine elements
of majoritarian and proportional representation systems (for instance by electing one fraction of legislators with a
system and another fraction with another system).

In majority representation systems, the candidates with the most votes are elected. These systems have as their
main advantages the clarity and simplicity of the formula and the fact that they generate clarity of responsibility.

1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Their main disadvantages is that they “waste” a lot of votes (all those who do not support the winning candidate),
failing to give them representation, and may thus generate a winner-takes-all political culture. Within majority
electoral formulas, there are two big types: plurality systems and absolute majority systems.
- Plurality systems, where the winning candidate need not obtain more than half of the votes. An example of
these is the system known as single-member first-past-the-post (or single-member district plurality, SMDP),
commonly used in the United Kingdom and many of its former colonies: here electoral races are for one seat
only, and the candidate with most votes wins it.
- Absolute majority systems force the winning candidate to have more than 50% of the votes. This can be
implemented with a variety of systems, most typically a two-round electoral system (TRS, where there is a
second round election among the top two candidates in the first round, if none of them achieves more than
50% of the votes) or with alternative vote (AV, where voters rank candidates and weaker candidates are
sequentially disqualified, giving that voter’s choice to the candidate they ranked next – this system is also called
an “instant-runoff”). These systems incentivize parties and candidates to win over not just their core supporters
but also the second-preferences of voters who may have other preferred candidates.

Proportional representation (or PR) systems, in contrast to majoritarian systems, distribute votes approximately in
proportion to votes. This family of electoral formulas seek to distribute votes not just to one winner but to several,
roughly in proportion to their electoral strength. It is important to note however that no electoral system is
completely proportional, since elections distribute a finite (and typically small) number of offices or seats, and can
only distribute seats in whole numbers (i.e., a party may get 0, 1, or 2 seats, but not 1.5 seats). The main advantages
of PR systems are that they give political representation to a wider range of options, which produces a more
complete representation of citizen preferences and may foster a sense of inclusion in a larger share of the
population. This is particularly important in societies where ethnic, religious, or other societal divisions may make
winner-takes-all politics dangerous. The main disadvantage of PR systems is that they diffuse clarity of responsibility,
since they tend to produce a more fragmented party system, which may lead to coalition governments and therefore
political turnover depending not just on election results but also, to varying degrees, on negotiations across parties.
There are two large subtypes of PR systems:
- Quota systems distribute seats through a more intuitive procedure in three steps. First, a quota is calculated
using the formula Q = V / (M + n), where Q is the quota, V is the total number of votes cast in the district, M is
the number of seats to be distributed, and n is a modifier. Different quota systems differ by their modifier, which
equals 0 in the simples system of the Hare quota. Second, each party’s total number of votes is divided by the
quota, producing numbers equal to the total number of seats a party would be able to afford given the quota.
Then, “automatic” seats are distributed to each party in a number equal to their quotas rounded down (such
that we do not distribute more seats than there are). Third, we distribute the “remainder” seats according to
the remaining quotas left from each party. To do so, we first calculate the number of remaining seats by taking
the total number of seats and subtracting the number of seats distributed in the automatic phase. Then we
calculate the remaining “quotas” left for each party by taking their total quotas from step 2 and subtracting the
number of seats they got. These remainders will range between 0 and 1. Then, we distribute the remaining seats
to the parties with the largest remainders. The final total number of seats for each party is equal to the seats
they obtained in the “automatic” phase and those obtained in the “remainder” phase.
- Divisor systems distribute seats according to a less intuitive procedure that produces less proportional results,
i.e. tends to give less seats to smaller parties and more seats to the larger parties. Different divisor systems differ
in the sequence of numbers that they use; the simplest divisor system is called d’Hondt and uses the sequence
{1, 2, 3, 4, 5, …}. All divisor systems proceed as follows. First, the total number of votes of each party is divided
by the sequence of numbers corresponding to that divisor system, up until the number is above the total number
of seats to be distributed. Then, seats are given to the cells in the resulting matrix with the highest numbers,
until we run out of seats to be distributed. The total number of seats for each party corresponds to the total
number of cells, corresponding to the vote count divided by the sequence of numbers, that got seats.

District magnitude

2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

District magnitude refers to the number of candidates to be elected (or seats to be distributed) in a given
constituency. The key distinction in this component of electoral systems is whether districts are single-member or
multi-member. Single-member districts are winner-takes-all, which has both disadvantages (e.g. misrepresentation,
votes waste) and advantages (clarity of responsibility, constituency service, some moderation). Multi-member
districts can be majoritarian (such as in single nontransferable vote or SNTV, where several candidates are elected
with a majoritarian formula) or more frequently proportional. In any case, multi-member districts give
representation to a larger number of candidates. The magnitude of the district the most direct lever by which an
electoral system can produce more proportional results – other things being equal, larger districts (i.e., districts
electing more candidates) lead to a more proportional distribution of seats.

Ballot structure

The ballot structure refers to how the ballot is presented to voters. When ballots let voters choose multiple
candidates, the key distinction is whether they are categorical (i.e., asking voters to choose candidates) or ordinal
(i.e., asking voters to choose and rank candidates). Ordinal ballots are appealing because they allow us to collect
richer data on voter preferences, but they can be cognitively cumbersome for voters and may take significantly
longer to count, especially if they are on paper and there are many candidates in the race. It is important to make
ballots easily digestible to voters in a way that avoids confusion and minimizes invalid votes.

Electoral threshold

The electoral threshold refers to the minimum performance a party needs to reach in order to participate in the
distribution of seats. The key distinction here is between natural vs explicit electoral thresholds. A natural threshold
applies when no explicit threshold applies, and equals 1 divided by the number of seats to be distributed. Many
countries establish higher explicit electoral thresholds (e.g., 3% in XX, 5% in XX). These explicit thresholds constrain
the representation of smaller political parties, or parties whose support is concentrated in a small part of the
territory. This constrains the diversity of views represented in the legislature, and may exclude from the legislature
parties that are only strong in a small fraction of the territory because they represent territorially concentrated
ethnic or religious minorities. On the other hand, electoral thresholds limit the fragmentation of the legislature and
the party system, which contributes to making government formation easier and coalition politics less complicated.

List openness

Electoral systems using open lists allow voters to select or rank candidates from within party lists. This is in contrast
to the more common close party lists, where voters can choose a list but the order of candidates is determined by
the party according to internal procedures. On the other extremes, free party lists allow voters to vote for candidates
from not just one but several different parties. Open list systems give more choice to voters, but they also weaken
party organizations and promote open factionalism through within-party competition. By weakening parties, they
may generate instability in the party system and may hinder processes of government formation, coalition building,
and law making.

Franchise

The franchise refers to the legal limits on who is and is not allowed to vote in an election. Franchise reforms refers
to legal reforms to extend (or contract) who is allowed to vote. Historically, the franchise was typically restricted
along the lines of gender and class, with only wealthy men being allowed to vote. Over time, the franchise was
extended and nowadays democracies typically allow all citizens aged 18 and above to vote. That said, debates about
the franchise and franchise extensions continue to exist in consolidated democracies. For instance, a growing
number of countries allow citizens aged 16 and 17 to vote, which may have consequences for the kind of issues that
are debated in campaigns and the kinds of policies that make it into the agenda. Another type of contemporary
franchise extension concerns immigrants – a growing number of countries allow non-citizens who are legally residing
in the country to vote, sometimes only in local elections but sometimes in elections at all levels.

3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Compulsory vs voluntary voting

While voting is generally voluntary, some countries make it compulsory. The rationale of compulsory voting rules is
that they may help socialize all citizens into the democratic process (including citizens from lower socioeconomic
backgrounds, which tend to have lower levels of turnout) and thus consolidate democracy, and may reduce the cost
of campaigning and the biases of electoral finances, since parties would not need to convince their supporters to
show up to vote. In practice, countries differ in how they operationalize compulsory voting – often it comes down
to small fines, which people from higher socioeconomic status are more sensitive to.

The consequences of electoral systems

Electoral systems are tremendously consequential for a country’s politics, governance, and stability. First, electoral
systems are a key driver of the fragmentation of the party system. According to Duverger’s law, first-past-the-post
single-member systems lead to a 2-party system, whereas more proportional PR systems with lower thresholds lead
to party systems with a larger effective number of parties. Second, electoral systems are a key contributor to the
strength and stability of political parties, which is essential for political competition to be legible to voters, and for
the bargaining processes after elections (be it to elect a head of state, or to pass legislation) to be manageable. Third,
the electoral system can shape voter political behavior, including whether people vote (with PR systems generally
producing higher levels of turnout) and how they vote (with PR systems generally leading to less “strategic” voting).
Last but not least, electoral systems can influence political selection, i.e. what kinds of leaders run and make it to
political office.

Evaluating electoral systems

It is not possible to unambiguously evaluate electoral systems, as the design of an electoral system often comes with
important trade-offs. First, electoral systems differ in the extent to which they encourage a direct relationship
between voters and their representatives: on one extreme, single-member districts with majoritarian systems
encourage it, while multi-member PR systems discourage or diffuse that relationship. Second, electoral systems
differ in the extent to which they produce an accurate representation of voter preferences. Systems with larger
districts, more proportional rules, and lower thresholds encourage such accuracy, while the opposite is true for
systems that have less proportionality and/or higher thresholds. Third, electoral systems differ in the clarity of
responsibility that they produce. If majoritarian systems, especially in single-member districts, tend to produce
clarity of political responsibility, such clarity is diffused in more proportional systems that tend to produce a more
fragmented party system, especially when such fragmentation causes governments to depend on forming and
sustaining coalitions. Fourth, electoral systems differ in the complexity of voting and of counting. More complex
procedures may be required to gather more data on voter preferences (e.g., their ranking of different options), which
may come at the cost of longer counting periods and in some cases a lower trust in the process. Last but not least,
there is a trade-off between allowing more voter choice in ranking party candidates, which may encourage turnout
and intra-party democracy, and on the other hand strengthening party organizations and encouraging stable parties.

Empirically evaluating the effects of different electoral systems is challenging because around the globe electoral
reforms are rare and not randomized. In fact, many variables may act as confounders in the relationship between a
given element of an electoral system (e.g., the district magnitude, the electoral formula, or the threshold) and a key
dependent variable of interest (such as the fragmentation of the party system, the stability of cabinets, or the
presence of underrepresented minorities in the legislature). There is also a major problem of reverse causality, since
electoral systems are chosen by political elites (typically in bargains at the moment of independence, transition to
democracy, or major reforms), so the outcome may have a causal effect on the independent variable. To learn
scientifically, therefore, it is useful to rely on so-called quasi-experiments, i.e., research designs that leverage
naturally occurring variation in a given setting to approximate an experimental design. As an example, recent studies
have exploited variation across Spanish municipalities’ electoral systems to measure their effects on turnout and on
the representation of women. In Spain, municipalities of between 101 and 249 residents elect city councilors by

4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

open list plurality at-large, whereas municipalities of 250 residents or more elect them through closed-list PR system.
By examining what happens in municipalities right below and right above the 250 threshold, we can get credible
causal estimates of the effect of open vs closed lists on outcomes of interest. Using this design, a recent study found
that in Spanish municipal elections the closed list system increases the share of female candidates, of female
councilors, and of female mayors in substantively and statistically significant numbers. Another recent study, on the
other hand, found that open lists increase voter turnout and the number of parties in the race. These two sets of
results illustrate the trade-offs that the design of electoral systems often face.

Referenda

Other than elections to select executive, legislative, or even judicial or bureaucratic officials, referendums or
referenda are an important form of electoral participation in some countries. In fact, referenda are increasingly
common in democracies, although they are still rare. Across countries, there is wide variation in the use of referenda
– some have never held a national-level referendum, while others (famously, Switzerland) use them regularly. In
some countries, laws mandate referenda on certain issues (e.g., EU treaties). Referenda can be justified according
to process-related arguments (e.g., there is an intrinsic value in collecting direct voter input, and in socializing voters
in the process of policy making) as well as outcome-related arguments (e.g., decisions taken in line with referendum
results may be more legitimate and better implemented). On the other hand, critics point to how referendum results
may be driven by the vocal opinions of the few (or the low participation of the many), and that it may be used by
political leaders or groups to give democratic legitimacy to a decision that hinders the rights of minorities.

Electoral integrity

Electoral integrity refers to extent to which elections are held according to international standards. Violations of
electoral integrity (e.g., vote buying, electoral violence, voter intimidation, obstacles for registering to vote or for
casting votes, unfair campaign finance regulations, etc.) exist to some degree in all democracies, although the nature
and gravity of violations varies across contexts. Electoral integrity is important not only because violations can alter
who gets to office, and the incentives of those seeking to gain office, but also because they hurt citizens’ trust in the
political process, depress electoral participation, and may exacerbate societal conflicts.

Questions for reflection and discussion

 In what sense are majoritarian electoral systems more democratic? In what sense are proportional
representation systems more democratic?
 What are the key trade-offs in the design of electoral systems?
 Should electoral systems be stable across time, or should we promote experimentation with electoral rules?
 How can improvements in the electoral system be promoted, given that reforms need to be approved by
the political parties that the current system produced?

Required readings

 Caramani. Comparative Politics. Oxford University Press, 2020. Chapter 10.

 Clark, Golder, and Golder. Foundations of Comparative Politics. CQ Press, 2019. Chapter 11.

Suggested materials for further exploration

 Video – Which voting systems is the best? https://www.youtube.com/watch?v=PaxVCsnox_4

 Interactive website – ACE’s comparative data on elections systems: https://aceproject.org/epic-en/
 Academic article – Gonzalez-Eiras, Martin, and Carlos Sanz. "Women’s representation in politics: The
effect of electoral systems." Journal of Public Economics 198 (2021): 104399.
https://www.sciencedirect.com/science/article/abs/pii/S0047272721000359

5
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Parties and party systems

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

Conceptualizing political parties

Parties are key players in contemporary politics – in fact, it is hard to think of democratic politics in modern societies
without political parties. Yet we often look at parties with suspicion, as sources of division, polarization, or even
corruption. A look at the modern history of political thought reveals that this tension has long been at the core of
discussions about political parties. Here are some classical definitions:
- Edmund Burke (1770) defined a political party as “a body of men united, for promoting by their joint
endeavors the national interest, upon some particular principle in which they all agreed.” Burke is writing
in the early phases of British parliamentary democracy, at a time where parties are associated with
factionalism and division, and is trying to highlight the beneficial effects parties can have over a country’s
governance. In his view, there is no division within parties and parties promote the national interest.
- Robert Michels (1911) wrote “a modern party is a fighting organization in the political sense of the term,
and must as such conform to the law of tactics.” Michels is interested in the internal organization of parties,
and draws attention to how they all necessarily become oligarchic and undemocratic (he even talks about
an “iron law of oligarchy” in parties), despite their claims of being democratic. In his definition, Michels
draws attention to how parties are organizations that, if successful, must be effective in the electoral
battleground.
- Joseph Schumpeter (1950) writes that “a party is a group whose members propose to act in concert in the
competitive struggle for political power.” Consistent with his definition of democracy as being mostly about
elite competition, Schumpeter sees parties as dedicated to that competition, and mostly office seeking. But
his definition also introduces an element of within-party division or competition across members.
- Anthony Downs (1957) goes further by saying that “a political party is a coalition of men seeking to control
the governing apparatus by legal means.” Downs brings our attention to the potential for divisions inside
the party (note the word “coalition”) and how they compete to control the state but do so through certain
legal channels (unlike say a rebel group).
- John Aldrich (1995) advances that idea further by saying that a party is not just any coalition but an
“institutionalized coalition, one that has adopted rules, norms, and procedures.” The role of
institutionalization is key for successful parties – this involves an organization, a brand, and a relatively
stable policy platform.

As this series of conceptualizations make clear, political parties are very unique organizations. For one, they are the
key players in the electoral arena – even in candidate-centric electoral systems, parties play a key role in the
financing, organizing, and monitoring of elections. Second, parties are private organizations with certain interests,
but if successful they “occupy” the state and get to use the state’s coercive, extractive, and coordinating capacities
to push certain policy agendas. Third, parties are in constant competition both amongst and within them.

The role of political parties in democracy

Despite the long-standing suspicious or cynical views about political parties, they are fundamental for the well-
functioning of a democracy. They play unique roles in the following tasks that political systems must solve:
- Linkage with society. Parties represent citizens and societal demands. Through their dual nature of being in
civil society and in the state (legislative and/or executive power), they connect other civil society
organizations (such as churches, unions, associations, businesses, etc.) to the state.

1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

- Aggregating, articulating, and bundling societal demands. Through their articulation of policy platforms,
party brands, and party identities, political parties make political conflicts clearer, more legible, and more
actionable. By aggregating and bundling societal demands, they enable political demands and
accountability.
- Coordinating political elites and policy action. Through private, long-standing relationships of trust and
connections, political parties make it easier for political elites to coordinate, thus enabling policy action.
Coordination by parties is most visible within legislatures, but is also key in executives at all levels, and in
the coordination between branches and levels of government.
- Giving continuity to policy and political action. Through their permanence across time, political parties give
continuity to policy and political debates, policymaking, and political narratives. This enables political and
economic development.
- Enabling political accountability. By clarifying policy platforms, reproducing political imageries and brands,
and grouping political action, parties enable political accountability. They make it easier for voters to follow
political action and to make decisions about who to support.
- Selection, recruitment, and organization of political elites. Parties play a key role in the selection and
grooming of candidates, and the socialization of political elites and their training for contesting elections
and for holding public office.

Another way to see the value of political parties is to examine the politics of places with no or weak political parties.
For example, Palau and other states in the Micronesia do not have political parties. Yet their politics are not less
divisive or more efficient – in fact researchers have found that the role of parties is played by clans, with the ensuing
challenges for political accountability. 1 The US state of Nebraska does not recognize parties in its legislature.
Researchers found that patterns of legislator behavior lack recognizable patterns, which hinders political action and
accountability.2 In Peru, political parties have for a long time now been very weak and uninstituionalized. The
Peruvian party system is characterized by high levels of volatility and fluidity, very weak party leaderships and
organizations (partly due to an open list PR system), and the emergence and disappearance of a wide range of
parties.3 This party system and the challenges and dysfunctionalities it creates helps explain the emergence of
Alberto Fujimori, his self-coup in 1992, and the authoritarian regime that ensued, as well as the more recent self-
coup attempt by Pedro Castillo in 2022, the ongoing political crisis and the quick deterioration of Peruvian
democracy. These three cases alert us about the undesirable effects of having weak parties, and the intimate
connection between the removal of parties and authoritarianism.

The origins of parties

Modern parties originate in the representative assemblies that grew with the formation of modern states, from the
16th to the 19th century. Parties emerged first within parliament, most notably in the British parliament throughout
the 17th century with the growing organization and recognition of parties, and their progressive engagement with
campaigns. These are known as elite or cadre parties. The expansion of the franchise to include progressively larger
fractions of the masses gave origin to parties of extra-parliamentary origin. These parties were fundamentally
different because they emerged at fundamentally different points in time (relative to the extension of the franchise).
Parties of extra-parliamentary origin had different social bases, different organizations, and different internal
dynamics in reflection to where they came from.

We can therefore classify parties according to their type, with types appearing at different points in time in history
and under different circumstances:

1 Veenendaal, Wouter P. "How democracy functions without parties: The Republic of Palau." Party Politics 22.1
(2016): 27-36. https://doi.org/10.1177/1354068813509524
2 Wright, Gerald C., and Brian F. Schaffner. "The influence of party: Evidence from the state legislatures." American

Political Science Review 96.2 (2002): 367-379. https://doi.org/10.1017/S0003055402000229

3 Levitsky, Steven. "Peru: The institutionalization of politics without parties." Party systems in Latin America.

Institutionalization, decay, and collapse 2018 (2018): 326-356.

2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

- Elite or cadre parties. They emerge within parliament, as parliamentary democracy strengthens. They have
minimal organizational structures. Their only members are the elites (e.g., parliamentarians). The primary
resource base is wealth, connections, and patronage.
- Mass parties. They emerge outside parliament, as the franchise extends and political entrepreneurs seek
to mobilize the masses for accessing political office. They have a stronger organization, with local branches
and participation of elites. Membership is large, with some form of internal participation and democracy,
and a leadership that is to some extent accountable to the masses. The members pay fees, and the key
resources are membership as well as connections to other organizations such as unions.
- Catch-all parties. This is the form that major parties take in the second half of the 20 th century, with the
evolution of party competition and the consolidation of mass media. Parties have member structures but
the members are not central players; the party organization is subjected to the party in office; and the role
of members and bases is mostly to provide legitimacy and support for the elites.
- Cartel parties. This is a form of party that some authors identify in the later decades of the 20th century up
until the present, where bases are further marginalized, parties depend more on state subsidies, and
membership structures and their role in internal governance further blurred.
- Anti-cartel parties. These are parties that emerge over the past two decades or so that challenge catch-all
and cartel-parties, or even challenge the party system as a whole. These can be anti-establishment, left- or
right-wing populist parties. They tend to have stronger membership bases which they seek to mobilize more
than traditional parties.

Party systems and their origins

A party system is the typical pattern of political competition and cooperation between parties in a given territory
(typically, a country – but there can be subnational party systems as well which are in some cases differentiated
from those at the national level). The most important dimensions of a party system are the number of parties, their
relative size, their types, and the proximity and behavior amongst them.

Party systems emerge from three main drivers: the electoral system, the timing of franchise extensions, and
cleavages. Within these three institutional and macro-sociological conditions, there is some role for political
entrepreneurs, particularly at times of party system crisis such as the one many democracies are currently
experiencing.

A cleavage is a significant, socio-political division in society that is mobilized politically, with political conflicts
organized around it and party organizations reflecting such divisions. There are three key elements in a cleavage.
First, a social division that separates citizens by economic, cultural, or other social features. Second, awareness by
key actors about their collective identity, and willingness to act based on it. Third, an organizational reflection of that
division. Critical to understand the stability of party systems is that cleavages are deep societal divisions that have a
reflection of politics – by creating political parties which defend those cleavages (because they benefit from them
electorally), cleavages and party systems have significant inertia.

In a classical book, Lipset and Rokkan (1967) theorized the role of cleaevages in the creation and reproduction of
parties and party systems. They see cleavages emerging from very deep societal transformations or revolutions,
mostly in the 19th century at the time when modern states are consolidating and democracies emerging.
- The national revolution generally starting in the early 19th century, with the consolidation of nation states
and the ensuing conflicts, leads to two cleavages. First, the center-periphery cleavage, with conflicts
between liberals and state-makers, on the one hand, and more local and regional elites which resist efforts
of centralization and standardization (e.g., with national schooling systems, national languages, etc.). Party
families resulting from this cleavage are regionalist, ethnic, linguistic, and minority parties, such as the Bloc
Québecois or the Partido Nacionalista Vasco. Second, the state-church chelavage, with conflicts between
liberal and secularizing state elites and religious authorities, around conflicts about the role of religion and
churches in public life (e.g., schools, family policy, etc.). Party families resulting from this cleavage are
conservative and religious parties, such as the Christian Democratic Union.

3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

- The industrial revolution generally starting in the late 19th century, with conflicts ensuing of the economic
transformations of industrialization, leads to another two cleavages. First, the rural-urban cleavage, with
conflicts between industrial and agricultural sectors of the economy and society, with divergent policy
preferences on issues of trade and economic policy more broadly. Agrarian parties come from this cleavage,
some of which still surviving to date like the Finnish Center Party. Second, the workers-employers cleavage,
or proletariat-bourgeoisie cleavage, with conflicts between the working class and the industrialists, with
conflicts around working conditions, social policy, welfare, etc. Workers’ parties, labor parties, and social
democratic parties come from this cleavage, such as the Labour Party.
- The Russian revolution, in the early 20th century, with conflicts in the left between social democrats and
communists, leads to a cleavage between socialists and communists. This cleavage led to the emergence of
communist parties which often played an important role in the second part of the 20 th century and are in
any case still present today in many democracies, for example the Partito Comunista Italiano.

Lipset and Rokkan argued that because of the transformations of society and politics ensuing from these cleavages
and from the expansion of electorates, and because the resulting parties reinforce such cleavages, party systems in
Europe “froze” in the first few decades of the 20th century. They noted how party systems were pretty much the
same, and even the organizations names and brands of parties were the same in the 1960s as they had been decades
before.

Nowadays however party systems are experiencing significant levels of volatility and change, with some very well
established parties coming under a lot of stress and even becoming marginal. Party systems even in consolidated
democracies like France or Germany are undergoing significant change. A leading factor of change is the emergence
of a potential new cleavage, which could be called that between open and closed societies, or globalist vs
nationalists, with conflicts ensuing around issues of trade, migration, and globalization. Many populist parties on the
left and on the right can be seen as resulting from, and reinforcing, this cleavage. However it is not obvious this will
consolidate as a separate cleavage, because traditional parties typically seek to “collapse” partisan competition
along the traditional cleavages. In many cases we see realignments of electorates but traditional parties still
capturing conflicts and competition along these new axes, for instance in the United States or Spain.

A key factor contributing to the stability of party systems is electoral rules, which can only be changed with the
participation of major parties. Duverger famously wrote about two “laws”. First, that first past the post electoral
systems (or more generally majoritarian systems) produce two-party systems. Second, that proportional
representation (PR) electoral systems produce multi-party systems. The more proportional an electoral system (e.g.,
with larger districts, lower thresholds, quota rather than divisor systems) the more fragmented the party system.
Duverger theorized there are two key mechanisms for the electoral system to shape the party system:
- Mechanical effects. These have to do with how votes are translated into seats. Majoritarian systems
mechanically constrain the representation of parties, and tend towards two-party systems.
- Psychological or strategic effects. Both voters and candidates adjust their decisions strategically based on
their expectations of who will get representation. Voters are more likely to vote to a major party in
majoritarian systems, and candidates are more likely to run with a major party (rather than a minor one, or
a new one) in majoritarian systems.

Morphology of party systems

A categorical classification of party systems based on the number and size of parties is useful to differentiate
qualitatively among party systems.
- Single-party systems have only one legal party – in these authoritarian settings (e.g., the Soviet Union or
China) the party plays a major role not just in facilitating governance but also in authoritarian control,
repression, and cooptation.
- Dominant-party system – these are systems in which a party recurrently gets an overwhelming share of the
electoral support, leading to no turnover in executives. Examples include both electoral autocracies (e.g.,

4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Mexico under PRI rule) and democracies (e.g., Japan in the second half of the 20th century or South Africa
after the transition).
- Two-party systems – these are party systems with two major parties that have a large majority (say, 80% or
more) of votes and seats, with alternation in power between them, and generally one-party governments.
Examples include the United States, Spain until about 2015, or Canada..
- Multi-party systems have a larger number of parties, with no party getting close to getting a majority of
seats or votes. Parties are of different sizes, not necessarily all equally sized. These party systems tend to
produce coalition governments, and political turnover is therefore more dependent on bargains across
parties. Examples include Italy, Belgium, and Colombia.

There are important trade-offs between two- and multi-party systems. Multi-party systems are better for
representing a wider range of electoral preferences and for giving political power to underrepresented groups.
However two-party systems tend to have more clarity of responsibility, a more agile formation of governments, and
a more direct connection between the results of the election and political turnover.

Quantitative measures are useful to get a more continuous measure of party system fragmentation.
- The fractionalization index, which goes from 0 to 1, where 0 means total concentration (1 party only) and
1 means total fragmentation (each seat corresponding to one party).
- The effective number of parties (ENP), which counts parties with representation but weighting them by the
share of votes or the share of seats that they have. Countries with an effective number of parties close to 1
are likely electoral autocracies with limited competition. Countries with a very high effective number of
parties (close to 10) are those with very high levels of fragmentation, typically resulting from a very
proportional electoral system, open lists, and less institutionalized parties (e.g., Brazil, Colombia, Peru) and
sometimes also from high levels of ethnic, religious, and/or regional fractionalization (e.g., Belgium, Israel).
Historically, high ENP figures were common in Europe in the interwar period, which contributed to
government dysfunctionality, conflict, and in the extreme the rise of totalitarian governments and civil and
international wars.

Dynamics of party systems

To make sense of parties’ placement within party systems, political scientists often make use of market analogies,
where parties act as firms providing policy platforms (and when in office policies) as products to voters who act as
consumers with policy preferences. The classic specification of this analogy is that of Anthony Down’s Economic
theory of democracy (1957). His model assumes a unimodal distribution of voters along a one-dimensional policy
space (left-right) and a two party system. From this perspective, parties observe voter preferences and move
(typically along a one-dimensional space) to maximize their electoral support. A key prediction of this model is that
parties will converge towards the middle of that distribution (getting closer and closer to each other in policy
positions), with voters in the middle being the determinant ones for the outcome of the election. This is known as
the median voter theorem.

This model can be extended to more complicated scenarios, such as one with a multi-party system, different
distributions of voters (for example, bimodal with two poles), and a more complicated policy space (for example one
with a left-right cleavage and a nationalist-globalist cleavage).

In these scientific efforts scholars measure the policy positions of parties, and the policy preferences of voters, and
how they relate to each other. Recent research for example has uncovered that political elites in Europe and the
United States often have biased perceptions of voter preferences, which may help explain some of their strategies
and observed levels of partisan polarization. 4

4Broockman, David E., and Christopher Skovron. "Bias in perceptions of public opinion among political
elites." American Political Science Review 112.3 (2018): 542-563. https://doi.org/10.1017/S0003055418000011
5
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Questions for reflection and discussion

 How do parties contribute to the health of a democracy?

 What features do institutionalized parties have, and how do features of a political regime, form of
government, or electoral system contribute to party institutionalization?
 What are the key trade-offs between two- and multi-party systems?
 What explains the emergence, resilience, and crisis of political parties?

Required readings

 Caramani. Comparative Politics. Oxford University Press, 2020. Chapters 12 and 13.

Suggested materials for further exploration

 Interactive website – Party Manifesto Project: https://manifesto-project.wzb.eu/

 Video – Book presentation on “Political entrepreneurs: The rise of challenger parties in Europe”
https://www.youtube.com/watch?v=3Fpx0sd-34Y
 TV show – Borgen (season 3 focuses more on the emergence of new parties)

6
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Voters and political behavior

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

Political behavior

While the course so far has mostly focused on institutions, comparative politics is also interested in individual-level
outcomes, especially political behavior. Political behavior refers to individual (or groups’) involvement in politics,
including any activities that have consequences in relation to politics, government, or policy. Behaviors of interest
in this subfield of comparative politics include vote choice (who people vote for), turnout (whether people vote),
participation in unions, participation in parties, petitioning, demonstrations, activism, protest, and even violence.
These different forms of political behavior vary on a number of meaningful dimensions, including whether it is
conventional (i.e., within formal institutions) or unconventional (outside them); more versus less involved, more
versus less public (or observable to others), more versus less risky (or dangerous to citizens, in either physical or
legal terms), and more versus less legal.

The field of political behavior was fundamentally transformed from the 1950s onward through what was called the
"behavioral revolution". This was a move in the discipline from examining formal institutions to examining how the
actors operating within them behaved. Progressively, political scientists since the middle of the 20th century turned
to collect and analyze mass surveys, systematic qualitative work (e.g., interviews, focus groups) and more recently
to do field and survey experiments that allow us to measure the causal impact of treatments of interest. In the study
of citizens' or elites' political behaviors, we are less constrained methodologically than when it comes to studying
macrostructures such as the form of government or the electoral system. By focusing on politics at the microlevel,
we can collect and analyze invaluable microdata (e.g. through surveys, interviews, or ethnography). Because the unit
of observation is often an individual citizen / politician / bureaucrat, it is easier to design and implement randomized
experiments. By randomly assigning a treatment of interest and comparing outcomes across the treatment and the
control group, we can credibly draw causal inferences (i.e., learn about the causal effect of a given intervention such
as a campaign ad or a misinformation campaign)

In substantive terms, the field of political behavior is often preoccupied with understanding how people choose who
to vote for, why and who chooses whether to vote, how people form their political opinions and beliefs, and who
joins social movements, parties and other costly forms of political engagement. There is also a lot of research on the
interplay between behavior and institutions, for instance by examining how political behavior responds to, but can
also transform, macro-level political institutions. Some recurrent themes in the academic research about political
behavior are biases in political behavior (e.g., some people vote more than others), the role and limits of rationality
(e.g., how the economy shapes or doesn't shape perceptions of the economy), the role of identity and emotion (e.g.,
how identity shapes vote choice), and the role of elites (e.g., how political entrepreneurs can influence individuals'
political behavior)

The paradox of participation

A key insight from the early literature in political behavior is that, from an economic standpoint, we would not expect
people to participate in politics – yet they do. Economic theory would start by noting that political behavior is costly
– it is time consuming, cognitively costly, risky (especially in authoritarian settings) and can even be burdensome
emotionally at least in polarized contexts. Another reason economic theory would predict no political participation
is that it often evolves around the production of public goods. Through participation, citizens often seek to produce
public goods such as a better government, a change in public policy (e.g., universal healthcare) or a better

1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

environment. One key feature of public goods is that they are nonexcludable: individuals who did not contribute to
the production of the good cannot be excluded from enjoying it. The other key feature of public goods is that they
are non-rivalrous: consumption of the good by one individual does not reduce the amount available for consumption
by others. With this in mind, we can see that collective action problems are central to political behavior. Self-
interested individuals would find it rational to not engage in costly behavior to contribute to the production of a
public good, in the hope that others will do.

Mancur Olson highlighted in his book on The logic of collective action that, counter to what was assumed at the time
at least by pluralists, the existence of a common interest among a set of people (one that could be advanced through
collective action) does not lead to voluntary collective action to advance the interest. Olson highlighted how rational
individuals would therefore fail to contribute to the production of public goods, and groups will therefore not
produce optimal results. This freeriding behavior is more serious the larger the group, because in larger groups your
individual decision to contribute to the public good is less likely to be the decisive factor. Olson here went counter
to the idea that the big threat in a democracy is the tyranny of the majority. From his perspective, the majority is
likely unorganized.

The paradox comes when we contrast that prediction with the empirical reality that many people do participate in
politics. Collective action in general, and political behavior more specifically, do happen. This is the collective action
paradox. What can explain this paradox? Olson thought collective action problems are solved, in large enough
groups, mostly through either coercion (for instance, compulsory contributions through taxes, ultimately based on
coercion) or selective incentives (for instance, the Ghent system or other institutions that distribute material benefits
to those who contribute to a public good). Yet we also observe political participation in the absence of coercion and
selective incentives – for instance, many people vote even when voting is not compulsory.

Other solutions to the collective action problem go through political leaders (who may change people's perceptions
of costs and benefits, or mobilize beyond the rational calculus), ideas (such as nationhood, solidarity, political
ideologies), social networks (since people's political behavior is often embedded in networks of friends, family,
coworkers) and social psychology (such as feelings of belonging and other intrinsic benefits of participation).

Vote choice

Understanding vote choice is central to comparative politics, for both theoretical and practical reasons, including
predicting the outcome of an election, designing electoral campaigns, understanding the rise of extremist parties,
among others. What drives people to vote for one party or another, and changes in those reasons, may shape
dynamics of democratic governance and accountability

There are three main perspectives on elections and vote choice: elections as popular sovereignty (whereby observers
see “the people” determine government policy by electing representatives), elections as leadership selection (where
elections as understood as political elites competing for the vote, and voters deciding based on their assessment of
incumbent performance), and elections as an expression of identity (a view through which people’s belonging and
identity drives their vote and their opinions).

The view of elections as popular sovereignty is appealing because it is normatively aligned with a democratic ideal
("government by the people and for the people"). This view understands citizens as informed, interested in politics,
and engaged with the political system. In turn, this view understands elected politicians as delegates representing
the will of the people. In this perspective, vote choice works in the following sequence: 1) citizens have a self
placement in an ideological or policy space on one or multiple substantive issues; 2) citizens assess the placements
of the different candidates / parties; 3) citizens choose the party / candidate closest to them. From this perspective,
what we would need to do to improve elections is giving more power to voters, strengthen deliberation, establish
avenues for direct democracy, and do civic education programs. As appealing and mainstream as this view is, there
are important limitations to it in terms of its consistency with empirical research about political behavior. First,
survey data consistently show people have low levels of interest in politics and information. Second, ordinary people

2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

often have no coherent policy preferences consistent with issue voting. Third, expressed preferences are very
sensitive to question wording. Last, when given opportunities of more direct democracy (e.g., referenda,
participatory budgeting), participation is often low and biased.

The view of elections as leadership selection is appealing because it has less demanding assumptions about citizens,
focus on elites. This perspective understands citizens as not necessarily informed about policy issues, but they know
how well incumbents did, or what kind of character candidates have. Consistent with this, politicians are understood
as trustees – they enjoy the confidence of citizens only temporarily. Vote choice, in turn, is understood as a
retrospective or prospective evaluation of political elites. In theories of retrospective voting, citizens assess the
performance of the incumbent (typically in the economy) and decide whether to renew their trust. In turn, in
theories of prospective voting citizens examine the character and trustworthiness of candidates to decide who to
vote. From this perspective, programs seeking to improve elections should provide more and more timely
information on relevant issues; raise education. Even though this perspective is less demanding on voters, it is
normatively less desirable as it is more elitists, and it faces some challenges from empirical research. First, surveys
show people have trouble attributing responsibility among levels of government. Second, voters are very myopic in
their assessments of incumbents – they are easily influenced by what happens immediately before the election.
Incumbents know this so they typically seek to produce economic booms, or to announce reforms or works right
before the election. Last but not least, there is ample empirical evidence of voters punishing incumbents for facts
governments do not control, such as shark attacks or defeats of local sports teams.

A final, more recent perspective is to examine vote choice as an expression of identity and groups. This view is
appealing because it does not assume or require that voters be informed, and can better explain patterns that we
observe particularly in more polarized societies. From this perspective, citizens are members of groups; those
overlapping memberships shape their identity and their view of the world; political attitudes are about groups /
factions. Politicians, in turn, are seen as either group members (which connects to the idea of descriptive
representation) or mobilizers of group identity. Vote choice in this perspective happens in a very different sequence
than the other two: 1) citizens have social identities, which shape their partisan allegiances; 2) allegiances, in turn,
shape how citizens assess the reality and choose their vote; 3) choices are rationalized ex post (motivated reasoning).
From this perspective, then, identities drive perceptions and choices, which in turn drive policy preferences. From
this perspective, improving elections requires facilitating the work of political parties and interest groups, limiting
the influence of money in politics, and helping the unorganized get organized. Normatively, this perspective is quite
troubling for a variety of reasons (including the multilayered dangers of polarization), but it also faces some problems
empirically. First, many voters have weak(er) identities, or less clear ideas of which group / party is consistent with
their identities. Second, swing voters are often the most relevant to determine electoral outcomes.

Turnout

Turnout, that is, electoral participation, is a backbone of representative democracy. Turnout helps legitimize
legislative and executive powers, and the system of government. Moreover, turnout directly connects citizen
preferences and political selection.

Yet there is wide variation across and within countries in rates of turnout, with systematic correlates (both at the
macro and the microlevels) of the decision to vote. This matters because differential propensities in turnout in a
society bias the selectorate, who gets elected into office, how political officials perceive the preferences of the
electorate, and the policies that get passed.

Globally, turnout has been decreasing over the past few decades, likely to a variety of reasons. First, post-materialism
has fundamentally transformed political behavior, with lower levels of trust and interest in formal political processes.
Second, political behavior has diversified, taking a wider range of forms, including online. Third, political allegiances
and identities have been transforming, leading some to disengage from traditional parties and their competition.

3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

At the individual level, turnout is, generally speaking, positively correlated with the following variables: gender, age,
socioeconomic background, income, ethnic majority status, turnout in the previous election, party identification,
union membership, attendance of religious services, and social networks more broadly.

At the macro-level, turnout is, generally speaking, positively correlated with the following variables: ease of voting
(automatic registration, early voting, etc.), proportional representation electoral systems, compulsory voting,
whether the election is first order, the timing of the election (whether it happens on a weekend, or whether it is
concurrent with other elecitons), competitiveness of the election, the electorate being small, and the campaign
being expensive.

Required readings

 Caramani. Comparative Politics. Oxford University Press, 2020. Chapter 18.

Suggested materials for further exploration

 Podcast – “What’s wrong with democracy? A conversation with Larry Bartels.”

https://podcasts.apple.com/gb/podcast/whats-wrong-with-democracy-a-
conversation/id1402038733?i=1000443444196

4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Subnational authorities

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

Territorial divisions of state authority

We typically think of the state as a unitary actor – and a dangerous one. As illustrated by Hobbes’ theory of the state,
we tend to think of a monocratic state concentrating all power. Yet, many of our political institutions in effect divide
state authority. The ideas of separation of powers and checks and balances stem from this notion. This helps us
avoid tyranny and its consequences – but also creates its own problems. For example, checks and balances,
accountability institutions, and constraints on the executive may make it harder to innovate, to adapt, and to
respond quickly.

Common ways in which state authority is divided include branches of government (executive – legislative – judicial),
politicians versus bureaucrats, national versus subnational (state, regional, local, etc.), national versus supranational
(e.g. the European Union, the Organization of American States, the United Nations, NATO), lower versus upper
legislative houses (in bicameral states), politicians versus judges, budget spenders versus auditors, internal versus
external auditors, etc.

Veto player theory

Veto player theory provides a useful framework from which to make sense of the origins and consequences of these
divisions of state authority. A veto player is an actor whose agreement is necessary to implement changes in the
political status quo. Institutional veto players are those generated by the constitution – e.g., the upper house,
federated governments, constitutional court. Noninstitutional veto players include different members of a
government's parliamentary coalition, among others.

Countries' political systems can therefore be described by the number of veto players there are. Veto player theory
makes a number of predictions through this perspective on political systems. The first prediction or hypothesis
stemming from veto player theory is that the more veto players in the country, and the more ideologically distant
they are, the less likely and smaller policy changes will be (and vice versa). The second one is that the more veto
players there are, the more regime instability there will be – this argument applies particularly to presidentialism,
where there may be clashes between popularly elected heads of government and veto players that are harder to
solve from within institutional channels. A third prediction from veto player theory is that the more veto players
there are, the more likely it is that the country will have judicial and bureaucratic activism – that is, judges and
bureaucrats actively engaging in policy making.

Veto player theory is an influential approach to examine politics, but it is not without problems. First, it is thin on
politics – it does not clarify or examine who the actors occupying veto player roles are, what their sources of power
are, or how they are connected. Second, veto player theory ignores how parties and interest groups can help
coordinate actors. Third, veto player theory ignores that in some cases, by diluting responsibility and facilitating
blame shifting, the existence of "veto players" can actually facilitate unpopular reform in some cases.

Another key question that is left unanswered by veto player theory is where veto players come from. Common
answers to this question refer to a normative commitment to limited government, the influence of an assumed
welfare-maximizer writer of the constitution, or to the pressures of isomorphism or foreign powers. All these
perspectives ignore that veto players are ultimately produced or established by political elites, at least in transitions

1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

to democracy or constitution-making moments. This suggests that veto players emerge from conflicts and bargains
among political elites. This idea invites us to take more seriously the political origins and dynamics of veto players,
and how they shape states and state capacity. There are multiple reasons why politicians would constrain executive
authority by establishing veto players, out of self-interest: as a way to preserve the political community (and allow
for long-term planning); as a way to constrain their opponents when they are out of office; as a way to credibly
commit, yet with flexibility, to fundamental rights (or solid monetary policy); or as a way to address distributive
conflicts.

When thinking about the consequences of veto players, it is important to realize that in some ways rather than
simply limiting the scope of politics, they create new political actors and arenas. For example, federalism creates and
empowers subnational governments; bicameralism creates and empowers senators; constitutionalism creates and
empowers constitutional courts. These arenas are not a mere continuation or extension of "politics as usual"; they
have their own resources, norms, and strategies. Their shape and effects are not set definitively in the constitution
/ law. The informal dimension of these institutions matters as much as their formal or legal powers.

Multilevel governance

Multilevel governance refers to the geographical dispersion of authority across levels of government be it within or
beyond states. There are two fundamentally different ways to examine multilevel governance. First, a binary
classification of federal vs unitary: the key criterion is constitutional, and in particular whether the national
government can take back subnational authority. Second, a continuous measurement of regional authority: law
making, executive control, fiscal control, and borrowing control.

The regional authority index measures these different dimensions of regional authority and shows that there is wide
variation in regional authority across both federal and unitary states, and that regional authority has been on the
rise in time. Other patterns are the growing divergence in political, administrative, and fiscal powers across
governments in the same level, and the merging or scaling up of subnational units.

There are three main approaches to thinking about drivers of decentralization. First generation theories, or the
functionalist approach, starts by thinking of economies of scale for public service provision. The key then is to provide
services at the optimal level, keeping into account that citizens in different regions may prefer different goods or
service, and that in some cases subnational governments may be better informed about (and closer to) what people
want. From this perspective, there is a correct level at which to provide goods and services. The challenge then is to
find the optimal design where costs and benefits balance. A key limitation of this approach is that it leaves little
space for political conflicts. It assumes decisions are taken by a welfare maximizing social planner. Decentralization
is seen as a technical matter.

Second generation theories, or the economic approach, take a fundamentally different perspective by assuming that
multilevel governance is about distributive and redistributive conflicts among regions. In this approach,
decentralization is driven by self-interested political and bureaucratic elites, as well as citizens in different regions.
Multilevel governance is therefore something they can use to their advantage. From this view, deviations from what
would be socially optimal emerge from the interests of subnational citizens and elites. A key insight from this second
generation literature is that imbalances in fiscal and policymaking capacity create problems. For example,
subnational borrowing may create moral hazard, and a risk to macroeconomic stability. Another key insight is that
multilevel governance may fuel inequality in access to public services. And last but not least, this approach
emphasizes that political games of credit claiming and blame shifting are central to politics. A key limitation of this
approach is that it ignores the role of differentiated identities.

A third, identity-based approach to multilevel governance fills this gap by assuming that subnational / territorial
identities are a key driver of decentralization demands. Here, it is not just about competencies but also about
boundaries and, in the extreme, secession. The boundaries of jurisdictions, as well as their competencies, may be
subject of heated political debates, and demands for self-rule may be independent of optimal task allocation or pure

2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

economic self-interest. Ethnic or linguistic differentiation and geographic isolation are typical correlates of identity-
based decentralization. But the role of identity and ethnicity is mediated by subnational party systems. Party
competition and party system differentiation are central regional parties' success fuels decentralization. Issues of
identity and recognition are key in that feedback loop between party competition and decentralization.

Federalism

Federalism is in essence an institutionalized division of state authority between national and subnational units. In
unitary systems, the central government remains the sole sovereign, subnational governments are subsidiary, and
their powers (to make or implement policy) can be curtailed by the central government. In federal systems, central
and subnational governments have independent powers and responsibilities; the central government cannot abolish
or curtail subnational authority

The key difference between federal and unitary states is not the existence or the policy responsibilities of
subnational governments. The key is whether the central government can abolish / suspend / reshape the power of
subnational governments. In federal systems, subnational governments are constitutionally protected, their power
is inherent rather than delegated, and can therefore not be curtailed by the central government. Some countries
such as Spain and the UK are quasi-federal because they have many of the features of federal countries but the
central government retains the ability to suspend the autonomy of regions.

Federalism can be classified along several dimensions.

- Congruent vs incongruent. Whether subnational units differ on their sociodemographic composition.
Congruent federations include Brazil, the USA. Incongruent federations include Belgium, Switzerland
- Symmetric vs asymmetric federalism. This refers to whether subnational units differ on their powers.
Symmetric federations include the United States and Brazil, and asymmetric federations include Belgium,
Canada, Russia, and Switzerland.
- More versus less decentralized. This refers to the extent to which subnational authorities have legal
authority to pass policies, to collect and spend taxes, and state capacity to implement policies.

There are two main narratives about the origins of federalism. In coming-together federalism, federalist
arrangements emerge as the result of bottom-up bargains in which previously sovereign unities form a federation.
The cases of Australia and the United States are consistent with this. By contrast, in holding-together federalism,
federalist arrangements emerge as the result of top-down bargaining process in which the central government
decentralizes power to subnational governments. Holding-together federalism is typical of multiethnic states in
which federalism is seen as a strategy to appease secessionist movements or parties, as in Belgium.

Bicameralism

Bicameralism is in essence an institutionalized division of legislative authority between two chambers. In unicameral
systems, the legislature has just one chamber or body. In bicameral systems, there are two distinct assemblies in
charge of legislative activity. Names of the second (or upper) house vary significantly across systems.

Bicameralism can be congruent or incongruent. It is congruent when the political composition of both chambers is
similar, because the system to choose members of the upper house is largely aligned with that of the lower house.
It is incongruent when they differ, for instance if the upper house is not elected by voters but delegated by
subnational authorities, or if the electoral system is fundamentally different across the two. There are four main
methods for selecting upper house members: heredity, appointment, indirect election, or direct election

Bicameralism can be symmetric or asymmetric dependent on whether the upper house is an equal partner in the
lawmaking process. If it is, bicameralism is symmetric. In many cases however the upper house plays a secondary
role. In the extreme, the lower house has ultimate decision-making power, being authorized to pass legislation even
if the upper house votes against it.

3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

A country has strong bicameralism only when it is symmetric and incongruent. The combination of symmetry and
incongruence is what makes it likely that there will be conflict or disagreement between the two houses, and that
such disagreement will matter for policy making.

Historically, bicameralism emerged as a status-quo-preserving institution whereby aristocrats would counterbalance

the deliberations and decisions of the lower house. Around the time when the legitimacy of class-based
bicameralism declined, emerged a model (first in the US and Switzerland) of federalism-protecting and enhancing
bicameralism, where the upper house would represent subnational units. Many unitary systems retained
(democratized) upper houses still on the idea of tempering (and improving) the deliberations of the lower house.
Others eventually got rid of the upper house, or reformed the system to make it strongly asymmetric.

Questions for reflection and discussion

 In what ways does multilevel governance deepen democracy? In what ways may it hinder democracy?
 Are citizens better off in federal than unitary systems? How does that depend on subnational political
institutions?
 How do differentiated political party systems alter the nature of federalism and decentralization?

Required readings

 Clark, Golder, and Golder. Foundations of Comparative Politics. CQ Press, 2019. Chapter 13.
 Caramani. Comparative Politics. Oxford University Press, 2020. Chapter 11.

Suggested materials for further exploration

 Dataset – Regional authority index: https://garymarks.web.unc.edu/data/regionalauthority2/

 Opinion piece – “Why federalism is hard”, Vox:
http://www.vox.com/polyarchy/2018/2/27/17058498/federalismhard

4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Courts

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

Constitutions as political products

A constitution is the formal, legal source of state authority. Constitutions typically establish the structure, procedure,
powers, and duties of governmental institutions. More recent constitutions also include a list of guaranteed rights.
It is useful to see constitutions as meta-norms, that is norms about how norms are produced, interpreted, and
enforced.

Constitutions are political products – while they structure politics, they are also created and structured by politics.
Counter to the legal perspective on constitutions, from a political science perspective it is important to realize that
they are not self-made, self-enforcing, nor definitive. First, they are not self-made or given externally, at least not in
most democracies (note however in some cases constitutions are imposed by invading powers). In general,
constitutions are instead designed by political elites. Second, constitutions are not self-enforcing. From a social
science perspective, it is key to realize that provisions in a constitution are not respected simply because they are in
the constitution – the mandates and rights contained within must be implemented by political, bureaucratic, and
judicial elites. Finally, constitutions are not definitive. Even if they are not amended, constitutions are constantly
subject to political battles around their interpretation.

There are several useful ways to classify constitutions.

- Codified vs uncodified. A codified constitution is contained within a single, written document. Uncodified
constitutions may be contained in several documents or even have fractions that are not formall written
down. The United Kingdom, New Zealand, and Israel have uncodified constitutions.
- Entrenched vs unentrenched. Entrenched constitutions can be modified through a special procedure of
constitutional amendment. Unentrenched constitutions can be reformed by simple legislative majority.
- Absolutist vs legislative supremacy vs higher law constitutions. Absolutist constitutions often reject rights
and separation of powers, but may include a blueprint of how authoritarian power is exercised. Legislative
supremacy constitutions have no constitutional review, and are not entrenched. Higher law constitutions
are entrenched, have a system of fundamental rights, and a system of constitutional justice.

New constitutionalism is an institutional combination of an entrenched constitution, a bill of fundamental rights

included in the constitution, and constitutional review of legislation. New constitutionalism also refers to the
growing norm that this combination is desirable and preferable to alternatives. With the rise of new
constitutionalism around the globe, the last century has seen a growth in the number of constitutions written, the
share of states that have higher law constitutions, and also the number of constitutional amendments.

Constitutions typically emerge in times of large-scale political crisis, including after international or civil wars, during
a transition to democracy, or during periods of societal transformation. Constitutions are typically written by
constitutional assemblies, although others are written in regular legislatures or directly by executives. While
constitutional assemblies often count with input from legal scholars, civil society organizations, interest groups, etc.,
the final product is eminently political, as it results from the bargains among the political elites involved. In some
cases, particularly after international defeat or under occupation, the process is largely influenced by external actors.

In any case, constitutions are a deliberate creation by political elites. They typically emerge from political bargains
after a period of disruption or fundamental change, and therefore in a context of mutual suspicion among political

1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

elites. That is one key reason constitutions are in many ways incomplete contracts, even when they detail many
different aspects of a country’s political life. As a result, constitutions are to some extent vague, ambiguous, or even
contradictory. Constitutional review can be seen as a way to address that incomplete nature of constitutional
contracts.

Judicial review

Provisions for constitutional or judicial review, that is the review the alignment of laws and other government
decisions with the constitution, can be seen as a way to deal with the incompleteness of the constitution. There are
three main political rationales for establishing judicial review. First, judicial review may be used as a strategy to
preserve hegemony (e.g., of a racial group or a socioeconomic group), particularly in times of change in the political
regime. Second, judicial review may be used as a political tool to solve credible commitment problems. Political
actors may establish courts to make commitments to private and/or international actors that otherwise would not
be credible (e.g., authoritarian governments who want to commit to protecting property rights to boost investment).
A final rationale for politicians to establish judicial review is insurance. Political actors may establish courts as a way
to protect their own basic interests (e.g., political and or ethnic survival) when they are in the opposition

Judicial review can bhe characterized along several dimensions. First, it can be abstract (i.e., not with reference to a
particular litigation case) or concrete (that is, with reference to a particular case). Second, it can be done a priori
(that is, before a law is formally enacted) or ex post (after it is). Third, it can be centralized (where a court has the
monopoly over constitutional review) or decentralized (where review is done by lower level courts).

There are two main models of judicial review, the so-called American and so-called European models. These are just
ideal types and in reality models of review vary also within continents. In the so-called American model, review
authority is decentralized, the supreme court has general jurisdiction, and review is concrete. In this context,
strategic litigation in lower courts remains a key way to influence what gets to the supreme court. By contrast, in
the so-called European model of judicial review, review authority is centralized, the constitutional court has only
(and exclusive) constitutional and fundamental rights jurisdiction, and review is abstract.

The impact of judges in state capacity and development

Courts play central roles in maintaining order and promoting development. They decide on all sorts of cases
(including criminal cases, civil disputes, disputes within and across families, administrative disputes between
governments and citizens, disputes between branches or agencies or levels of governments, and constitutional
controversies). Throughout these decisions, courts play a central role for state capacity and effectiveness, and
therefore for politics. The existence of effective courts plays a fundamental deterring role in promoting the state’s
coercive, extractive, and coordinating capacities. In so doing, courts contribute to rule of law, to private sector
development, and to the maintenance of democracy.

Consistent with the role of judges in promoting economic development, cross-national regressions have found a
positive association between de facto judicial independence (rather than de jure) and economic growth. 1 It is hard
however to rule out confounders and reverse causality with this design. A recent quasi-experimental study of judicial
reforms in Africa made progress on causal inference by doing within-country comparisons, before and after the
reform, in perceptions of members of ethnic minorities versus the majority. The author found that judicial reforms
have a positive impact on perceptions of governance, on satisfaction with democracy, and on self-reported economic
conditions.2

1 Voigt, Stefan, Jerg Gutmann, and Lars P. Feld. "Economic growth and judicial independence, a dozen years on:
Cross-country evidence using an updated set of indicators." European Journal of Political Economy 38 (2015): 197-
211. https://doi.org/10.1016/j.ejpoleco.2015.01.004
2 Chemin, Matthieu. "Can judiciaries constrain executive power? Evidence from judicial reforms." Journal of Public

Economics 199 (2021): 104428. https://doi.org/10.1016/j.jpubeco.2021.104428

2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Judicial selection

To effectively fulfill these important functions, judicial systems depend on the recruitment and deployment of
specialized agents, judges. Countries use a variety of systems to select judges and to select justices (i.e., judges of
higher courts), including political appointment by the executive, political appointment by the legislature,
appointment by lower courts, civil service exams, elections, and a combination of the above. Other key dimensions
of variation of systems for judicial selection, in particular for supreme and constitutional courts, is the length of their
tenure, whether they can be reappointed, and whether there is a compulsory retirement age.

Here again obtaining empirical evidence on the effects of selection systems is challenging, because simple
regressions with controls are unlikely to uncover unbiased estimates of the true relationship between a selection
system and outcomes of interest such as quality or number of decisions. A recent quasi-experimental study used
data from Pakistan, exploiting a reform that mandated that some judges be selected not by presidential appointment
but through selection by judges. The author found that removing the presidential appointment led to a significant
decrease in the share of cases where judges side with the federal government, which suggests that political
appointment causes direct accountability in judicial decisions. They also found that the quality of decisions went up,
without significantly increasing the time it took judges to make decisions.3

Judicial behavior

There are multiple approaches to examining judicial behavior, including:

- Legalism. From this perspective, judges "find" the meaning of legal rules through politically neutral
methods.
- Labor economics. In this view, judges are motivated and constrained by (mostly) nonmonetary costs like
effort or criticism, and benefits such as esteem, influence, self-expression.
- Sociology. Here, judges' biographies, personal characteristics, and identities shape decisions. For example,
research has shown that judges that have one daughter (compared to those that have one son) are more
likely to make decisions protective of women’s rights.
- Attitudinal. In this approach, more predominant in political science, judges' decisions reflect their
ideological attitudes on cases. Researchers in the United States especially have used data on judges’ own
decisions and data on judges’ campaign contributions as two competing approaches to measure the
ideology of federal judges or supreme court justices.
- Strategic. From this perspective, common in political science and economics, judges consider the
preferences of other actors, what they think others will do, and what the institutional environment is when
making decisions. s

While legal criteria are paramount in judicial behavior, their decisions are likely influenced by ideology, background,
and strategic considerations. Actors that can influence judicial behavior include their colleagues, their superiors,
politicians, the media, public opinion, and social movements.

A variety of reasons may explain why judges are influenced by external actors. Most directly, judges may be
accountable to their principals (or actors who appointed them), e.g. presidents or parties that appointed them, or
citizens who elected them. Second, judges may anticipate influence of external actors over their careers, and adapt
strategically. Politicians, higher-level justices, and the public can influence judicial careers. Less obviously, judges
may be influenced by their strategic consideration of how other actors will respond in the enforcement of their
decisions. Judges often need other powerful actors to collaborate for decisions to be enforced (especially in high-

3Mehmood, Sultan. "The impact of Presidential appointment of judges: Montesquieu or the

Federalists?." American Economic Journal: Applied Economics 14.4 (2022): 411-445.
https://www.aeaweb.org/articles?id=10.1257/app.20210176
3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

level cases including subnational governments, government agencies, etc.), and may adapt decisions to avoid weak
or inexistent enforcement.

Questions for reflection and discussion

 What are the main sources of tension between politicians and high-court justices? What kinds of conflicts
can emerge among them? How may politicians react to backlash to their decisions from judges? And vice
versa?
 How can poli
 How should judges be selected? Does your answer differ for regular judges and for highercourt judges?
How about the justices of the supreme court or the constitutional court?

Required readings

 Caramani. Comparative Politics. Oxford University Press, 2020. Chapter 9.

Suggested materials for further exploration

 Video – “Key facts about the Federal Constitutional Court.”

https://www.youtube.com/watch?v=1UJ7YVTCqA
 Podcast – “Should judges be elected or appointed?” Not Another Politics Podcast.
https://podcasts.apple.com/cy/podcast/shouldjudgesbeelectedorappointed/id1496412542?i=100064986
0307
 Dataset – The Supreme Court Database. Law School of Washington University.
http://scdb.wustl.edu/index.php

4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Interest groups and social movements

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

Civil society

Politics is often understood as the realm of governments, political parties and politicians – with an (indirect or
infrequent) appeal to citizens. Yet a lot of contemporary politics is played out in the sphere of civil society, by actors
that are not directly competing for votes but are directly involved in political conflicts. Interest groups, social
movements, and the media are some of the most important civil society actors – each with their own unique ways
of shaping politics

A key feature of political systems is thus how much space they leave for civil society to operate. International
observers have recently warned of shrinking civic space globally. This is reflected in V-Dem’s civil society index, but
more markedly in other bottom-up measures of civic space, such as the CIVICUS Monitor. The CIVICUS Monitor is a
participatory research platform that examines civic space around the world. They define civic space as “the respect
in law, policy and practice for freedoms of association, peaceful assembly and expression and the extent to which
the state protects these fundamental rights.” Through partnerships with civil society research organizations, they
produce yearly reports on incidents and violations of civic space, as a way to identify trends.

Debates about civil society and its role in politics, for instance with regards to protests, social movements, lobbying
activities, or associations, recurringly go back to some fundamental debates. How we evaluate specific instances of
civil society politics depends on how we stand in these basic questions, which include:
- What is civil society? How does it relate to the state, the market, political leaders, culture?
- How can civil society be most effective at controlling the state? And at controlling markets? And political
parties?
- How can collective action problems (which are endemic to civil society and citizen participation) be
overcome?
- How autonomous should civil society be from the state? Can genuine civil society exist when it is directly
supported and financed by the state?
- What is the ideal shape of a healthy civil society? A few, strong organizations? Many fragmented initiatives?
- Can there be civil society without organization or leaders?
- What should be the relationship between civil society leaders and street-level citizens?

From a philosophical standpoint, there are three main perspectives on civil society, which provide fundamentally
different answers on these debates above. First, a republican or unitarist perspective. From this point of view,
associations are seen as a threat to popular sovereignty; and large associations are seen with special suspicion as an
obstacle to the expression of popular sovereignty. Second, a liberal or pluralist perspective, from which associations
of all kind (including those based on religion or sports) are fa source of liberty and social development, with the
bottom-up organization and association of citizens seen as contributing to democracy. From the pluralist
perspective, the key policy implication would be to defend rights of association and to avoid any state interference
in a “free market” of civil society. Third, a corporatist perspective. From the corporatist view, associations play a
unique role in democratic governance, different from that of citizens and from the state. But counter to the pluralist
view, corporatism assumes that free competition among associations would lead to bias, since not all interest or
sectors of society have the same ease of organization, resources to do so, etc. The key difference is that corporatism
calls for intervention of the state, by recognizing and supporting interests while balancing forces across them. So if
from a pluralist perspective all the state has to do is to protect the right of association, from a corporatist one the

1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

state needs to intervene more actively in promoting and strengthening associations. Corporatist traditions are
particularly preoccupied with the balance between the interests of labor and capital, and seek to balance them while
incorporating them into the policy process.

Empirical research about civil society addresses questions at the level of the country, the individual, and the civil
society actor. At the country level, the two key questions are 1) what country-level factors shape the emergence,
strength, and impact of interest groups and social movements?, and 2) how do specific social movements and
interest groups condition the functioning of formal political structures? At the level of the individual, the key
questions in the empirical literature are who decides to join social movements and/or interest groups, and why?,
and 2) how do citizens evaluate civil society action and actors? Finally, at the level of the association, the key
questions have to do with country-specific successes and strategies, namely: 1) why are some social movements and
interest groups more successful than others?, 2) what are the key challenges they need to overcome?; and 3) what
strategies do they use?

Contentious politics

The phrase contentious politics refers to collective political struggle or non-routine collective action. Contrary to
other forms of political behavior, contentious politics is episodic (rather than continuous), occurs in public (rather
than online or in private), involves interactions between claim-makers and targets, and appeals to the state as either
a target or a mediator. Examples of contentious politics include protests, strikes, riots, insurgencies, and revolutions.

Different empirical measures of contentious politics have detected an uptick in the incidence of these forms of
political behavior, both in the global north and in the global south. A key empirical question is therefore what is
driving these changes, and how formal political institutions respond to the challenge of protests and strikes.

Collective action problems and solutions

Mancur Olson famously theorized the “logic of collective action”. At the core of the problem of collective action, in
his theory, is that self-interested individuals would find it rational to not engage in costly behavior to contribute to
the production of a public good, in the hope that others will do. Therefore, the existence of a common interest
among a set of people (one that could be advanced through collective action) does not lead to voluntary collective
action to advance the interest. Olson highlighted how rational individuals would therefore fail to contribute to the
production of public goods, and groups will therefore not produce optimal results.

Collective action problems permeate politics, and in particular contentious politics, because it is about the
production of public goods (i.e., goods such as national defense or a new public policy, which are non-rivalrous and
non-excludable).

Olson theorized that collective action problems could be addressed through two key strategies: 1) coercion, and 2)
selective incentives. By providing selective incentives, he argued, organizations would be able to overcome the free-
riding problem. A key example of selective incentives is the Ghent system, used in some countries of northern Europe
to alleviate the collective action problem of unions by having them manage unemployment insurance for their
members. Ultimately, Olson’s theory of collective action is deeply economic in its understanding of human behavior.
From an Olsonian perspective, the recurrence of collective action is in fact a paradox – people do participate even
in the absence of coercion and of selective incentives, although it is true that many do not.

Other solutions to the collective action problem (and theoretical explanations to the collective action paradox) can
tap on the role of leadership, culture, identity, and networks. Social science research has shown that leaders and
their appeals can fundamentally alter the way people perceive the costs and benefits of collective action. Identity
can also be a powerful driver of collective action, particularly when complemented with the role of social networks.
We are, at the end of the day, social beings and not mere “homo economicus” utility maximizers.

2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Social movements versus interest groups, and their strength

Social movements can be thought of as collective claim-making efforts with some degree of organization and
sustained in time. Social movements challenge the status quo in some policy area or idea, the power of government
authorities, and/or the power of other civil society groups. Ultimately, social movements often make an appeal to a
"better society". Social movements vary in their scale, range, organization, and ideology, but they all have a network
structure, some form of collective identity, the aim to change society or resist changes to society, and an explicit or
implicit resource to collective and public protest.

Interest groups differ in that they tend to focus on obtaining concrete benefits for their members (enticements),
often through specific policy changes. Unlike many social movements, interest groups often have a formal structure,
with a professional staff rather than activist membership. They often focus on influencing the outputs of the formal
political process.

When thinking about interest groups and social movements, it is relevant to consider the different channels and
abilities they have to influence the state. Here the theory of Offe and Wisenthal about the “two logics of collective
action” is useful because it articulates a simple yet powerful theory to differentiate between interest groups and
social movements, and among types of interest groups and movements. Their key insight was that for firms and
investors it would be easier to influence governments, because they can simply refer to their investment decisions
(e.g., a decision on whether to relocate a firm). This power makes it a lot easier for businesses to act collectively,
because it allows them to skip contentious politics and other forms of costly collective action. Contrast that to unions
and labor groups, which require the willingness and ability of workers to act together (ultimately, by striking, a much
costlier form of behavior) to influence public policy.

Beyond these structural differences on their dependence on collective action, social movements and interest groups
on the types of repertoires of collective action they have access to. Direct lobbying, that is the direct access to
policymakers, is most typical of interest groups, particularly those with endowments of resources like money and
expertise. Political exchanges, in turn, have more to do with trading of goods, policies, or legitimacy between
associations (such as trade unions or business associations) and government, for instance in the negotiation of
economic policy. Contentious politics are more costly forms of influence but potentially more disruptive (think for
example of a general strike). On the other end we have private interest government, where the state delegates
policy making authority to civil society organizations, particularly business and labor associations or other
professional associatinos.

The strength and effectiveness of social movements and interest groups is analyzed from four main perspectives.
First, resource mobilization theories, which put the emphasis on how endowments (such as money, activists,
expertise, or time) can be mobilized, particularly by movement entrepreneurs, to achieve goals. Second, political
opportunity structure theories, which focus on how social movements and interest groups can take advantage of
openings in the formal political system, which may enable or restrict their activity. From a political opportunity
structure, the key is not so much (or exclusively) the resources, but the obstacles and opportunities provided by the
formal political system. Key features of the formal political system include openness to challengers, cohesiveness of
political elites, party system crisis, and capacity and willingness to repress movements. A third perspective on the
success of interest groups and social movements is framing theories, which focus on how they communicate their
causes. For scholars of framing, effective frames provide three main things: a diagnosis (identifying a policy or
political problem, and culprits); a prognosis (a vision on what would happen without action, and a plan on how the
problem could be solved); and a motivation to join the action. Finally, identity-based theories focus on how identity,
individual and especially collective, drive individuals' decisions to join a movement, the organization of the
movement, and its behavior and effectiveness. Constructivists emphasize how identity is the main currency of
movements, and efforts to redefine it are therefore central.

3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Questions for reflection and discussion

 To what extent is there a fully autonomous sphere of civil society, or are social movements and interest
groups nowadays dependent on political parties and state institutions?
 What explains that collective action often takes place in “waves”, with times of relative quiet interrupted
by episodic and short but transformative periods of collective action such as the Arab spring or the occupy
movement?
 What are the channels through which social movements and interest groups can reshape society?

Required readings

 Caramani. Comparative Politics. Oxford University Press, 2020. Chapters 14 and 16.

Suggested materials for further exploration

 Visualization – Global Protest Tracker. https://carnegieendowment.org/features/global-protest-tracker

 Dataset – LobbyView portal. https://www.lobbyview.org/

4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

Bureaucrats

This handout is intended to supplement –not replace—the assigned readings, lectures, and class discussions.
For a complete understanding of the material, it is essential that students engage with all course resources.

Key concepts and theories

Conceptualizing bureaucrats

Bureaucrats are, in many senses, the face of the state. While few citizens interact directly with politicians or judges,
most of them have regular interactions with bureaucrats, be it teachers, doctors, or police officers. By “bureaucrats”
we refer to public employees in the executive branch who are not political leaders. Rather than political leadership,
these public officials are in charge of administrative leadership, service delivery, and other bureaucratic tasks.

The centrality of bureaucrats for politics

Popular images of the bureaucracy are often negative. Bureaucrats are often seen as rigid, unresponsive, and
unaccountable. Yet bureaucrats are central to the working of governments, for better and for worse. While
politicians produce laws and lead public agencies, it is bureaucrats that implement rules and programs. In fact,
bureaucrats are the ones in charge of delivering on the promises of a state, including order and safety (think of police
officers, soldiers, etc.), human development (think of doctors, teachers, etc.), and economic growth (think of
regulators, tax inspectors, etc.). In fact, whatever our view of development, solving contemporary challenges (like
climate change, inequality, or disease) will require bureaucratic capacity. Few of our problems are solvable by
decree. Critically, markets can provide only a very partial solution to these public policy challenges because they
have collective action problems at their core. From this perspective, then, the recruitment and promotion of
bureaucrats is a key challenge to solve for states and for those interesting in making states more capable and more
effective.

Bureaucratic recruitment

Historically, there have been different models of recruiting public sector employees, with important implications for
state capacity and political order. China famously established the first robust meritocratic system, with bureaucrats
being selected on the basis of performance in selective written exams, which historians argued greatly helped
imperial stability. By contrast, European states until the modern era typically recruited civil servants through
personal or aristocratic connections with the monarchs, and understood bureaucratic offices as personal property.
In the early United States, bureaucrats were often recruited through partisan connections. In the extreme spoils
system, presidents replaced most of the federal bureaucracy with co-partisans. This is typically understood as
introducing corruption and rent seeking in the bureaucracy, but also democratic responsiveness.

Throughout the 19th and 20th centuries, many states have reformed their systems of bureaucratic recruitment to
introduce meritocratic recruitment and to reduce the influence of political input. These civil service reforms were
roughly aligned with the way Max Weber understood modern bureaucracies as characterized by meritocratic
recruitment and promotion, strict hierarchy and rules, insulation from politics, and continuity. For Weber, this model
would progressively impose itself due to its advantages for public administration, making states more capable and
effective.

Hiring also differs by the degree of centralization. There are two ideal types: unified and departmental recruitment.
In unified hiring, recruitment is done for the state as a whole. In this tradition, hiring is supposed to be driven by
overall capability rather than fit for one particular post. Rather than joining one particular job in one particular
division, bureaucrats are recruited into a “corps” or body of civil servants, and develop their career with horizontal

1
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

and vertical moves across government units. By contrast, in departmental recruitment hiring is done with an eye
towards identifying bureaucrats with post-specific skills and expertise, with limited mobility throughout their
careers.

Management of bureaucrats

Throughout the 20th century, and in particular since the 1980s, states have experimented with different strategies
to make bureaucracies more responsive and capable. Rather than following Weber’s prediction of making
bureaucracies more hierarchical, tenured, and insulated, states have been introducing market reforms, to varying
degrees and with varying levels of success.

New public management is the term used for the introduction of private-sector-style reforms to public
administration, with components such as competition, performance bonuses, accountability, and autonomy. These
reforms, initially led by the Reagan and Thatcher governments, have garnered support across all sorts of
governments partly due to their electoral appeal. The goal is often to make individual bureaucrats as well as
bureaucratic agencies more responsive, giving them autonomy but also tying their tenure and salary to their
performance.

Outsourcing is a different strategy that consists of contracting out public services to the private sector. The idea
behind this strategy is that by having firms provide services rather than provide them “in-house”, the government
can be more responsive, fast, and efficient. Proponents often make reference to the potential for increased
competition among contractors, which would help depress prices. Additionally, it is assumed that private contractors
would be able to provide the services with more agility since they can use private sector rules to recruit an d manage
their employees. Opponents often point to the lack of accountability, the depletion of government expertise and
know-how, and the potential for corruption in the contracting process as weaknesses of outsourcing.

A final strategy that has emerged in recent decades to manage bureaucrats is called e-government, by which we
refer to the use of information technologies to manage and deliver public services. Online and AI technologies can
be introduced in the provision of information, in interactions with citizens, in transactions by citizens, or throughout
all processes of citizen-government interaction.

Organization of bureaucracies

Bureaucracies everywhere are characterized by high levels of organizational complexity and hierarchy. In general,
executive branches are led by a head of government and their cabinet, composed of ministers or secretaries
appointed by the head of government. That political leadership of the executive branch is typically considered not
the bureaucracy. Under them, there are ministries or departments, which typically focus on one particular area of
policy, such as foreign affairs, finance, education, or defense. Each ministry is typically composed of a series of
hierarchically organized departments, units, or divisions, with bureaucrats closer to the top being more likely to be
politically appointed and directly responsible to the minister.

There is an important trade-off in the degree of penetration of political appointments within ministries. On the one
hand, if most senior officials are civil servants, that may protect bureaucratic expertise and capacity, reducing the
costs of transitions from one government to the next. On the other hand, having more senior officials politically
appointed can reduce problems of bureaucratic unresponsiveness, and conflicts between the interests of
democratically elected political leaders and civil servants.

Of growing importance are non-departmental public bodies, or autonomous government agencies. These are public
organizations that are one-step removed from ministries or departments, and a special legal status that grants them
more autonomy. They are still within the government, and typically report to a minister or secretary, but are not
subject to the same level of hierarchical control as a within-government division. This status is typically used for

2
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

government units that are understood to require extra levels of autonomy, such as research agencies, intelligence
agencies, or regulatory agencies.

Models of bureaucratic governance

There exist two main paradigms from which to talk about bureaucratic governance: Weberian theories and principal-
agent models. It is useful to contrast these two since they provide key, competing insights and ideas for policy
reforms to make our bureaucracies better. Like other social science theories, both perspectives make assumptions
which are to some degree unrealistic. By making those assumptions explicit, we can problematize them and think of
other ways to improve our bureaucracies.

Max Weber articulated the classical theory on modern bureaucracies. In his view, a modern bureaucracy combines
merit-based recruitment and promotion; the movement of bureaucrats across hierarchical and predictable careers,
and life-long tenure. In Weber’s view, public organizations would increasingly adopt this bureaucratic model due to
its superiority in ensuring capacity, efficiency, and speed. From a Weberian perspective, there are three main policy
prescriptions to improve bureaucratic performance. First, staff bureaucracies on merit: recruit based on capacity,
ideally through competitive exams. Second, give bureaucrats tenure: hire bureaucrats for the long-term, and protect
public organizations from turnover. Finally, protect politicians from political interference: politicians' interventions
in the bureaucracy are seen as harmful. These theories therefore have a positive view of bureaucracies, which they
are seen as endowed with a strong ethos and commitment to the state. What this perspective misses is that
autonomous and tenured bureaucrats may become unaccountable – be it because they become lazy, or because
they engage in bureaucratic activism, deciding policies on the margins of political authority. This perspective also
misses that politicians are not always ill-intentioned when seeking to intervene in the bureaucracy, and in fact some
degree of political input may be necessary to ensure responsiveness to democratic demands.

Principal-agent theories provide an alternative view, and set of prescriptions, on bureaucratic governance. Principal-
agent models are used in microeconomic theory to model the relationship between a principal who delegates a task
and an agent who receives the authority to implement that task. Examples of principal-agent relationships include
that between an employer and an employee, or a politician and a bureaucrat. These theories see delegation as a
contract. The principal wants the agent to act in the principal's best interest, but agency problems arise because of
imperfect contracting. A first problem is adverse selection: the principal may unknowingly hire the wrong type (e.g.,
someone who's not very capable). A second problem is moral hazard: the agent acts against the principal's best
interest (e.g., slack). A third problem is information asymmetry: the agent knows things that the principal ignores,
which hinders monitoring. From a principal-agent perspective, the policy prescriptions to improve bureaucratic
governance are fundamentally different from the Weberian recipe. First, you would want to ensure the right types
are selected: since there are agency problems, make sure to select agents who will be motivated to act in the
principal's best interest. Second, this perspective calls for setting up the right incentives: make sure to set up material
incentives so agents are pushed to deliver. Finally, you would want to reduce information asymmetries (by improving
the quantity, quality, and flow of information to facilitate monitoring and incentives), and to implement reforms of
new public management including pay for performance, decentralization, and competition. The principal-agent
approach to public bureaucracies is not without problems, either. First, bureaucrats often operate in policy spaces
where markets fail, which means market reforms likely hard to implement. Second, many important tasks in a public
bureaucracy are hard to monitor and thus incentivize; bureaucrats in fact operate in a complex task space. Finally,
principals may not be capable or able to hold agents accountable.

The costs and the benefits of patronage

While the term patronage is used with a variety of meanings, in the context of bureaucratic governance it refers to
the political allocation of public employment. A more precise definition of patronage is the discretionary hiring and
firing of bureaucrats based, at least partly, on political criteria. An extreme version of patronage is the so-called
spoils system, where most of the administration is changed with the winner's political supporters after political
turnover.

3
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

All political systems have some degree of patronage. For example, advisors and leaders of administrative units are
typically political appointees. This presence of patronage in the bureaucracy arguably helps politicians transmit
mandates and effectively lead public administration to implement their chosen policies. That said, there is very wide
variation in the prevalence and nature of patronage across and within countries.

Patronage can of course be costly for governance and for development. First, patronage can lead to the selection of
worse types – in fact, political appointees are often less educated, less capable, or both, and there may well be a
competence – loyalty tradeoff. Second, patronage can lead to the misallocation of bureaucrats, deploying connected
people to more desirable / better paid jobs. Last but not least, patronage can depress bureaucratic effort: if
government jobs are used to transfer public money to the pockets of political supporters, government employees
would work less on actual tasks.

On the other hand, patronage can have benefits too.

Demonstrating the costs and benefits of patronage empirically is challenging. Simple regressions of cross-country or
within-country data are unlikely to pick up the effects of connections. For instance, while cross-country regression
show a positive association between meritocratic recruitment and economic growth, it is unclear whether this is
driven by confounders (such as levels of economic or educational development) or reverse causality.

A recent study of the costs of patronage uses quasi-experimental variation of connections between British colonial
governors and the secretary of state in London to demonstrate that patronage can hurt bureaucratic effort,
performance, and development. 1 The author conceptualizes patronage as social connections between bureaucrat
and political superior, and bureaucratic performance as the results of the territory covered by the governor. In
particular, the author examines the effect of whether the governor and the secretary of state share aristocratic,
family, or elite schooling backgrounds, on the tax revenue of the governorship. The empirical design focuses on
comparing the performance of connected versus unconnected governors, exploiting the turnover of the secretary
of state in London to focus on within-governor changes. In essence, his work shows that connected bureaucrats
perform worse.

In contrast, another recent study uses quasi-experimental variation of connections between municipal school
directors and mayors across Brazil to show that patronage can, in some circumstances, increase bureaucratic
effectiveness and accountability. 2 The author’s theory is that patronage can change how bureaucrats work, providing
some key governance resources: monitoring, the ability of applying sanctions and rewards, alignment of values and
incentives between bureaucrats and politicians, and mutual trust. A first empirical design focuses on within-director
changes in political connections induced by the turnover of the mayor in close elections. By comparing what happens
after mayoral turnover to the performance of schools with appointed and unappointed directors, the author shows
that connections can facilitate school performance. A second empirical design focuses on the effect of meeting a
school quality performance target on the likelihood that appointed vs unappointed directors. By showing that
appointed directors who miss the target are more likely to be replaced, whereas unappointed directors see no effect,
the author shows that patronage can facilitate bureaucratic accountability.

Questions for reflection and discussion

 How do bureaucrats matter for the different dimensions of state capacity, i.e. coercion, coordination,
compliance, legibility?

1 Xu, Guo. "The costs of patronage: Evidence from the British empire." American Economic Review 108.11 (2018):
3170-3198. http://guoxu.org/docs/EmpireJMP_Xu.pdf
2 Toral, Guillermo. "How patronage delivers: Political appointments, bureaucratic accountability, and service

delivery in Brazil." American Journal of Political Science 68.2 (2024): 797-815.

https://www.guillermotoral.com/benefits_of_patronage.pdf
4
 Comparative Politics Guillermo Toral
BBA BIR Assistant Professor of Political Science
Fall 2024 www.guillermotoral.com

 What makes bureaucrats more or less effective, and how can governments be reformed to nurture more
capable bureaucracies?
 How should we think of the tension between elected leaders and bureaucracies? Can politicians be a
positive influence on the bureaucracy?

Required readings

 McCormick, John, Rod Hague, and Martin Harrop. Comparative Government and Politics. Bloomsbury,
2019. Chapter 10.

Suggested materials for further exploration

 Blog post – “Getting someone else to do the task: When accountability initiatives create shadow states
instead of building bureaucratic capacity.” Political Science Now: https://politicalsciencenow.com/getting-
someone-else-to-do-the-task-when-accountability-initiatives-create-shadow-states-instead-of-building-
bureaucratic-capacity/
 Dataset – Worldwide Bureaucracy Indicators, World Bank:
https://datacatalog.worldbank.org/search/dataset/0038132/worldwide-bureaucracy-indicators
 TV show – The Wire (especially season 3)