Oxford Open Economics, 2023, 00, 1–6

https://doi.org/10.1093/ooec/odad002
Research Article

The morphing of dictators: why dictators get worse over

time
Kaushik Basu*

Downloaded from https://academic.oup.com/ooec/article/doi/10.1093/ooec/odad002/7036634 by guest on 30 May 2024

Department of Economics and SC Johnson College of Business, Cornell University, 422 Uris Hall Tower Road, Ithaca, New York 14853, USA
*Correspondence address. E-mail: [email protected]

Abstract
Dictators, even those who seize power with the intention of helping the nation, frequently morph over time into tyrants. There may
be many reasons for this. This paper focuses on one interesting and arguably pervasive driver behind this process. A model is developed
which shows that the series of decisions taken over time by an authoritarian leader concerning how much political intrigue and evil to
indulge in in order to stay in power leads to a dynamic inconsistency converting the leader into a tyrant. It is possible that the dictator
will, eventually, come to regret this, but by then they have no exit options. The analysis prompts us to think about ex ante rules and
term-limit provisions to prevent this from happening.

Keywords: dynamic inconsistency, procrastination, term limit, tyrant, dictator

INTRODUCTION For strong political leaders who stay in power for a long time,
there are many changes that occur in their environment that
Through history, we have seen examples of dictators who came to
can be cause for concern. We know that the information and
power with some genuine intention to lead the nation to a better
news that such leaders get become increasingly biased. Their staff
future and morphed, over time, into a ruthless tyrant, oppressing
and minions, concerned not to upset the leader, give their boss
the citizenry and willing to do anything to remain in power.
the information the boss would like to hear. It is likely that Mao
Usually, when a leader reaches this ruthless phase it becomes
Zedong did not learn for a long time that his Great Leap Forward
politically incorrect to say that they may have once been reason-
had failed and was causing one of the great famines in world
able, or at least more reasonable, with some genuine intention
history, instead. It is likely that what Putin believes is happening in
to provide good leadership to their country. I am not suggesting
Ukraine is far from what is happening in Ukraine. His reputation
that the initial good intention is always the case. The main result
for tyranny and brutality must ensure that his staff will give him
in this paper is to show that no matter where an authoritarian
the news he wants to hear. Further the psychology of tyrants
leader starts from there will be a tendency to deteriorate over
is often very different from ordinary mortals. Some of them are
time. However, I believe that to deny that there may be good inten-
probably unaware of their own brutality or under the belief that
tions at the start, as though evil intention is an axiomatic truth,
it is done in the best interest of the nation. It is impossible to know
hurts an objective analysis of dictators and totalitarian regimes
if they actually believe in this or create these delusions to be able
and thereby thwarts the development of policies and strategy to
to live with themselves.
block such unwanted turns of events as we have seen repeatedly
The aim of this paper is to focus on a slender cut of this reality
through history. Examples jump out at us: Joseph Stalin, Benito
and draw attention to a natural result. No matter what the initial
Mussolini, Hugo Chavez, Fidel Castro, Kim Il-sung, Robert Mugabe,
intention of a leader seeking power, it is easy to take steps to
Zine El Abidine Ben Ali, Muammar Gaddafi, Daniel Ortega. In case
prolong their tenure, which one step at a time may seem natural,
these examples seem inadequate, after I began working on this
but in their totality traps the leader in political evil and oppression
paper, Putin invaded Ukraine, giving us a reminder that he had
that they cannot escape. The most poignant statement of this
been left out. So let me add: and Vladimir Putin.
tragic human predicament occurs in Shakespeare’s Macbeth (Act
The paper uses simple, plausible assumptions and provides, by
III, Scene IV), when Macbeth tells his wife:
a step-by-step analysis, a surprisingly clear explanation of this
propensity of leaders in power for long to morph into someone “I am in blood
tyrannical and evil. What the paper tries to explain is the puzzle Stepp’d in so far that, should I wade no more,
summed up by Stephen Kinzer (2021): ‘For most of his life [Daniel Returning were as tedious as go o’er.”
Ortega] was in passionate rebellion against [Somoza’s dictator- If the leaders are innately evil, their ending up practicing evil,
ship and] everything it represented. Then, in what seemed an would not be a surprising result. By focusing on the case of leaders
astonishing about-face, he began replicating it. With precision and with good intention and then showing how they end up as tyrants,
design, he has created an insular, dynastic tyranny that eerily this paper explains why this transformatio of leaders who remain
resembles the one against which he fought decades ago.’ long in power is so ubiquitous. It should be emphasized that the

Received: August 17, 2021. Revised: December 24, 2021. Accepted: January 24, 2022
© The Author(s) 2023. Published by Oxford University Press.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which
permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
2 | Oxford Open Economics, 2023, Vol. 00, No. 00

model is not about where the leaders start—good or bad, but on in office, and that will make the exit option even worse, and
about the process by which they deteriorate, wherever they start. the inexorable logic of this continues. Gaddafi must have realized,
midway through, that voluntary exit was no longer an option for
him. If he had done the full dynamic optimization at the start he
DICTATOR’S CHOICE may never have gone down this route.
Though the paper is an exercise in theory, it owes its origin to a What makes the ongoing Ukraine war so dangerous is that
personal experience. In September 2013 I met Daniel Ortega in Putin is fast losing all exit options. We have enough evidence and
Managua. The meeting had nothing to do with the World Bank analysis from behavioral economics to know that from procrasti-
mission that had taken me to Nicaragua. As a student in India nators to drug addicts many fall into this trap. What the present
and, later, England, I had heard and admired the achievement of paper argues is so may dictators and the final outcome can be bad
Ortega in overthrowing Somoza’s corrupt regime. It was a long for all.
meeting where we talked about the Sandinista revolution and the Conversely, leaders who have got out of power have on occa-

Downloaded from https://academic.oup.com/ooec/article/doi/10.1093/ooec/odad002/7036634 by guest on 30 May 2024

challenges faced by Nicaragua. Given the sordid accounts that sions paid a heavy price for it. Cooter & Schäfer (2012) refer to
have subsequently emerged of Ortega’s oppressive and tyrannical this as the ‘dictator’s dilemma.’ A good illustration of this comes
behavior (Kinzer, 2021), I regret the decision to meet him. I was also from Chile’s General Augusto Pinochet. He came to power after
left with a troubling question: How could this happen? How could a brutal military coup in 1973, with non-negligible help from the
Daniel Ortega, who struggled, suffered and eventually overthrew CIA, against the democratically-elected leader Salvador Allende.
the corrupt tyrant, Anastasio Somoza Debayle, morph into a He quickly morphed into a dictator, and ruled the country with an
creature similar to the one he overthrew? Was it a case of one iron hand over more than a decade, killing thousands of opponent
freak individual and his psychology, or is there something generic and torturing hundreds and thousands of his opponents, mainly
about authoritarianism that makes it deteriorate over time. The leftists and socialists, during his term as leader of the nation. In
model that follows is an exercise in atonement through model 1990 he got out of power by offering his resignation. However,
building. it was not a good life after exiting from power. He was soon
The paper presents a theoretical structure inspired by the charged with human rights violation, murder and torture during
above reality. What the paper shows is that a modicum of dynamic his term as president and was arrested in 2004. Citing the case
inconsistency or present bias, which is a widely noted trait in of Augusto Pinochet, Cooter and Schäfer (2012, p. 219) point out,
human beings (Phelps and Pollak, 1968; O’Donoghue and Rabin, ‘This story depicts a dilemma: an aging dictator wants to resign
1999; Kleinberg and Oren, 2014; Chakraborty, 2021), gives us this from power, but he fears prosecution for crimes. His only effective
result. As one ponders about this, it becomes clear that this is guarantee against prosecution is to retain the power that he wants
not just a theoretical possibility but, being based on realistic to relinquish.’
assumptions, it gives us insight into this troubling phenomenon,
of strong leaders morphing into tyrants. Often big problems have
simple root causes. What this paper shows is that the reason
DICTATOR MORPHING: A MODEL
dictators morph into tyrants is, in terms of logic, the same as To capture this dilemma simply, let us assume that the political
why George Akerlof procrastinated mailing Joseph Stiglitz’s parcel calendar is broken up into terms, say 4 years for each term. Once
from India, after Stiglitz returned home and Akerlof stayed on for in office, you are there for one term. During that term you do
a while more (Akerlof, 1991). Every morning, the transactions cost many things as leader but you also have to decide whether to
of going to the post-office and standing in the queue that day fight to stay on for another term. Consider the problem in the
acquired a salience in his mind, which made him decide he would first term. It is a new leader and he has to decide his strategy.
mail it the next day. It was this series of present-biased decisions Suppose the amount of evil you indulge in is an element of
that led to the long procrastination. [0, 1], where 0 means no evil and 1 is maximum possible evil.
The gist of the argument in the context of political leaders is the If you find it difficult to imagine what 1 means, look around
following. Consider a person who comes to power by overthrowing the world. You will not have to imagine; you will find empirical
a bad leader, as Daniel Ortega did by overthrowing Anastasio examples.
Somoza Debayle. The new leader, even while helping the country The leader has to choose e ∈ [0, 1], where e represents the
to do better, will soon face the decision of whether to seek more extent of evil they indulge in in order to stay in power. Suppose
time in office. Politics is a tough game, where you often have to for simplicity that the probability of staying in power for another
indulge in evil acts in order to escape the fate of being evicted term is equal to the chosen e. The more evil you do the more likely
from power. So the person has to decide how far to go in terms you are to stay in power. Also assume, once one exits office, one
of political intrigue and ruthless acts in order to stay in power. A does not return again. This simplifies the analysis.
common feature in making this optimization decision is dynamic Let the joy or payoff of being in power for one term be denoted
inconsistency. by x and the payoff of being out of power be y. Assume x > y.2 Few
If the leader indulges in a certain amount of evil in order to like to do evil. Let the guilt and remorse of doing evil, e, be given
remain in power and does thereby manage to stay in power, then
during the next term or further in the future there may be an
additional reason to want to stay in power. The evil acts of the
earlier terms in office could mean that exit from office will be
more costly. You will face inquiries about your acts while in office help create such a make belief world. For simplicity of analysis, I shall stay away
and may be punished1 . This will prompt greater evil acts to stay from these complexities.
2 The constituents of the joy of being in office are not modeled here. One
joy could come from being able to do what you believe is good for the nation.
1 It is also possible that as people indulge in such acts they internalize these The reason why you may be able to do good in office is that being in office is
as not evil acts but hard decisions one has to make for the good of the nation. a way to get legitimacy over others. You can simply give orders to incentivize
Such delusions make it easier to live with oneself. Their persistence is also others (Akerlof, 2017a). This is what leaders do and it is not hard to model this
made more likely by the advisers who surround the leader and misinform and (Basu, 2022b).
 RRH: Basu et al. | 3

Now think of the problem more generally as consisting of T-

terms, where T > 1. The leader has to take a decision in T-1 terms,
that is, in all but the last term. The dynamically inconsistent
person, which is the way most human beings are, treats each
decision, without detailed calculation of the full future. If the
leader thought of the full life’s decisions as one big optimization
problem, the calculation would be different. At each period, she
would have to recognize the next period’s payoff from exiting
from office is a decreasing function of the level of e chosen in
this period. If in period t, the leader chooses evil et and the payoff
from leaving power in the next period is yt + 1 , what I am asserting
Figure 1. Dictator’s choice. is that4 :

Downloaded from https://academic.oup.com/ooec/article/doi/10.1093/ooec/odad002/7036634 by guest on 30 May 2024

yt+1 = f (et ) , f  (et ) < 0.
3
by r(e), where r’(e) > 0, for all e. Finally, I assume r”(e) > 0 and
If we denote the net payoff in each period, t, by πt , which is
r(1) > x and r (0) < x–y. (1) defined the same way as (1), above, then the full optimization
Condition 1 is made simply to ensure that there is an interior exercise consists of choosing (e1 , e2 , . . . eT−1 ) to maximize
solution. I will work with y > 0 but this is not necessary.
π1 + π2 + · · · + πT−1 .
It seems likely the politician, trying to decide whether to stay
on in office for another term, will do the following optimization. The main theorem is the following: There exist parameters,
He will try to choose e to maximize: under which, if this full optimization is done, the political leader
π ≡ xe + y (1 − e) − r(e). (2) will choose et = 0, for all t. However, if the choice is made in each
period, unmindful of the fact that this term’s choice of e, will have
Hence, e is derived from the first-order condition of maximizing an effect on next term’s value of y, or treating this term and the
π: next term as the only two terms of life, then the politician will
 
x − y = r (e). choose (e1 , e2 , . . . eT−1 ) such that e1 < e2 < · · · < eT−1 . That is,
the extent of evil actions chosen will increase monotonically with
Figure 1 illustrates this maximization exercise, with the equi- each term in office.
librium choice of evil, e∗. The process of morphing into a tyrant is inexorable. The conse-
The line yP depicts xe + y(1 – e); and the r(e) graph is also quence of the cost of exit getting higher in each term is that you
shown in the figure. The politician wants to maximize the gap will have to indulge in even greater political intrigue, ruthlessness
between the two, described by (2). This happens at e∗. Note that (1) and evil to make sure you do not get thrown out of office. The
implies that e∗ > 0. Hence, this level of evil causes remorse given following period you have to do even more. Ultimately, you morph
by r(e∗), which is greater than 0 and is shown in the picture by Be∗. into late stage Alexander Lukashenko, Vladimir Putin or Daniel
However, it has a net benefit of AB. AB is the largest gap between Ortega.
the line yP and curve r(e). Hence, AB is greater than y, shown in If they knew how to do dynamic programming right they may
the figure by the line segment y0. This means that indulging in not have got into the process at the start, done one good term in
no evil, that is e = 0, lowers the cost of remorse to 0, but you lose office and left. But this kind of short-sightedness, whether we call
power for sure after one term and that is not attractive. it hyperbolic discounting or just unawareness of the full impact
There is however a ‘mistake’ the leader makes here, which is of your action into the future, is a common human mistake. The
common and natural. That is the mistake of not being far-sighted problem is that this has devastating consequences for the well-
enough. Typically, if you indulge in evil and stay in power, your cost being of society and the world, as we can see from the Ukraine
of exit in the next period will rise. That is, in the next period, the war.
payoff from exiting from power (that is, the y of the next period) One can give a formal proof of the theorem for the T-period
will be lower. There will be inquiries about why you, as president game by solving the full optimization. But the intuition is easy
or prime minister, indulged in unlawful, evil acts? There may be enough that the algebra is redundant. I give a proof here for the
cases brought against you; and you may even be incarcerated. case where T = 3; so that there are two decision points. The entire
This has happened to leaders in history. The fear of this must proof is obvious using one diagram, Figure 2.
stalk Bolsonaro, Erdogan, Lukashenko, Orban, Ortega and Putin— A part of Fig. 2 is a reproduction of Fig. 1, where the leader chose
presented, I hasten to add, in alphabetical order. e∗ in period 1, when the exit option was y, where y = f(0), since at
To understand this, note that the above analysis was done the start there is no history of evil. Denote f(e∗) by y’. We know
assuming that life (in politics) consists of 2 (potential) terms, that y’ < y, since f’ < 0. It is easy to see that as y falls, the line yP
which means there is only one decision point which is the first becomes steeper by pivoting around point P. This is because, as
term. This is because there is no decision to be taken in the final one can see from (1), at e = 1, π is independent of y. Hence, as y
term, since in this model the only decision concerns how much falls to y’, the earlier line, yP, pivots at P and become y’P. If we now
effort you are willing to spend for one more term in office. In life’s
last term, that decision does not arise. To put it another way, in
the last term, you will mechanically choose e = 0. 4 The assumption that the cost of evil acts upon exit from office happens
in the next period is made for algebraic simplicity. It is arguable if one acts evil
and fails to remain in power one will have to pay for the evil right away. This will
3 It is possible that some of the power-hungry leaders, feel no remorse or yield the same result as in the paper if we make the payoff upon exit to depend
guilt. That will simply reinforce the result. What makes the result obtained in on the accumulation of all past terms in office including the immediate one. In
this paper more interesting is that it survives even if the leader does have guilt- other words, we could assume, yt = f (e1 + e2 + . . . et ) , where f  < 0. We would
feeling. end up with a little more algebra and the same result.
4 | Oxford Open Economics, 2023, Vol. 00, No. 00

Taking cue from this, we can create models where dynamic incon-
sistency is compatible with rational behavior. That is, it is possible
to conceive of contexts where there is a set of decisions to be taken
over time, each decision is taken fully rationally, but at the end
of it all, the individual has reason to regret the choices. This is
possible in contexts where there is an infinite number of decision
points5 . In the present context, if we assume that this is an infinite
decision problem, that is, during a politician’s lifetime, there are
an infinite number of occasions where she can decide what to do
to enhance her political power, we can create a game akin to some
of the art of M. C. Escher, where each step enhances the players
Figure 2. The Stepladder of authoritarianism. utility, but the player ends up with a utility level, below where she

Downloaded from https://academic.oup.com/ooec/article/doi/10.1093/ooec/odad002/7036634 by guest on 30 May 2024

had started. I argued elsewhere (Basu, 1994), this can happen not
just in art, but in games.
Second, the above model showing the morphing of tyrants was
denote the optimal choice of e by e’, it must be that e’ > e∗, since presented using many strong assumptions. The reason is that
r”(e) > 0. my claim is not that the morphing will always happen, which
Now it is entirely possible to think of parameters such that, would require proving it under the most general conditions, but
the leader, if she did the full optimization for the 3-term game, that there is a propensity for this to happen under reasonable
she would prefer to choose e = 0 in both decision points (that is, in situations. However, it is worth pointing out that many of the
terms 1 and 2), rather than choosing e∗ and e’ as he does when he assumptions can be relaxed without losing the result.
is short-sighted and ignores the effect on his exit option when he For instance, the assumption that the function r(e) is increasing
chooses 2. and convex can be generalized if we relax the assumption that the
To see this, note that, if she chooses e = 0 in both the first term probability of winning another term is equal to e, as assumed in
and the second term, the total benefit he gets over time is x + y + y. the above model. Allowing for non-linearity in this will allow us
This is because, in term 1, she is in office and so gets x. Since she to relax some of the restrictions on r(e) and to get the same result.
chooses e = 0, in term1, she incurs no remorse from choosing evil Third, the fact that it was assumed that it is political intrigue
but she has to leave office and so gets y. Once out of office, always and evil, variable e in the model, that helps authoritarian leaders
out of office; so in term 3 she again gets y. to prolong their hold on power, may suggest that it is being
Next suppose he chooses e∗ and e’ in term 1 and term 2. The assumed that doing good does not help retain power. That is not
total payoff she gets over time is x + AB + (1 – e∗) y + e∗CD, where however an assumption being made here. In fact, doing good, or
the AB and CD refer to the line segments in Fig. 2. This is easy to at least doing visible good, will typically help prolong a leader’s
see. In term 1, she is already in office and gets x. By choosing e∗ in term in office. However, that is a tactic that all leaders will use
term 2, she gets AB. But choosing e∗ means she has a probability because doing good will make them feel good and help society.
(1-e∗) of losing power, in which case she gets y or probability e∗ of Hence, this is a strategy that will be used by all. The interesting
staying in power, in which case she gets an expected payoff of CD. challenge arises when this strategy is exhausted and the leader
All we need to show is that it is possible that: x + y + y > x + AB has to choose e, and hence that is what the model focusses on.
+ (1 – e∗) y + e∗CD. In other words, we have to show: (1 + e∗)y > AB + Finally, there is one important assumption used in the paper
e∗CD. that may be worth trying to generalise in the future. Note that in
From the fact of the leader’s optimization, we know that in the above model it is assumed that the leader chooses e, keeping
Fig. 2, AB is greater than both y and CD. But this is compatible in mind that the payoff of failing to remain in office is given by
with y being close to AB and CD being much smaller than y, which y. In reality, the y will itself depend on the e, and as e increases,
would make the above inequality valid. This completes the proof. y will fall, that is, the exit option for the dictator will get worse.
It is intuitively obvious that, in the general case, with T > 2, even Leaders with a little far-sightedness, even if they are not doing the
if a leader does the optimization for a few periods into the future full optimization over time, should be aware of this. This injection
at a time (instead of just one period at a time, as in the above of reality will complicate the model. However, if we assume that
example), but not the whole length of the future, she could end y depends on the accumulated evil indulged by the leader in the
up on a path where she has to indulge in increasing evil acts to previous terms in office, plus the evil done now, that is, e, to stay
stay in power, morphing fully into a tyrant after some stretch of in office for another term, it should be possible to build a more
time in office. sophisticated model which will continue to give a result similar
Before commenting on the normative and policy implications to the one obtained in the present paper.
of the above claim, I want to add a few technical comments and
caveats.
First, note that some neoclassical economists assume people
SOME NORMATIVE QUESTIONS
are always rational and so they reject some of the premises of This paper is part of a larger discourse on the power of the
behavioral economics. At times this becomes a tautology. They leader or the state, and the risk of oppression and violation of
assume people are always rational and, no matter how they basic human rights of individuals that goes back to the 17th
behave, they describe this as rational behavior. This leads to century, but has been a topic of contemporary interest, as part of
contortions of the model to ensure that dynamic inconsistency,
which implies a form of intransitivity of preference, cannot occur, 5 There is a literature that shows how this can happen and, more generally,
by definition. Interestingly, there is work in philosophy (see Parfit, discusses the case of infinite decision points, with one player making an infinite
number of decisions or a countably infinite set of players, each making one or a
1984), that asserts how, especially in domains of moral decision- finite number of decisions (Voorneveld, 2010; Cingiz et al., 2016; Rachmilevitch,
making, people may be fully rational and violate transitivity. 2020).
 RRH: Basu et al. | 5

a larger discourse on collective action and power (Hobbes, 1651; in mainstream neoclassical economics is that these moral and
Acemoglu and Robinson, 2019; Binmore, 2020; Ferguson, 2020; normative moorings are often critical for markets to function and
Lopez-Calva and Bolch, 2022; Basu, 2022b). It is possible to argue society to prosper (Basu, 2000, Chapter 4; Bowles and Gintis, 1998;
that leadership is often needed to achieve the social good, but Bowles, 2016). Further, these norms may change and get modified
there is always the risk of the leader going astray, or morphing over time, solving and occasionally creating new problems. These
into a force of evil, as the above model illustrates. Given this very norms are often so embedded in our psyche that we abide by them
real problem of political economy, what can we do by way of policy unwittingly. We do not bite others, even when that would clearly
and prior regulation to prevent such a predicament? How do we help us, not because this action is not available in the game of
put checks and balances on the leader? life, nor because on doing cost–benefit analysis, we find this does
Most problems of political economy are unlikely to be fully not give us a positive payoff, but simply because our norms have
solvable. This is because, in reality, there is no such thing as the programmed us not to even consider this action.
ultimate game of life. As you devise rules to prevent people or There are other ways in which the notion (common in eco-

Downloaded from https://academic.oup.com/ooec/article/doi/10.1093/ooec/odad002/7036634 by guest on 30 May 2024

players from indulging in bad behavior, they can keep finding new nomics) of human beings having exogenously given preferences
strategies, beyond the original game of life6 , to perpetrate evil. The or payoff functions, has been contested. We know, for instance,
only hope for well-meaning policymakers is to try to stay ahead that people often choose actions in order to enhance their self-
of the curve. It is in this spirit that I offer some closing comments, esteem (see Crocker and Wolfe, 2001; R. Akerlof, 2017b). Moreover,
aware that they raise more questions than answers. there is often a tussle between self-esteem and the esteem of
At this time, what the above theoretical model points to is peers. If being in power raises esteem (of oneself or of the peers),
the need for ex ante agreements, akin to constitutions, which a political leader may try harder to stay in power. If, on the other
debar certain kinds of actions by human beings. The key is to hand, giving up power, when one could have clung on to it, raises
have agreements in advance which deter certain behaviors in the esteem, the leader may be more inclined to quit office.
future. The awareness of these norms and endogenous elements in
Why would such agreements work? One reason, rooted in human preferences opens up a vast terrain of research and anal-
economics and game theory but not widely understood, is that ysis that goes beyond mainstream economics and straddles the
the social, political and economic world we inhabit has multi- terrain between sociology and economics (Kliemt, 2020; Coyle,
ple equilibria. Prior agreements are ways of nudging ourselves 2021; Basu, 2022a). This can also bring to light new avenues
away from the current equilibrium to another collectively more for modifying our behavior in order to create a better world. By
desirable equilibrium. The key word here is ‘equilibrium’. If this changing what we esteem in our peers, we can prompt our peers
is done successfully, the question of why people will adhere to to behave differently. This is a matter of some urgency in our
a constitution is easy to answer. If the constitution is a set of troubled world. Though the definitive answer to this may take
rules which has the property of constituting an equilibrium, then time to evolve, it is a direction of research worth pursuing in the
once it becomes focal it will be adhered to simply because it is in future.
each person’s self-interest to abide by the constitution if others Let me close by addressing a narrower matter, which arises
abide by the constitution. I am using the term ‘equilibrium’ in a directly from the above model. This concerns having global rules
somewhat unconventional way, to denote a stable vector of sets of concerning term limit. A tyrannical leader in one nation can
omitted strategies. Each player agrees not to use certain strategies destabilize the world. So, ideally, we need a term-limit require-
if other players agree not to use some of their strategies7 . ment built into an international charter or a global constitution.
By this criterion the challenge is to draft a constitution that If the T, in the above game, is capped by a charter at 2, there will
happens to be one of the equilibria of this society and then create be only that much evil a leader can morph into. What is more
pressures so that this equilibrium becomes a focal point to which important, is that leaders, being aware that they will have to exit,
people are willing to shift. Further, in our globalizing world the will be more conscious of their behavior when in office.8
need is for a global constitution and global agreements. Globaliza- It is possible to think of more immediate actions, that is, even
tion raises important questions of global justice (Hassoun, 2012). before a global agreement is reached. For this, we have to get away
Just as we have an obligation toward the global poor, we have an from our propensity to think of nations as individuals or players.
obligation toward those whose basic human rights are violated by Thus the Cuban missile crisis is typically modeled as a two-player
dictators and tyrants, no matter where they live. We are at a stage game, USA and USSR locked in a deadly Hawk-Dove scenario.
of history, where we have no escape but to be more intrusive in Likewise, there is a body of writing on the war games in the Korean
other nations. There will have to be lines, which, if nations cross Peninsula and what we can do to avert disaster. This is conducted
them, others acquire the right to intervene. mostly in terms of a game between North Korea, USA, China
While striving for such agreements and constitutions, it is and Russia, occasionally tossing in South Korea. Much of this
worth keeping in mind that there are other options, which require is done again treating nations as players. When individuals are
us to step beyond the conventional idea of equilibrium. The belief brought in, like Kim Jong-un, they are treated as having interests
that individuals choose from the universal set of all available totally aligned with their own nation—‘Kim Jong-un is an ardent
actions, and they choose rationally from this set, as mainstream nationalist who regularly responds to threats by upping the ante’
economics assumes, is deeply f lawed. First of all, we all have social (Menon, 2017).
norms and moral hard-wirings that we adhere to. What is ignored What is not always recognized in popular discourse but is
a central tenet of modern political economy, at least since the

6 Thereby raising troubling questions about the very meaning of ‘the game
of life’ (Basu, 2022a). 8 All real-life solutions come with caveats. It must be recognized that
7 The idea of an equilibrium of omitted strategies (instead of a vector of term limits have some disadvantages. It makes politicians focus on the short
selected strategies as in conventional game theory) leads to various concepts duration at the expense of long-term gains for society. On the other side, having
of set-valued equilibria in games (see Basu and Weibull, 1991; Arad and term limits can encourage dynasties, which act as a substitute for one leader
Rubinstein, 2019). These set-valued notions are germane to understanding the carrying on for too long. The reign of Kim Il-sung, Kim Jong-il and Kim Jong-un,
meaning of a constitution in a formal way. from 1953 to now, is, for all practical purpose, rule by one person.
6 | Oxford Open Economics, 2023, Vol. 00, No. 00

seminal work of Downs (1957), is that leaders have their own Basu, K. (2022a) ‘Conventions, Morals and Strategy: Greta’s Dilemma
interests, which are not aligned with the interest of the nation. and Incarceration Game’, Synthese, 100: 1–19.
Many dictators who have long been in power may want to exit Basu, K. (2022b) ‘Why Have Leaders at All? Hume and Hobbes, with
purely for their own interest, but they may not have a viable exit a Dash of Nash’, Homo Oeconomicus, 39: 1–26.
option9 . They know that, once out of power, they will be punished Basu, K., and Weibull, J. (1991) ‘Strategy Subsets Closed Under Ratio-
by their own people or even killed by the generals. nal Behavior’, Economics Letters, 36: 141–6.
It is not immediately clear how to solve this. Some may argue Binmore, K. (2020) Crooked Thinking or Straight Talk? Cham: Springer.
that the way to solve this is to create attractive exit options. Bowles, S. (2016) The Moral Economy: Why Good Incentives are No Substi-
This would amount to the US telling Kim Jong-un, ‘If you stop tute for Good Citizens, New Haven: Yale University Press.
oppressing your people and threatening other nations, we will Bowles, S., and Gintis, H. (1998) ‘The Moral Economy of Communities:
protect you by helping you leave your country, and give you a Structured Populations and the Evolution of Pro-Social Norms’,
castle on a pacific island to settle in.’ This could help us deal with Evolution and Economic Behavior, 19: 3–25.

Downloaded from https://academic.oup.com/ooec/article/doi/10.1093/ooec/odad002/7036634 by guest on 30 May 2024

Kim Jong-un and promote peace in his nation and the world. But Chakraborty, A. (2021) ‘Present Bias’, Econometrica, 89: 1921–61.
no problem in economics comes with the absolute final answer. Cingiz, K. et al. (2016) ‘Doing It, Now, Later or Never’, Games and
If the US uses this strategy regularly for world peace, we will Economic Behavior, 97: 174–85.
have another problem. Individuals with no interest in power and Cooter, R., and Schäfer, H. -B. (2012) Solomon’s Knot:
tyranny may now strive to become tyrants for no other reason but How Law can end the Poverty of Nations, Princeton: Princeton
to get that castle in the Pacific Island. University Press.
Coyle, D. (2021) Cogs and Monsters: What Economics is and What it Should
be? Princeton: Princeton University Press.
Acknowledgement Crocker, J., and Wolfe, C. T. (2001) ‘Contingencies of Self-worth’,
The paper began with a presentation at the Fifth Conference Psychological Review, 108: 593–623.
of Philosophy, Poli tics & Economics (PPE) Society, New Orleans, Downs, A. (1957) An Economic Theory of Democracy, New York: Harper.
3–5 February 2022, and was developed further as part of my Ferguson, W. (2020) Political Economy of Collective Action, Inequality and
W. Edmond Clark Distinguished Lecture at Queen’s University, Development, Stanford: Stanford University Press.
Canada, delivered on 6 April 2022. The nudge to bring the paper to Hassoun, N. (2012) Globalization and Global Justice: Shrinking
closure needed the free-f lowing conversation I had with Hartmut Distance, Expanding Obligation, Cambridge: Cambridge University
Kliemt. En route to completing the paper, I had conversations Press.
with and received comments from many. I owe special thanks Hobbes, T. (1651) Leviathan, 1994th edn edition, edited by E. Curley,
to George Akerlof, Nancy Chau, Chris Cotton, Joe Halpern, Nicole Hackett, Cambridge, MA.
Hassoun, Chenyang Li, Hans-Bernd Schäfer, Arunava Sen and Kinzer, S. (2021) ‘Ortega in His Labyrinth’, in New York Review of Books
Saloni Vadeyar. I want to thank the Bucerius Law School, Ham- September 23.
burg, for hosting me during the last lap of work and Cornell Kleinberg, J., and Oren, S. (2014) ‘Time-Inconsistent Planning: A Com-
University’s Microeconomic Theory workshop where I presented putational Problem in Behavioral Economics’, Proceedings 15th
the paper. Finally, I would like to thank two anonymous referees of ACM Conference on Economics and Computation (EC).
the journal for valuable comments, many of which are now part Kliemt, H. (2020) ‘Economic and Sociological Accounts of Social
of the text. Norms’, Analyse & Kritik, 42: 41–96.
Lopez-Calva, L. -F., and Bolch, K. (2022) ‘A Giant Blob of Oily
Ambiguity’, in Basu, K., and Mishra, A., (eds.) On the Use of the
References Concept of Power in Economics, Palgrave Macmillan: forthcoming:
Acemoglu, D., and Robinson, J. (2019) The Narrow Corridor: States, Law and Economic Development.
Society, and the Fate of Liberty, New York: Penguin. Menon, R. (2017) ‘Avoiding Apocalypse on the Korean Peninsula’,
Akerlof, G. (1991) ‘Procrastination and Obedience’, American Economic Salon, June 11.
Review, 81: 1–19. O’Donoghue, T., and Rabin, M. (1999) ‘Doing it Now or Later’, American
Akerlof, R. (2017a) ‘The Importance of Legitimacy’, World Bank Eco- Economic Review, 89: 103–124.
nomic Review, 30: S157–S165. Parfit, D. (1984) Reasons and Persons, Oxford: Clarendon Press.
Akerlof, R. (2017b) ‘Value Formation: the Role of Esteem’, Games and Phelps, E. S., and Pollak, R. A. (1968) ‘On Second-best National Saving
Economic Behavior, 102: 1–19. and Game-Equilibrium Growth’, Review of Economic Studies, 35:
Arad, A., and Rubinstein, A. (2019) ‘Multi-Dimensional Reasoning 185–199.
in Games: Framework, Equilibrium and Applications’, American Rachmilevitch, S. (2020) ‘A Note on Discontinuity and Approximate
Economic Journal: Microeconomics, 11: 285–318. Equilibria in Games with Infinitely Many Players’, Economics Let-
Basu, K. (1994) ‘Group Rationality, Utilitarianism and Escher’s ters, 193: 1–4.
Waterfall’, Games and Economic Behavior, 7: 1–9. Swedberg, R. (2005) Interest, New York: Open University Press.
Basu, K. (2000) Prelude to Political Economy: A Study of the Social and Voorneveld, M. (2010) ‘The Possibility of Impossible Stairways: Tail-
Political Foundations of Economics, Oxford and New York: Oxford Events and Countable Player Sets’, Games and Economic Behavior,
University Press. 68: 403–10.

9 It is worth emphasizing here that ‘interest’ has come to acquire a rather

unidimensional notion of economic interest and this has had wide influence
in the way we do our analysis and formulate policy. In contexts such as
the one analyzed in this paper, we have to remind ourselves that ‘interest’
is a complicated multi-dimensional idea (Swedberg, 2005). This may make
policymaking harder, but it can make policymaking more effective.