DEMOCRATIZATION

https://doi.org/10.1080/13510347.2021.1960314

RESEARCH ARTICLE

Costs and benefits of accepting presidential term limits:

“should I stay or should I go?”
a b
Anna Fruhstorfer and Alexander Hudson
a
Faculty of Economics and Social Sciences, University of Potsdam, Potsdam, Germany; bDemocracy
Assessment Unit, International IDEA, Stockholm, Sweden

ABSTRACT
As presidents approach the end of their constitutionally defined term in office, they
face a number of difficulties, most importantly the deprivation of sources of power,
personal enrichment, and protection from prosecution. This leads many of them to
attempt to circumvent their term limits. Recent studies explain both the reasons for
the extension or full abolition of term limits, and failed attempts to do so. Key
explanations include electoral competition and the post-term fate of previous post
holders. What we do not know yet is how compliance with term limits may be tied
to the current president’s expectations for their post-term fate. In particular, we do
not know whether leaders who attempt to remove term limits and fail to do so
jeopardize their post-term career as a result, and conversely, whether leaders who
comply will have better outcomes in terms of security, prestige, and economic gain.
Hence, we ask how the decision of a leader to comply or not comply with term
limits is conditioned by the expectation of their post-term fate. To address this
question, this article introduces new data on the career trajectories of term-limited
presidents and its systematic effect on term limit compliance.

ARTICLE HISTORY Received 11 December 2020; Accepted 1 July 2021

KEYWORDS Presidents; head of state; term limits; executives; corruption; prestige; institutional change;
constitutions

Introduction
An aspect of presidential term limits that has to this point probably received more
attention in discussions in quiet corners of pubs than in the pages of scholarly journals
is what presidents (and in particular those with authoritarian sympathies) might actu-
ally do if they retired. Few, it seems, agree with Benjamin Franklin that it would be a
promotion to leave office and return to being a citizen-ruler.1 The examples are
obvious, some hypothetical, some all too real. What might Vladimir Putin have
done with the remaining years of his life if the Putin-Medvedev tandem in 2008 had
not worked? Similarly, as scandals rocked his administration (and a hard term limit
approached), many in South Africa mused that Jacob Zuma might spend the rest of
his days in quiet luxury in Dubai (his mansion was in fact a few doors down from

CONTACT Anna Fruhstorfer [email protected]

© 2021 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives
License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduc-
tion in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way.
2 A. FRUHSTORFER AND A. HUDSON

Robert Mugabe’s).2 At the same time, we can only think of a few presidents who were
able to actually maintain power in a different capacity after leaving office, even less
when they first attempted to extend their term.
What most of the previous work on executive term limits shares is an assumption
about the motivations of the term-limited executive. One of the dominant views of the
interests of holders of political office is that they seek to stay in power as long as poss-
ible.3 This is certainly a justifiable assumption, and one that has held currency at least
since the eighteenth century.4 Even a passing acquaintance with the political class in
both democracies and autocracies confirms that the self-interest of many executives
and legislators dictates that they attempt to remain in office. The specific reasons for
this evaluation of their interests vary, ranging perhaps from a genuine belief that
they are the best person for the job, to worries about exposure to criminal prosecution
or civil liability after leaving office.
However, it is certainly not always the case that an executive would view staying in
office as their best option. Many former executives do quite well for themselves after
leaving elected office. US presidents all the way back to Ulysses Grant have made sig-
nificant amounts of money through writing memoirs and giving speeches after leaving
office.5 Former presidents do not only pursue wealth, but sometimes also have interests
in personal prestige and reputation. After leaving office following both her first and
second (non-consecutive) terms, Michelle Bachelet held prestigious posts with the
United Nations (UN) first as head of UN Women, and now as UN High Commissioner
for Human Rights.
In this article, we hold an executive’s evaluation of their best option to be an open
question, and pursue the possibility that in different political contexts, even executives
with autocratic ambitions would choose to leave office as established term limits would
dictate. In essence, we ask: When will executives follow term limits? We argue that
leaders who attempt to remove limits and fail to do so, will jeopardize their ability
to find a suitable, prestigious, or financially attractive post-term activity. Yet, presi-
dents’ willingness to comply with term limits could be rewarded with more attractive
possibilities, and anticipation of these two options will have an impact on a president’s
decision.
This article begins with a short review of the current state of the field in research on
term limits. There have been a number of notable contributions to our knowledge in
this area even in the last two years, and this review helps to place our contribution
within this growing field. We then introduce some new data in which we systematically
code the post-office activities of former presidents. We use these new data in combi-
nation with data collected by McKie6 and Baturo7 to model the conditions under
which term limits will be respected, with particular attention to the soft or hard
landing that former executives manage.

A theory of term limit compliance

Research on term limits for executives has taken on a new urgency in the past several
years, as events in a number of countries have provoked scholars to consider the possi-
bility that we are entering an era of widespread democratic backsliding.8 For most of
the academic debate, the positive side of constitutional term limits (mainly with regard
to presidents) and their impact on successful and sustaining democratization has been
emphasized. In line with previous findings,9 scholars have consistently found a positive
 DEMOCRATIZATION 3

effect on the rule of law and horizontal accountability, but also as protection against
democratic backsliding or re-autocratization.
Research on formal institutions shows that the stability or instability of an insti-
tution can either be affected by political power dynamics,10 or create its own stability
through self-enforcement and/or creating high costs for change.11 While the scholar-
ship on institutional stability in broader terms is reasonably clear, research on the stab-
ility or instability of term limits in particular is still largely focused on individualist
explanations about presidential behaviour and credentials (for example, military
ties) and popularity or support from the ruling party.12 However, some more recent
studies have emphasized a rather institutionalist perspective on the power relations
that encourage term limit evasion or compliance.13 While structural explanations
usually focus on the political, social or economic context of a decision, agency based
explanations emphasize individual human behaviour. Moving between these classical
approaches, we view presidents as strategic actors, but also emphasize that they usually
pursue complex goals without complete information and that their actions occur
within institutions that favour certain strategies over others.14 These institutional pre-
ferences are in turn shaped by the context and also by the interests of the actors’ roles
within these institutions.15 Therefore, we must consider the interaction between insti-
tutions and individual decision making.16
While previous studies have made an important contribution by shedding light on
this structure-agency relationship, particularly between term limit compliance, party
system and individual characteristics of presidents, we still do not know if or how
the decision to comply or not comply with term limits is tied to the respective presi-
dent’s post-term fate.17
To address this lacuna, we argue that leaders who attempt to remove limits and fail
to do so, seriously jeopardize or limit their post-term opportunities as a result.
However, leaders who do comply will have better chances for prestigious or economi-
cally beneficial post-term activities. We explain the effect as follows: First, building on
the literature on career choice and career trajectories of leaders,18 we posit that presi-
dents have some knowledge about their potential post-term activities when they decide
to stay or to go. Second, building on institutional arrangements theory,19 we argue that
the institutional setting in which the president encounters a term limit and its contex-
tual embeddedness within the country and the region condition the post-term fate of
these head of states.
In particular, we rely on arguments about term limit compliance in the literature on
electoral hegemony and the intra-elite cohesion thesis,20 but define a different causal
mechanism. In this way, we link institutional arrangements, partisan power-balances,
and intra-elite cohesion with the question of term-limit compliance or non-compli-
ance. We argue that these contextual factors causally influence the probability of
post-term prestige, success, and safety for term-limited presidents.
Bringing these arguments together, we posit that not all post-term activities are
open to all presidents due to both their own varying individual characteristics and
because of the prevalent social and political conditions in their respective home
country. As a result, presidents with less-promising prospects find term limit compli-
ance to be unappealing, while the prospect of post-term wealth, prestige or immunity
will make a president more likely to comply with term limits. We thus argue that post-
term career options are framed by the specific characteristics of a given country and
that these options then in turn shape the responses to an approaching term limit.
4 A. FRUHSTORFER AND A. HUDSON

Personalities, incentives, and career calculations

While it takes a certain kind of person to pursue the highest political office available at
the domestic level, presidents are in the end persons like the rest of us. They will
pursue, as Maslow described it, goals to fulfil their needs that relate to “physiological
[concerns], safety, love, esteem, and self-actualization.”21 Presidents face a limited
choice set as they approach the end of their term, with opportunities to maximize
one or two of these needs, but not all. First, whether motivated by physiological
(wealth) or safety (perhaps from prosecution) they may pursue continued political
power. A focus on physiological concerns could also drive an individual to pursue
opportunities outside formal political office. Second, feeling reasonably secure in
those matters, an individual may be led by needs for love, esteem, or self-actualization
to pursue opportunities that would increase their personal prestige, or allow them to
feel that they are contributing to some higher good.
While former presidents aim to protect their interests, their ability and willingness
to do so depends on their post-term fate.22 We expect that the post-tenure career
opportunities available to former leaders will have a different, mitigating effect.
Baturo has provided evidence for a general theory of term limit compliance for presi-
dents in democratic regimes, arguing that compliance may be higher when presidents
receive some benefit from their position after they leave office.23 He specifically argues
“that incentivizing political leaders with post-tenure opportunities and socializing
them into prestigious international forums may indeed be an important policy tool
to entice leaders to accept democratic norms.”24 Conversely, clinging to power is
related to the lack of post-term opportunities, for example, not finding a suitable
place for exile,25 or the looming loss of assets.26 Yet, we already noted a limited
choice set above. Specifically, not all opportunities are available to all executives.
While Michelle Bachelet had the opportunity to become the executive director of a
new United Nations entity after she left office, we suspect that such an office would
not have been offered to Hugo Chavez (for example).
Similarly, opportunities for generating wealth in “retirement” are linked to various
aspects of the state, such as overall level of wealth, extractive industries, and levels of
corruption. Several authors have pointed to both personal attributes and endogenous
circumstances as influences on the level of corruption.27 From a purely observational
point of view, the individual “well-being,” “status” and “enormous economic advan-
tages”28 that come with a position of power are an obvious reason for trying to stay
there. Baturo emphasized this point as follows: “It is not sufficient to focus on how pol-
itical institutions constrain or empower presidents who desire to take over political
regimes. Instead, we also need to understand what is it that individual presidents
can gain or lose, in or out of political office.”29 More broadly, scholars have been
seeking to understand levels and predictors of corrupt behaviour for some time. Yet,
the difficulty of gathering reliable cross-country data has been a serious impediment,
and we still know comparatively little about the incentives for powerful people to
pursue corrupt or dishonest activities.

Institutional arrangement and political context

If one considers term limit compliance and evasion as deeply individualistic reac-
tions, the question of economic opportunities for the person in power is key.
While institutional constraints might forestall corruption, we argue that there is a
 DEMOCRATIZATION 5

relationship between individual enrichment, protection against prosecution, and rule

compliance. While from a comparative-historical perspective we know that since
World War II the post-term fate of dictators often means exile,30 recent research
has shown that the precedents set by a previous post-holder provide a clear forecast-
ing tool for the decision of any successor (also Hartmann in this special issue).31 The
level of actual threat to a political party’s future electoral success has a profound effect
for the stability of political institutions,32 and more specifically for institutional
choice.33 It is also vital to explain the probability of achieving different post-term pos-
itions while considering the probability of an attempt at term limit evasion. With
regard to term limits, McKie shows that the level of electoral competition has a sig-
nificant effect on term limit adherence.34 And, while the degree of electoral compe-
tition is a key factor, it is as yet unclear whether the endogenous aspects that lead to
changes or stability in electoral competition are more important for the acceptance of
term limits.
Furthermore, within-party hegemony plays an important role for a fluctuation in
electoral competition (see, for example, in Taiwan, Mexico, Russia, or Azerbaijan).
Parties are relevant institutions in modern political systems, and we increasingly see
their importance in the survival of autocratic regimes.35 In particular, ruling parties
co-opted by leaders contribute to the stability of the hegemonic power distribution.36
Scholars have also argued that autocratic leaders rely heavily on political institutions,
most notably partisan legislatures, to broaden the base of their support.37 Yet, for this
to be effective, the legislature has to be dominated by a more or less hegemonic party,
and the party system should have reached a minimum level of institutionalization.
Without that, and despite an already low level of democracy, the legislature is a poten-
tial source of opposition to the autocratic leader and his hegemonic rule. When it
comes to the role of parties within the organization and preservation of a hegemonic
elite, the institutionalization and dominance of one party is a key variable that differs
markedly, for example, in the post-Soviet space. In this sense, we assume that parties –
both in democracies and autocracies – usually constrain the president. If these parties
are not at all or only insufficiently institutionalized, the checks upon the presidency
decrease.
Furthermore, McKie has shown that “the power asymmetry affects the intra-ruling
party dynamics between the president, ruling party legislators, and the judiciary” and
she relies on differences in the “margin of victory.”38 While this is a key point, we think
that to fully engage with the mechanisms underlying this process, intra-party and
inter-institutional dynamics have to be understood within the logic of not yet conso-
lidated democratic regimes. Tying parties as “strategically-relevant actors”39 to the
leader or leadership group creates an incentive for potential opposition intra-elite
actors to avoid obstruction,40 and ensures “intra-elite cohesion” even within the
formal opposition.41 We assume that formal co-optation of parties and within the leg-
islature in particular influences the paths that presidents are willing and able to pursue
when they encounter a term limit.

Expectations concerning the interaction between institutional arrangements,

personal incentives, and term limit compliance
As we investigate the relationship between the choice that a term-limited president
may make between extending their term or pursuing other opportunities, an
6 A. FRUHSTORFER AND A. HUDSON

intriguing potential endogeneity becomes apparent. First, it seems clear that presi-
dents have some knowledge of their potential post-term activities when they make
a decision about whether or not to seek to overcome the term limit in some way
(time 1). Yet because this information is far from complete, actors’ decisions are
shaped by their perception of the institutional context. In addition, as we observe
their choices again at time 2, the choices available to them have been limited by
the interaction between their decision at time 1 and various factors relating to the
political context. Thus, term limit compliance (or non-compliance) is the result of
individual action, filtered by the limited information and perception of the insti-
tutional context: A choice to attempt to extend their term is likely to close a
number of doors in international organizations, jeopardize their ability to return to
politics in the future or to make money in other ways (both above and below
board), and may even increase the likelihood of criminal prosecution. The underlying
mechanism can be best illustrated by a specific case. Former President Evo Morales
was a decisive figure in Bolivian politics for close to 15 years. We observe several legal
and illegal attempts to extend his time in office that deeply influenced his post-term
options. The new constitution in 2009 already allowed Morales to evade his first term
limit, yet this would probably not have posed serious issues if he had intended to
comply with the new term limit provisions. One could imagine that the first indigen-
ous president in the Americas – as the embodiment of Latin American self-govern-
ance and the true end of colonialism – would have had a myriad of prestigious
opportunities in the international realm, had he wanted to pursue them. Yet, the
different attempts varying in their legality and their political success to evade
approaching term limits (with a proposed constitutional amendment, a failed refer-
endum, yet an accommodating ruling by the Plurinational Constitutional Court) pro-
ceeded to limit the options available to Morales and thwarted any hope for a
prestigious and influential post-presidential career. Similarly, presidents with a
heavily co-opted opposition, will be first more inclined to attempt a term limit
evasion, but may also be more likely to pursue different career choices (i.e. retaining
power outside of formal office).
The international reputation of the country may also matter. Influential inter-
national opportunities are likely to be more available to executives whose countries
have a relatively higher level of international prestige (or more soft power). Inter-
national prestige opens doors to international organizations and other means of main-
taining something like soft power in a personal sense as well as reputation. Presidents
coming from countries with a relatively high level of international prestige may be less
likely to attempt to overcome term limits in the first place, as they will have a keener
interest in remaining a respected public figure when they leave office. Additionally,
measures of national prestige are likely to be closely tied to perceptions of corruption,
human rights abuses, and other contextual factors that we would expect to be associ-
ated with respect for term limits.
It is also likely that the means a president uses in a term-limit evasion attempt will
have some impact on their post-term prospects. For example, a president who achieves
a constitutional amendment extending their term is likely to have better post-
extended-term prospects than one who suspends the constitution or uses an auto-
golpe to overturn the results of an election.
 DEMOCRATIZATION 7

Data
Previous research has provided ample evidence that presidents often attempt to cir-
cumvent their constitutionally assigned term limit.42 Since 2000, approximately half
of the presidents in hybrid and authoritarian regimes attempted to overcome a term
limit.43 A key reason for the attempts at non-compliance is the dangers some of
these presidents face after losing their power. Rationally, they have strong incentives
to stay in office longer – for example in case of patrimonialism or corruption. And
according to McKie, nearly 80% of the attempts to evade their term limits made by pre-
sidents of third-wave democracies have been successful.44
In this article we build on different datasets, particularly the data collected by
McKie, and her article on the approaches that would-be term limit evaders have
taken between 1975 and 2018.45 Each observation in these data is an incident in
which a president in Latin America, Africa, the Middle East, or Asia encountered a
term limit. She includes all heads of state in a presidential or semi-presidential
system,46 when the country is at least an electoral democracy and belonged to a
region that transitioned in the third wave of democratization (i.e. Latin America;
Sub-Saharan Africa; Central, South, Southeast, and East Asia; Eastern and Southern
Europe; Middle East and North Africa). These national presidents of individual and
sovereign states are the unit of analysis.
We add to this our own novel data on the post-term occupations or other personal
outcomes for these presidents. We coded the post-term position of the individual
presidents into exclusive categories as follows: (1) presidents who stayed in a
formal political office with high degree of power (for example as prime minister
like Vladimir Putin in Russia or Robert Kocharyan in Armenia), (2) presidents
that maintained power outside formal office (for example as party leader like Jean-
Bertrand Aristide in Haiti or Daniel Ortega in Nicaragua), (3) presidents who left
politics but retained some public profile as speaker, scholar or, lobbyist (for
example Fernando Henrique Cardoso in Brazil), (4) presidents that retired or other-
wise disappeared from public view (for example Carlos Salinas de Gortari in Mexico),
(5) presidents who took on some role in an international or intergovernmental
organization (for example Michelle Bachelet in Chile), or (6) presidents who were
convicted of a crime and sentenced to house arrest or prison (for example, Alberto
Fujimori in Peru).

Dependent variables
We are interested in what presidents do after the end of their term, focusing especially
on the post-term career trajectories of those who complied with term limits or failed to

Table 1. Term limits evasion attempts and post-term activities.

No attempt Attempt
Continued as president 0 18
International organization 13 0
No political office but public profile 38 6
Other formal political office 15 4
Outside formal office 29 6
Retired 48 20
Imprisoned 14 4
8 A. FRUHSTORFER AND A. HUDSON

stay in office beyond their constitutionally mandated term. Here, we complement the
data provided by McKie47 with new data on the post-term activities of these presidents
(see Tables 1 and 2). The dataset includes only those former presidents who encoun-
tered a term limit and have thus had to make a choice about whether they will leave
office (as the law requires) or attempt to extent their time as the chief executive.
The data are coded as no attempt, failed attempt, successful extension by one term,
and finally a successful extension with a full abolition of term limits.48 Table 1
shows the frequencies of various post-term outcomes and their relationship to
attempts to extend or abolish term limits.
In the statistical models that follow, we estimate the predictors of various kinds of office
that a former executive might take on (including retiring and continuing in office). As
Table 2 shows, outcomes are not always straightforward. For example, there are a
number of cases where a president was successful in changing the law to allow for an
extended term, but their actual post-term activity was different. The data for the coding
were drawn from media accounts, country reports and secondary literature.

Independent variables
Corruption
The overall level of corruption in a political system is closely related to term limits
compliance. At a basic level, a more corrupt system potentially allows for greater
levels of rent seeking on the part of the president. In one sense, this reduces their like-
lihood to abide by term limits, as they have incentives to remain in power to retain
access to these rents.49 At this level, we would expect higher levels of corruption to
predict attempts to evade term limits. However, corruption is also related to the activi-
ties that are available to presidents in the post-term period should they either decide
not to attempt to extend their term or fail in the attempt. Specifically, we would
expect to find that executives in more corrupt systems are more likely to find ways
to remain influential or extract rents. In our categorical coding of post-term activities,
we would expect to find these former presidents maintaining power outside of formal
office, or perhaps occupying some other formal office. The “retirement” category is
particularly interesting here, as we expect that the category may capture some forms
of influence that are not publicly reported.

Electoral competition, cooptation, and party dynamics

We know that leaders in weakly or non-democratic regimes also rely on partisan
legislatures, dominated by a more or less hegemonic party and a certain level of

Table 2. Term limit decisions and post-term activities.

Extend term Fail to extend term Lift term limits Step down
Continued as president 18 0 0 0
International organization 0 0 0 13
No political office but public profile 1 3 2 38
Other formal political office 0 2 2 15
Outside formal office 4 2 0 29
Retired 6 5 9 48
Imprisoned 3 1 0 14
 DEMOCRATIZATION 9

institutionalization.50 McKie has specified this argument and emphasizes the

“margin of victory” to explain term limit evasion and compliance.51 We expand
this argument as we think in the case of hybrid and autocratic regimes (the most
common sites of term limit evasions), inter-party relations and partisan dynamics
function in a substantially different way that cannot be measured only by the
margin of victory. We argue that presidents in these regimes frequently employ
different strategies to stay in power – and how they engage and treat possible oppo-
sition parties and potential challengers is key to understanding regime continuity
and survival.52 What they often do is create or engage strongly with quasi-demo-
cratic institutions like parliaments to co-opt possible opposition and guarantee
the cooperation of potentially threatening groups or individuals. With V-Dem
data on opposition party autonomy we test the effects of variation in the indepen-
dence and autonomy of opposition parties and their level of being selected or co-
opted by the ruling regime.53 We also assume that these broader dynamics may
be further complicated by the continuing role of the president as the founder of
the party (described further below).

Control variables
Economy
While we do not control for the economic situation per se, we have to consider litera-
ture that points to the influence of economic factors for the probability of corruption, a
broader, yet related frame to our “post-term activities” variable. We add control vari-
ables to our models that account for the level of GDP growth, and foreign aid as a share
of GNI in the year that the president’s term expired (usually the year of the attempt to
extend the term if one is made).54

Party institutionalization
At least two aspects of the political parties themselves (especially those tied to the pre-
sident) could matter for our statistical models. First, in cases where the president is the
founder of the political party, it is likely that they have a higher level of control over the
party. We suspect that this will make them both more likely to attempt to extend their
term, and more likely to succeed. We use data from Baturo and McKie to include a
binary coding of the president as party founder to control for this.55 Second, we
include the strength of the political party with a variable that accounts for “level and
depth of organization, links to civil society, cadres of party activists, party supporters
within the electorate, coherence of party platforms and ideologies, party-line voting
among representatives within the legislature.”56 We expect that party institutionaliza-
tion should be negatively related to attempts to alter or evade term limits, but positively
related to success when an attempt is made.

Level of democracy and horizontal accountability

To account for differences in the level of democracy, we include two measures that rep-
resent different perspectives on autocracy and democracy (both developed by V-Dem).
The high-level democracy measure of polyarchy is focused on the question of the
extent to which the ideal of electoral democracy is achieved. It combines core electoral
principles of democracy, effect of election on chief executives as well as the free oper-
ation of civil society, freedom of expression and media.57 We add to this a measure of
10 A. FRUHSTORFER AND A. HUDSON

horizontal accountability. Here we are interested in how well the checks between insti-
tutions, i.e. legislature, judiciary, and other oversight institutions, that prevent the
abuse of power, are established.58

Prior attempts
Not every observation in our data is a new individual. On the contrary, there are a
number of repeat players who, having successfully navigated either a graceful exit or
successful relaxation of constitutional term limits, encounter a term limit a second
or third time. We think that this learning experience will be highly predictive of
further attempts to evade term limits, and will also close a number of doors outside
politics. To both control for the ways in which these individuals are different from
single-incident individuals, we created a dummy variable that identifies term-limit
incidents that are an individual’s second (or more) encounter with these constitutional
limits. The experience of previous presidents may also be instructive for incumbents.
For example, if a previous president (perhaps even in a previous constitutional system)
was able to successful evade a term limit, this could be seen to erode the significance of
the term limit as a norm, not just as a law. In the opposite direction, a notably negative
outcome for a previous president might reinforce the norm of respecting term limits.
To that end, we also include a dummy variable that records whether or not a previous
president was imprisoned, exiled, or even assassinated.59

Method of attempt
We also account for the manner in which a term limit evasion was attempted and
differentiate between legal and illegal means of evasion. For this, we distinguish
between term limit removal or relaxation via a constitutional amendment, and a
term limit evasion by other means. We expect that the facial legality or rule-bounded-
ness of the term-limit evasion attempt will be influential on the post-term options
available to a president.

Statistical analyses
In the statistical models presented in Table 3 we consider the likelihood of each of
seven different potential post-term outcomes for presidents as a binary outcome.
We face some challenges in terms of rare combinations of events and use a generalized
linear model with bias reduction to estimate these relationships.60 Each model com-
pares that particular post-term activity to all the alternatives. The results here in
some ways serve to highlight the limitations of studying this particular question in a
cross-national quantitative approach, as we find that few potential predictors of
post-term outcomes reach statistical significance.
The first column in Table 3 presents some predictors of successfully continuing in
office. The most notable covariate in this model records whether or not this attempt to
extend or abolish term limits involved a constitutional amendment. Unsurprisingly,
this is strongly predictive of remaining in office, with presidents who are able to
amend the constitution 4 times more likely to be successful than those who do not
change the constitution. However, we should note that there are still a number of
cases where alternative means have been used, most notably via court rulings. We
also find that horizontal accountability is negatively related to the ability of a president
Table 3. Logistic regressions on post-term activities (odds ratios).
Dependent variable
Continued Retired Int’l org. Non-political public Other formal office Power outside form. office Imprisoned
(1) (2) (3) (4) (5) (6) (7)
Attempt to extend 0.941 0.116 0.553 1.815 1.150 0.865
(0.423) (0.175) (0.321) (1.225) (0.654) (0.600)
Prior attempt 1.362 1.092 1.109 1.713 0.759 0.436 0.475
(1.105) (0.583) (1.566) (1.091) (0.712) (0.394) (0.447)
Const’l amend. 4.606** 0.974 0.305 0.580 0.389 0.539 0.373
(2.892) (0.521) (0.659) (0.464) (0.386) (0.442) (0.364)
Horizon. account. 0.402* 0.915 2.166 1.090 1.001 1.297 1.227
(0.221) (0.287) (1.431) (0.411) (0.526) (0.545) (0.616)
Corruption 1.612 1.130 0.015** 0.769 1.794 11.786** 1.816
(2.839) (0.962) (0.025) (0.766) (2.410) (14.451) (2.470)
Oppo. autonomy 1.246 1.005 0.186*** 1.117 1.589 1.777 0.771
(0.442) (0.237) (0.097) (0.345) (0.721) (0.661) (0.294)
Party founder 0.956 1.106 0.662 0.750 0.437 0.965 2.993**
(0.593) (0.387) (0.512) (0.326) (0.291) (0.449) (1.462)
GDP growth 0.943 0.959 0.962 1.009 1.031 0.948 1.054
(0.063) (0.041) (0.085) (0.054) (0.075) (0.058) (0.065)
Aid % of GNI 0.998 1.011 1.032 0.988 0.935 1.011 0.968
(0.038) (0.018) (0.029) (0.022) (0.049) (0.023) (0.037)
Constant 0.039** 0.466 7.152 0.318 0.051** 0.017*** 0.090*
(0.055) (0.364) (11.012) (0.315) (0.070) (0.022) (0.114)
Observations 215 215 215 215 215 215 215
Log Likelihood −42.943 −132.255 −36.216 −100.667 −57.001 −82.603 −57.830
Akaike Inf. Crit. 103.887 284.510 92.433 221.334 134.002 185.206 135.661
Notes: *p < 0.1; **p < 0.05; ***p < 0.01.
DEMOCRATIZATION
11
12 A. FRUHSTORFER AND A. HUDSON

to continue in office.61 A one-unit increase in horizontal accountability (which in our

data ranges from −1.7 to 1.9) is associated with a roughly 60% reduction in the like-
lihood of the president successfully evading a term limit and remaining in office.
This is what we expect as the horizontal accountability measure indicates whether
key institutions are able to prevent the abuse of power.
Our sixth outcome is retaining power outside of formal office. This is probably the
most normatively troubling of the various post-term activities in our coding scheme. It
is not an altogether uncommon outcome, with 35 of the presidents in our data finding
a way to retain power while not occupying a formal political office. Included in that list
are a number of well-known (some infamous) presidents, including Corazon Aquino
(Philippines), Jean-Bertrand Aristide (Haiti), Leonid Kuchma (Ukraine), Olusegun
Obasanjo (Nigeria), and Daniel Ortega (Nicaragua). We do in fact find a very
strong connection between retaining power outside formal office and corruption.
This measure runs from 0 to 1, and our results indicate that the odds of a president
maintaining power outside formal office are 11 times higher moving from a non-
corrupt system to a fully corrupt system. At the same time, we do not find any
effect of opposition party autonomy.
Another outcome of substantive interest is the likelihood that a term-limited pre-
sident who is unable to remain in office will be imprisoned. Here we find that being
the founder of the ruling political party is positively associated with imprisonment,
and strongly so. While this may at first seem counterintuitive, it is likely that presidents
of this type leave office in under circumstances that entail a more complete defeat.
Moreover, their former influence may make them more likely to be targeted for pro-
secution by their political opponents, both from within the elite but also from the
opposition. Considering that the majority of parties in autocratic regimes do not
survive the first leadership change,62 we cannot expect that the party will be able to
provide protection to its former leader.
However, these models suffer an important limitation in the extent to which they
capture the theory we presented above. Obviously, it is not possible for a term-
limited president (i.e. those present in our sample) to continue in office without
making an attempt to extend or abolish term limits. Thus far, we have controlled
for this by including a dummy variable for attempts. Yet, this treats the attempts as
being exogenous to the outcomes we observe, which we know is not the case.63 A
better approach is to model the endogenous process through which an individual
enters the sample of cases in which an attempt was made, and then to model the pre-
dictors of the varying outcomes of that result.64 We can achieve this with a two-step
Heckman sample selection model in which attempts to extend or abolish term
limits are considered in a selection model, which then provides starting values for esti-
mating an outcome model of post-term activities.65
In Table 4 we present the probit selection model (predicting attempts to extend or
abolish term limits), and in Table 5 we present the type-two tobit outcome model (pre-
dicting post-term activities). In this implementation, the selection model is identical
for each outcome, but we have treated each of the outcomes as a binary dependent
variable, estimating an outcome model for each of them. Compared with the exhaus-
tive treatment in Table 3, we have simplified the outcomes here, and group the out-
comes into the broader categories of power, prestige, retirement, or imprisonment.
As is necessary, we vary the predictors somewhat between the selection and
 DEMOCRATIZATION 13

Table 4. Selection model: attempt to extend term limits.

Dependent variable
Term limit evasion attempt
Prior attempt 0.934**
(0.380)
Fate of previous pres. −0.040
(0.232)
Polyarchy −1.925
(1.177)
Horizontal accountability −0.621***
(0.235)
Corruption −0.900
(0.753)
Opposition autonomy −0.081
(0.229)
Party founder 0.557**
(0.253)
Party institutionalization −0.597
(0.666)
Constant 1.582*
(0.938)
Observations 214
Rho −0.893
Inverse Mills Ratio −0.491 (0.500)
Notes: *p < 0.1; **p < 0.05; ***p < 0.01.

outcome models (most notably accounting for the fate of previous presidents in the
selection model, and the legality of the attempt in the outcome model).
To account for the expectations a president may have when they make their choice
to attempt a term-limit evasion, the selection model includes a binary indicator of

Table 5. Outcome model: post-term activity.

Dependent variable:
Retain power Retain prestige Leave public life Imprisoned
(1) (2) (3) (4)
Const’l amend. −0.010 −0.095 −0.001 −0.060
(0.126) (0.078) (0.132) (0.067)
Prior attempt −0.399 0.007 −0.189 0.021
(0.264) (0.157) (0.250) (0.133)
Polyarchy 1.280 0.713 −0.168 −0.410
(0.899) (0.538) (0.869) (0.457)
Horizontal account. 0.223 0.055 0.177 −0.052
(0.235) (0.142) (0.231) (0.121)
Corruption 1.226** −0.008 −0.273 −0.207
(0.531) (0.322) (0.532) (0.276)
Oppo. Auton. 0.018 0.001 −0.055 0.001
(0.108) (0.064) (0.102) (0.054)
Party founder −0.302 −0.206 0.043 0.230**
(0.212) (0.128) (0.209) (0.109)
Party institutional. 1.347*** −0.570** 0.394 −0.174
(0.475) (0.287) (0.469) (0.245)
Constant −1.184** 0.426 0.663 0.269
(0.515) (0.308) (0.498) (0.262)
Observations 214 214 214 214
Rho −0.893 −0.781 −0.518 0.702
Inverse Mills Ratio −0.491 (0.500) −0.246 (0.302) −0.247 (0.497) 0.184 (0.259)
Notes: *p < 0.1; **p < 0.05; ***p < 0.01.
14 A. FRUHSTORFER AND A. HUDSON

whether or not a previous president (within the last 15 years) has been imprisoned,
exiled, or killed. The measure of corruption also provides information about presiden-
tial expectations regarding the desirability of remaining in office. The logistic
regressions reported in Table 3 suggest that corruption is both predictive of opportu-
nities to retain power (and possibly wealth) outside of formal office, and also a poten-
tial proxy measure for international prestige (as seen through its effects on the
probability of appointment to international organizations). While this is not a
perfect measure, it provides valuable information about the political context that
would inform presidential expectations. However, we do not find that these measures
are significant predictors of attempts.
The theory we outlined above suggested that opposition party autonomy and cor-
ruption would be strong predictors of post-term outcomes, all else equal. But the
coefficients estimated in both stages confirm only parts of our theoretical assumptions.
In the first stage, none of our theoretically expected coefficients reaches significance.
But most of our control variables (horizontal accountability, prior attempt, and the
president as party founder) are significant and have an effect in the expected direction.
At the same time, following up on the results from the logistic regressions above (in
Table 3), in the second stage we again find that corruption is positively associated
with the ability of a president to retain power (in this case either in office or outside
office, Table 5).
However, opposition party autonomy has no impact on any of the outcomes under
consideration. Based on our analysis, opposition party autonomy – which ranks how
independent and autonomous the opposition is from the ruling regime – does not have
the expected effect. While presidents often instrumentalize a co-opted parliament or
co-opted opposition groups to stabilize their rule, this instrumentalization does not
extend to term limit evasion and does not influence the post-term opportunities
available.
Political parties do seem to matter in another way. Presidents in contexts with more
highly institutionalized political parties are more likely to retain power (either by
remaining in formal office, or through some informal arrangement). At the same
time, party institutionalization has a negative effect on retaining prestigious positions
after presidents end their term. We assume that this is a combination of the measure
that captures characteristics of parties and not characteristics of party systems and the
nature of our dataset. Countries in our dataset were part of the third wave of democra-
tization. Since a high level of party institutionalization (i.e. existence of national and
local organizations, platforms, high levels of cohesion between parties and elites, as
well as programmatic linkages with citizens) takes time to develop, the at least
weakly democratic cases in our sample may not have had enough experience with
democracy to achieve a high level on this measure. We expect that the dynamic is
different in hybrid and authoritarian regimes, where party institutionalization is
mainly the result of a one-party dominance.
However, as we saw in the logistic regressions, presidents who are founders of the
ruling party are more likely to be imprisoned if they cannot stay in office. Taken
together, it seems that stronger parties increase the likelihood that a president can
retain power, but losing power is potentially more costly for leaders who have tra-
ditionally held a stronger position within their party. Similarly, we see a strong and
positive effect of party institutionalization on retaining power. The way a party
system is organized, whether it is professionalized and institutionalized has a strong
 DEMOCRATIZATION 15

impact on a post-term activity of the leaving president. While every point increase in
the institutionalization of the party system increases the odds for retaining power by
65%, it decreases the odds for retaining a prestigious position by 43%.

Conclusion
Term-limited presidents face a difficult decision regarding whether or not they will
attempt to extend their term in office. The costs and benefits attending this choice
are not equally obvious. We have argued that this calculation will be different for
each individual, as they decide whether to pursue power, status, or some more altruis-
tic goal. As we have seen, the choice to even attempt to extend their term closes other
doors (and may even increase the likelihood of ending up in prison). At the same time,
various elements of the political context serve to limit the choices available.
In our analysis, we show that leaders of some countries have opportunities to move
upwards and take on a role in an international organization. However, if they make an
unsuccessful attempt to extend their term, those opportunities are likely to be closed to
them. Our analysis also suggests that the costs and benefits of remaining in office are
not equally distributed. Some leaders (particularly party founders) may face criminal
liability if they lose power – either due to actions they took in an official capacity, or
as payback from their political opponents. On the other side, not all presidents have
the same opportunities to make money (either legitimately or corruptly) once they
have left office.
When we combined the two stages of our analysis in a Heckman sample selection
model, we found that there are few reasonably consistent predictors of post-term
activities for presidents. However, both party institutionalization and corruption
emerged as a positive predictor of the ability of a president to retain power. While
opposition party autonomy was not a significant predictor of post-term outcomes,
the centrality of political parties and whether elite loyalty is tied to a person or a
party is clear and suggests a particular point of focus in future work.66
These findings have important implications for the tension between presidents’ and
leaders’ dual goals of winning/remaining in office and at the same time protecting their
post-term interests. Given our findings that presidents’ expectations about post-term
perspectives influence their attempts to evade term limits, a president that primarily
cares about wealth or influence might consider the diminishing choices in the long
run of an evasion attempt. Thus, future research may include a more detailed analysis
of anti-corruption rules and measures as well as more practically the de-facto pension
entitlement. Second, this finding highlights the ways in which institutions and political
contexts more broadly shape the degree to which term-limited presidents are able to
maintain power or influence. We also note that our null findings in some areas
suggest potential pathways for future research. The null results in areas where we
have strong theoretical reasons to believe a relationship exists also suggest the necessity
of qualitative research, particularly as the data are sparse around some outcomes of
considerable normative importance.

Notes
1. Madison, Debates in the Federal Convention, 325.
2. Wa Afrika and Hofstatter, “Inside Zuma’s Duabi Palace.”
16 A. FRUHSTORFER AND A. HUDSON

3. Mayhew, Congress; Fiorina, Congress, Keystone of the.

4. Dawson, The Federalist, 55–64.
5. Anderson, “The Ex-Presidents.”
6. McKie, “Presidential Term Limit Contravention.”
7. Baturo, Democracy, Dictatorship, and Term Limits; Baturo, “Democracy, Development, and
Career.”
8. Graber, Levinson, and Tushnet, Constitutional Democracy in Crisis.
9. Ginsburg, Elkins, and Melton, “Do Executive Term Limits.”
10. Weyland, “Limitations of Rational-Choice”; McKie, “Presidential Term Limit Contravention”;
Corrales and Penfold, “Manipulating Term Limits in.”
11. North, Institutionen, Institutioneller Wandel Und; Elster, “Deliberation and Constitution
Making.”
12. Corrales, “Can Anyone Stop the”; Dulani, Personal Rule and Presidential Term Limits in Africa.
13. Corrales, Fixing Democracy; McKie, “Presidential Term Limit Contravention.”
14. Hay and Wincott, “Structure, Agency and Historical Institutionalism,” 954–5.
15. Mahoney and Snyder, “Rethinking Agency and Structure in the Study of Regime Change,” 13.
16. Giddens, Central Problems in Social Theory.
17. Although we have some idea about the post-term fate of democratic leaders, see in particular:
Baturo, Democracy Dictatorship, and Term Limits.
18. Baturo, “Democracy, Development, and Career”; Theakston and De Vries, Former Leaders in
Modern.
19. Thelen and Steinmo, “Historical Institutionalism in Comparative”; Pierson, “Increasing
Returns, Path Dependence, and the Study of Politics”; Hall and Taylor, “Political Science
and the.”
20. McKie, “Presidential Term Limit Contravention”; Baturo, Democracy, Dictatorship, and Term
Limits; Silitski, “Tools of Autocracy”; Schedler, “Electoral Authoritarianism: The Dynamics”;
Gandhi, Political Institutions Under Dictatorship; Gandhi and Przeworski, “Authoritarian
Institutions and the.”.
21. Maslow, “A Theory of Human,” 394.
22. Wright and Escribà-Folch, “Authoritarian Institutions and Regime.”
23. Baturo, “Democracy, Development, and Career.”
24. Ibid., 1048.
25. Escribà-Folch and Krcmaric, “Dictators in Exile: Explaining.”
26. Baturo, “Democracy, Development, and Career.”
27. Chang and Chu, “Corruption and Trust: Exceptionalism”; Davis, Camp, and Coleman, “The
Influence of Party”; Meier and Holbrook, “‘I Seen My Opportunities’.”
28. Baturo, “The Stakes of Losing,” 640.
29. Baturo, Democracy, Dictatorship, and Term Limits, 6.
30. Escribà-Folch and Krcmaric, “Dictators in Exile: Explaining.”.
31. Baturo, “Democracy, Development, and Career.”
32. Shepsle, “The Strategy of Ambiguity.”
33. Frye, “A Politics of Institutional”; Jung and Deering, “Constitutional Choices: Uncertainty
and”; Luong, Power, Perceptions, and Pacts.
34. McKie, “Presidential Term Limit Contravention,” 1525.
35. Levitsky and Way, Competitive Authoritarianism.
36. Gandhi, Political Institutions Under Dictatorship.
37. Gandhi and Przeworski, “Authoritarian Institutions and the,” 1279.
38. McKie, “Presidential Term Limit Contravention,” 1512–13.
39. Gerschewski, “The Three Pillars of,” 22.
40. Shleifer and Treisman, Without a Map, 8.
41. Gerschewski, “The Three Pillars of,” 22.
42. Ibid.; Dixon and Landau, “Constitutional End Games”; Corrales and Penfold, “Manipulating
Term Limits in”; Meng, When Do Autocrats Share; Baturo, “Democracy, Development, and
Career.”
43. Versteeg et al., “The Law and Politics,” 176.
44. McKie, “Presidential Term Limit Contravention,” 1501.
 DEMOCRATIZATION 17

45. McKie, “Presidential Term Limit Contravention”; Goemans, Gleditsch, and Chiozza, “Introdu-
cing Archigos”; Ginsburg, Melton, and Elkins, “On the Evasion of”; Baturo, Democracy, Dic-
tatorship, and Term Limits; Baturo, “Democracy, Development, and Career”; Baturo, “The
Stakes of Losing”; Versteeg et al., “The Law and Politics.”.
46. Elgie, “The Politics of Semi-Presidentialism.”.
47. Ibid.
48. McKie, “Presidential Term Limit Contravention.”
49. Baturo, Democracy, Dictatorship, and Term Limits.
50. Gandhi and Przeworski, “Authoritarian Institutions and the,” 1279.
51. McKie, “Presidential Term Limit Contravention,” 1513.
52. Gerschewski, “The Three Pillars of”; Przeworski, “Institutions Matter?”
53. Coppedge et al., “V-Dem Codebook V10,” 91.
54. McKie, “Presidential Term Limit Contravention”; The World Bank, World Development
Indicators.
55. Baturo, Democracy, Dictatorship, and Term Limits; McKie, “Presidential Term Limit
Contravention.”.
56. Coppedge et al., “V-Dem Codebook V10,” 293.
57. Ibid., 42.
58. Ibid., 270.
59. Mckie, “Presidential Term Limit Contravention,” 1515.
60. Firth, “Bias Reduction of Maximum Likelihood Estimates.”
61. McKie, “Presidential Term Limit Contravention,” 1516.
62. Meng, “Ruling Parties in Authoritarian,” 534.
63. Guo and Fraser, Propensity Score Analysis, 88.
64. Heckman, “Sample Selection Bias as a Specification Error.”
65. Toomet and Henningsen, “Sample Selection Models in R.”
66. Osei et al., “Presidential Term Limits and Regime Types,” 271.

Acknowledgements
We would first like to thank Charlotte Heyl and Mariana Llanos for their invitation to contribute to
this special issue. Thanks also to the participants at the workshop that they organized at the German
Institute of Global and Area Studies (online in Autumn 2020). We are also grateful to the editors of the
journal and the anonymous reviewers for their insights and generous engagement with our research.

Disclosure statement
No potential conflict of interest was reported by the author(s).

ORCID
Anna Fruhstorfer http://orcid.org/0000-0002-1193-2706
Alexander Hudson http://orcid.org/0000-0002-2301-3918

Bibliography
Anderson, Lisa. “The Ex-Presidents.” Journal of Democracy 21, no. 2 (2010): 64–78. doi:10.1353/jod.0.
0166.
Associated Press (2020). “Mexico Supreme Court OKs Vote on Prosecuting Ex-Presidents.” Associated
Press. https://apnews.com/article/virus-outbreak-mexico-archive-courts-referendums-
6bcaa317dc2690cfb229e37b9003ff1e.
Baturo, Alexander. “Democracy, Development, and Career Trajectories of Former Political Leaders.”
Comparative Political Studies 50, no. 8 (2017): 1023–1054. doi:10.1177/0010414016666835.
Baturo, Alexander. Democracy, Dictatorship, and Term Limits. Ann Arbor: University of Michigan
Press, 2014.
18 A. FRUHSTORFER AND A. HUDSON

Baturo, Alexander. “The Stakes of Losing Office, Term Limits and Democracy.” British Journal of
Political Science 40, no. 3 (2010): 635–662.
Bizzarro, Fernando, John Gerring, Carl Henrik Knutsen, Allen Hicken, Michael Bernhard, Svend-Erik
Skaaning, Michael Coppedge, and Staffan I. Lindberg. “Party Strength and Economic Growth.”
World Politics 70, no. 2 (2018): 275–320. doi:10.1017/S0043887117000375.
Chang, Eric C. C., and Yun-han Chu. “Corruption and Trust: Exceptionalism in Asian Democracies?”
Journal of Politics 68, no. 2: 259–271. https://onlinelibrary.wiley.com/doi/10.1111/j.1468-2508.
2006.00404.x/full.
Coppedge, Michael, John Gerring, Carl Henrik Knutsen, Staffan I. Lindberg, Svend-Erik Skaaning, Jan
Teorell, David Altman, et al. “V-Dem Codebook V10.” Varieties of Democracy (V-Dem) Project,
doi:10.23696.
Corrales, Javier. “Can Anyone Stop the President? Power Asymmetries and Term Limits in Latin
America, 1984–2016.” Latin American Politics and Society 58, no. 2 (2016): 3–25.
Corrales, Javier. Fixing Democracy: Why Constitutional Change Often Fails to Enhance Democracy in
Latin America. New York: Oxford University Press, 2018.
Corrales, Javier, and Michael Penfold. “Manipulating Term Limits in Latin America.” Journal of
Democracy 25, no. 4 (2014): 157–168.
Davis, Charles L., Ro Ai Camp, and Kenneth M. Coleman. “The Influence of Party Systems on
Citizens’ Perceptions of Corruption and Electoral Response in Latin America.” Comparative
Political Studies 37, no. 6 (2004): 677–703.
Dawson, Henry Barton, ed. The Federalist: A Collection of Essays Written in Favor of the New
Constitution as Agreed Upon by the Fedral Convention, September 17, 1787. New York: Charles
Scribner’s sons, 1864.
Dixon, Rosalind, and David Landau. “Constitutional End Games: Making Presidential Term Limits
Stick.” Hastings LJ 71 (2019): 359.
Dulani, Boniface Madalitso. Personal Rule and Presidential Term Limits in Africa. East Lansing:
Michigan State University. Political Science, 2011.
Elgie, Robert. “The Politics of Semi-Presidentialism.” In Semi-Presidentialism in Europe, edited by
Robert Elgie, 1–21. Comparative European Politics. Oxford [u.a.]: Oxford Univ. Press, 1999.
Elster, Jon. “Deliberation and Constitution Making.” In Deliberative Democracy, edited by Jon Elster,
97–111. Cambridge: Cambridge University Press, 1998.
Escribà-Folch, Abel, and Daniel Krcmaric. “Dictators in Exile: Explaining the Destinations of Ex-
Rulers.” The Journal of Politics 79, no. 2 (2017): 560–575.
Fiorina, Morris P. Congress, Keystone of the Washington Establishment. New Haven: Yale University
Press, 1989.
Firth, David. “Bias Reduction of Maximum Likelihood Estimates.” Biometrika 80, no. 1 (1993): 27–38.
doi:10.2307/2336755.
Frye, Timothy. “A Politics of Institutional Choice Post-Communist Presidencies.” Comparative
Political Studies 30, no. 5 (1997): 523–552.
Gandhi, Jennifer. Political Institutions under Dictatorship. New York: Cambridge University Press
Cambridge, 2008.
Gandhi, Jennifer, and Adam Przeworski. “Authoritarian Institutions and the Survival of Autocrats.”
Comparative Political Studies 40, no. 11 (2007): 1279–1301.
Gerschewski, Johannes. “The Three Pillars of Stability: Legitimation, Repression, and Co-Optation in
Autocratic Regimes.” Democratization 20, no. 1 (2013): 13–38.
Giddens, Anthony. Central Problems in Social Theory: Action, Structure, and Contradiction in Social
Analysis. Vol. 241. Los Angeles: Univ of California Press, 1979.
Ginsburg, Tom, Zachary Elkins, and James Melton. “Do Executive Term Limits Cause Constitutional
Crisis?” In Comparative Constitutional Design, edited by Tom Ginsburg, 350–379. Cambridge,
New York: Cambridge University Press, 2012.
Ginsburg, Tom, James Melton, and Zachary Elkins. “On the Evasion of Executive Term Limits.” Wm.
& Mary L. Rev. 52 (2010): 1807.
Goemans, Henk E., Kristian Skrede Gleditsch, and Giacomo Chiozza. “Introducing Archigos: A
Dataset of Political Leaders.” Journal of Peace Research 46, no. 2 (2009): 269–283. doi:10.1177/
0022343308100719.
Graber, Mark A., Sanford Levinson, and Mark Tushnet. Constitutional Democracy in Crisis? Oxford:
Oxford University Press, 2018.
 DEMOCRATIZATION 19

Guo, Shenyang, and Mark W. Fraser. Propensity Score Analysis: Statistical Methods and Applications.
London: SAGE, 2015.
Hall, Peter A., and Rosemary C. R. Taylor. “Political Science and the Three New Institutionalisms.”
Political Studies 44, no. 5 (1996): 936–957.
Hay, Colin, and Daniel Wincott. “Structure, Agency and Historical Institutionalism.” Political Studies
46, no. 5 (1998): 951–957.
Heckman, James J. “Sample Selection Bias as a Specification Error.” Econometrica 47, no. 1 (1979):
153–161. doi:10.2307/1912352.
Jung, Jai Kwan, and Christopher J. Deering. “Constitutional Choices: Uncertainty and Institutional
Design in Democratising Nations.” International Political Science Review 36, no. 1 (2015): 60–77.
Kurmanev, Anatoly. “Evo Morales and Bolivia: What We Know about the President’s Resignation.”
New York Times. 12 November 2019. https://www.nytimes.com/2019/11/12/world/americas/evo-
morales-resignation-bolivia-facts.html.
Landau, David, and Rosalind Dixon. “Abusive Judicial Review: Courts Against Democracy.”
University of California, Davis Law Review 53 (2019): 1313–1387.
Levitsky, Steven, and Lucan A. Way. Competitive Authoritarianism: Hybrid Regimes After the Cold
War. Cambridge: Cambridge University Press, 2010.
Luong, Pauline Jones. Institutional Change and Political Continuity in Post-Soviet Central Asia: Power,
Perceptions, and Pacts. Cambridge: Cambridge University Press, 2002.
Madison, James. The Debates in the Federal Convention of 1787 Which Framed the Constitution of the
United States of America. Edited by Gaillard Hunt and James Brown Scott. Clark, NJ: The Lawbook
Exchange, Ltd., 1999.
Mahoney, James, and Richard Snyder. “Rethinking Agency and Structure in the Study of Regime
Change.” Studies in Comparative International Development 34, no. 2 (1999): 3.
Maslow, A. H. “A Theory of Human Motivation.” Psychological Review 50 (1943): 370–396.
Mayhew, David. Congress: The Electoral Connection. New Haven: Yale University Press, 1974.
McKie, Kristin. “Presidential Term Limit Contravention: Abolish, Extend, Fail, or Respect?”
Comparative Political Studies 52, no. 10 (2019): 1500–1534.
Meier, Kenneth J, and Thomas M Holbrook. “‘I Seen My Opportunities and I Took’Em:’ Political
Corruption in the American States.” The Journal of Politics 54, no. 1 (1992): 135–155.
Meng, Anne. When Do Autocrats Share Power? A Theory of Party Institutionalization and Leader
Strength: Working Paper, 2017. https://www.annemeng.com/uploads/5/6/6/6/56666335/meng_
share_power.pdf.
Meng, Anne. “Ruling Parties in Authoritarian Regimes: Rethinking Institutional Strength.” British
Journal of Political Science 51, no. 2 (2021): 526–540.
North, Douglass C. Institutionen, Institutioneller Wandel Und Wirtschaftsleistung. Vol. 76. Die Einheit
Der Gesellschaftswissenschaften. Tübingen: Mohr, 1992.
Osei, Anja, Hervé Akinocho, and Stephen Mwombela. “Presidential Term Limits and Regime Types:
When Do Leaders Respect Constitutional Norms?” Africa Spectrum 55, no. 3 (2020): 251–271.
Pierson, Paul. “Increasing Returns, Path Dependence, and the Study of Politics.” The American
Political Science Review 94, no. 2 (2000): 251–267. doi:10.2307/2586011.
Przeworski, Adam. “Institutions Matter?” Government and Opposition 39, no. 4 (2004): 527–540.
Schedler, Andreas. “Electoral Authoritarianism: The Dynamics of Unfree Competition.”
Shepsle, Kenneth A. “The Strategy of Ambiguity: Uncertainty and Electoral Competition*.” American
Political Science Review 66, no. 2 (1972): 555–568. doi:10.2307/1957799.
Shleifer, Andrei, and Daniel Treisman. Without a Map: Political Tactics and Economic Reform in
Russia. Cambridge: MIT Press, 2001.
Silitski, Vitali. “Tools of Autocracy.” Journal of Democracy 20, no. 2 (2009): 42–46. doi:10.1353/jod.0.
0067.
The World Bank. World Development Indicators, 2020. http://data.worldbank.org/indicator/NY.GNP.
PCAP.PP.CD.
Theakston, Kevin, and Jouke de De Vries. Former Leaders in Modern Democracies. Basingstoke:
Springer, 2012.
Thelen, Kathleen, and Sven Steinmo. “Historical Institutionalism in Comparative Politics.” In
Structuring Politics: Historical Institutionalism in Comparative Analysis, edited by Sven Steinmo,
Kathleen Thelen, and Frank Longstreth, 1–32. Cambridge: Cambridge University Press, 1992.
20 A. FRUHSTORFER AND A. HUDSON

Toomet, Ott, and Arne Henningsen. “Sample Selection Models in R: Package SampleSelection.”
Journal of Statistical Software 27, no. 1 (2008): 1–23. doi:10.18637/jss.v027.i07.
Versteeg, Mila, Timothy Horley, Anne Meng, Mauricio Guim, and Marilyn Guirguis. “The Law and
Politics of Presidential Term Limit Evasion.” Columbia Law Review 120 (2020): 173–248.
Wa Afrika, Mzilikazi, and Stephan Hofstatter. 2017. “Inside Zuma’s Dubai Palace and Mugabe Lives
Next Door.” Sunday Times. https://www.timeslive.co.za/sunday-times/news/2017-06-04-inside-
zumas-dubai-palace–and-mugabe-lives-next-door/.
Weyland, Kurt. “Limitations of Rational-Choice Institutionalism for the Study of Latin American
Politics.” Studies in Comparative International Development 37, no. 1 (2002): 57–85.
Wright, Joseph, and Abel Escribà-Folch. “Authoritarian Institutions and Regime Survival:
Transitions to Democracy and Subsequent Autocracy.” British Journal of Political Science 42, no.
2 (2012): 283–309.