APPLICATION OF GAME THEORY TO CRIMINAL LAW

LAW AND ECONOMICS- I

SUBMITTED BY: SUBMITTED TO:

VRINDA VINAYAK DR. RADHA SESHAN
ROLL NO. 76LLB14

2015

NATIONAL LAW UNIVERSITY, DELHI

 ACKNOWLEDGMENTS

I feel very privileged in expressing my gratitude to Dr. Radha Seshan for permitting me to move
forward with this research project and providing valuable inputs. This project could not have been
completed but for her able guidance. I would also like to thank my parents and my classmates who
encouraged me and assisted me directly or indirectly with my research.

Vrinda Vinayak
IInd year
B.A. LL.B. (Hons.)
National Law University
Delhi
 TABLE OF CONTENTS

Introduction………………………………………………………………………………………..1

Part A: Game Theory as Applied to Strategies of Law Enforcement Agencies…………………..3

Part B: Are Punishments Effective in Curbing Crime? ..................................................................7

Conclusion……………………………………………………………………………………….13

Bibliography……………………………………………………………………………………..15
 INTRODUCTION

Game theory explains how rational individuals make decisions when they are mutually
interdependent, and studies strategic behaviour.1 It uses games, which are formal models of
interactive behaviour involving one or more players, in which every player chooses optimal
strategies of behaviour in the pursuit of self-interest. Game theory was first used by Cournot in a
study of duopolies in 1838, but found a firm footing only after the publication of Von Neuman and
Oskar Morgenstein’s The Theory of Games and Economic Behaviour in 1944.2

Game theory has much to offer to the field of law and economics. The law can substantially alter
the strategies of the players involved by trying to negate the problems of adverse selection and
moral hazard by trying to alter their payoffs.3 Law, particularly criminal law, imposes obligations
through sanctions. There is a conflict that exists between the law and enforcement agencies, and
those who want to commit the crime, and this project aims to study this very conflict.

The researcher has divided this paper into two major parts, followed by a conclusion which
addresses the hypothesis and proposes suggestions. Part A being an application of game theory to
the conflict between the criminal and the law enforcement agencies, and following from this, Part
B deals with whether sanctions and punishments proposed by the law are effective in controlling
criminal behaviour. The game theories used in this paper are those proposed by scholars such as
George Tsebelis, R. Smith, H. Rauhat, Shreedhar Sasikumar etc.

Hypothesis

Unpredictability is an effective strategy to make the police’s strategy superior to the criminal, and
increase efficiency of the police system. Further, strict penal policies do not necessarily reduce
crime, and the Prisoner’s Dilemma game also goes against stringent punishments.

Research Questions

1. How can game theory be used to explain the possible outcomes of the conflict between
criminals and the police?

1
GRAHAM ROMP, GAME THEORY: INTRODUCTION AND APPLICATIONS 05-09 (1997).
2
Id.
3
ROMP, supra note 1 at 57.

1
 2. Are strict punishments prescribed by criminal law actually effective in curbing crime?
3. Does the ‘Prisoner’s Dilemma’ support a strict penal policy?

Research Methodology

The researcher has used doctrinal methodology for this paper, which involves collection if relevant
information from secondary sources such as books, journal articles and newspaper articles. Games
propounded by famous economists have been studied and applied to the Indian context. This paper
is purely doctrinal, and does not involve any empirical elements.

Format of Paper:

The paper has been presented in the form of a research article with an introduction, two parts (i.e.
Part A and Part B), and a conclusion.

2
 PART A: GAME THEORY AS APPLIED TO STRATEGIES OF LAW ENFORCEMENT AGENCIES

To apply game theory effectively to criminal law, it is vital to apply it to the conflict between the
law and criminals. A simple theory applied to a single police officer (X), and a single criminal (Y),
has been presented to demonstrate this conflict.4 For illustrative purposes, the game shall be
applied to murder, which is a widespread crime in India today. The interests of both parties in
such a situation are that the criminal is interested in committing the crime, and the police officer
is interested in preventing the crime. Game theory should be applied to the advantage of the law,
and those who seek to enforce it. Since solutions like barricading the entire potentially-dangerous
area and evacuating all citizens, and deploying more and more police officers to that area are not
practically viable, game theory may be applied to help the police officer analyse his situation and
take advantage of it, or at the very least, prevent the criminal from taking advantage of it.

There are several elements of the police officer’s problem that have to be considered. X can only
patrol one place at one time, and may not be able to keep an eye on the entire neighbourhood at
once. 5 Secondly, whenever Y makes a move, X may not be able to play his turn as he may not be
available (he may have taken a break from duty). X also has to efficiently utilise the limited
resources at his disposal to prevent more serious crimes than others.6 If we assume that the
neighbourhood he is patrolling houses a banker, and the finance minister, he will accord more
weight to protecting the minister if need be. But this also makes his pattern predictable for Y, who
will not fail to take advantage of it. X’s main problem seems to be this: He must delegate his time
between the banker and the minister’s house, but remain unpredictable as to the pattern of his
distribution.7 While this cannot be absolutely guaranteed, the game should try to maximize the
minimum ‘winnings’ (that is, the points he scores for being where he should be) for each of his
moves in the conflict game.8 To maximize a minimum means keeping the least possible amount
to be won at each play of the game at its highest possible level.9 Y’s moves have certain constraints
as well. He must murder either the banker or the minister, but cannot do both. He will try to kill

4
This game is based on the model proposed by R. Dean Smith.
5
R. Dean Smith, Random Patrol: An Application of Game Theory to Police Problems, 53 JOURNAL OF CRIMINAL
LAW, CRIMINOLOGY, AND POLICE SCIENCE 256, 258 (1962).
6
Id.
7
JOHN P. CRANK, UNDERSTANDING POLICE CULTURE 79 (1998).
8
SMITH, supra note 5 at 260.
9
Id.

3
the minster, as it would wreak more havoc and get his point across, and accords more weight to
that move. We assume that X and Y both accord the same number of points to the banker and the
minster, i.e. 1 and 2 respectively. If, after attack, both persons are still alive, X would win 3 points.
If the banker is murdered then the winning would 2 (since the minister is still alive), and if the
minister is killed the winning would only be 1 point. X’s gains are Y’s losses and vice versa.
Scholar R. Dean Smith proposes a simple and interesting 2x2 ‘Game Matrix’ where each square
represents the point where X intercepts Y, and the consequent winnings of X the police officer.
This may be used as follows:

Column 1 Column 2

Row 1 3 Minister (1)

X

Row 2 Banker (2) 3

FIGURE 1

Assuming that X will be moving along the rows, and Y along the columns. X must analyse this
game matrix and make his move accordingly, in order to protect both individuals and hold his
minimum possible winnings to a maximum. X feels that he should spend more time protecting the
minster since the havoc caused in society due to his murder would be greater if an ordinary civilian
was killed, but the rules demand that he protect both from the criminal. If he spends all his time
protecting either, Y will soon become aware of his pattern, attack the other, and be at an advantage
over the officer. Since X is uncertain about Y’s movements, he must protect both persons against
what the criminal ‘might’ do, which gives rise to the ‘assumption of cleverness’ i.e. the police
officer must assume that the criminal is at least as intelligent as himself, and will not make stupid,
predictable mistakes. 10 The next stage is to compute strategies, for which Column 2 is subtracted
from Column 1 in each row, and then the rows are switched.11 The results are shown in FIGURE 2.

10
Ibid at 261.
11
SMITH, supra note 5 at 265.

4
 Y

Column 1 Column 2 X's Strategies

Row 1 3 Bureaucrat (2) -2

X

Row 2 Minister (1) 3 1

FIGURE 2

Now to test X’s strategies, they must be played against Y’s moves. One knows that the strategy is
sound if it reveals the same result against any of the opponent’s moves.12 X’s potential points are
first multiplied with his strategies, irrespective of the minus signs. The two results are added, and
then divided by the sum of the strategies to arrive at an ‘average payoff’.13 The result of the
strategies if Y chooses Column 1 or Column 2 is shown in FIGURE 3 and FIGURE 4 respectively.

Column 1 X's Strategies

Row 1 3 2 3x2= 6 6+1=7

X 7/3= 2.33

Row 2 1 1 1x1=1

FIGURE 3

12
JOHN A. ETERNO & ELI B SILVERMAN, THE CRIME NUMBERS GAME 30 (2012).
13
Id.

5
 Y

Column 2 X's Strategies

4+3=7
Row 1 2 2 2x2=4
7/3=2.33
X

Row 2 3 1 3x1=3

FIGURE 4

It is found that whether Y selects Column I or Column 2, X’s expected winnings would be equal
i.e. 2.33. However, the condition is that the strategies must be selected in a random manner. It is
commonly known that strategies are worthless if they are known to the opponent’ if a hockey team
uses the same strategy for play every time, the opposing team would practice a way to counter it
and win the match. Similarly, if Y is aware of the strategy X will adopt, that makes X a predictable
police officer and his patrolling would be rendered ineffective. But, X must come up with a system
to devise random action too, since running from one point to another randomly would be an utter
waste of time and energy, and would lessen the expected payoff. An example of such a method
could be: At any given point of time, if the second hand of X’s watch is between 0 and 30, he
should select Row 1 to patrol, and if it is between 30 and 0, he should select Row 2. This strategy
still makes his moves random when looked at from Y’s perspective, as Y would not know when
X would decide to check his watch, and even if he does, it is unlikely that X and Y’s watched be
coordinated to the very last second so as to make X’s moves predictable to Y.

Looking at the above, it may be concluded that this example game would function only if both
parties are uncertain about each other’s actions since certainty leads to predictability, leading to a
consequent increase in potential winnings. This can be worked around to the police officer’s
advantage, by the use of ‘informants’.14 From this game, it is concluded that the key to effective
crime-patrolling lies in unpredictability of the law enforcement agencies’ actions, coupled with
means and methods to devise such random and unpredictable action.

14
SMITH, supra note 5 at 270. Instead of designating more police officers to patrol the area, which would be an
expensive proposition and may come at the cost of increasing crimes in other areas, the officer already in-charge could
designate some informants to increase predictability of the criminal’s activity.

6
 PART B: ARE PUNISHMENTS EFFECTIVE IN CURBING CRIME?

Efficacy of Punishments in Preventing Crime

Now that effective strategies for law enforcement agencies to curb crime have been discussed, it
is necessary to discuss the impact of punishments in crime prevention as well. Law enforcement
in India is based on the ‘rational actor’ model of traditional economics which says that people are
likely to perform an action only if the benefits from that action exceed the costs. This model says
that crime will be deterred if the expected punishment is strong enough15, a prediction that has not
been borne out in practice. Although long sentences are now common, the incidence of cognizable
crimes in India has increased by an astounding 212% since 1953.16 An example of such a
punishment could be the death penalty. The recent case of Yakub Memon’s hanging has fanned
the already raging debate about the death penalty in India, and abolitionists question if inflicting
capital punishment helps to deter perpetrators of heinous crimes such as rapes and murders. This
feeds into the larger question of whether punishments for crimes have a deterrent effect at all.

A positive answer to this question corresponds to a conservative stand, while a negative answer is
indicative of a liberal stance.17 The experts are divided too; the economists answer in the
affirmative, based on the method of intentional explanations, while sociologists maintain a
reserved stance based on causal relations.18 Methodologically, crime may be explained in two
ways:19

 Firstly, pre-existing factors may operate on the criminal (like negative socialization), or on
the victim (such as wealth). This forms the causal explanation, which suggests
improvement in living conditions as a remedy.
 Secondly, it may be a matter of choice on the criminal’s part, who may not act as a rational
individual at all times. This is the international explanation of crime, which is in favour of

15
Irving Piliavin et al., Crime, Deterrence, and Rational Choice, 51 AMERICAN SOCIOLOGICAL REVIEW 101, 109
(1986).
16
National Crime Records Bureau, Snapshots 1953-2006 (2012), http://ncrb.nic.in/ciiprevious/Data/CD-CII2006/cii-
2006/Snapshots%201953-2006.pdf (last visited Sep 20, 2015).
17
G. Tsebelis, Penalty has no Impact on Crime: A Game-Theoretic Analysis, 2 RATIONALITY AND SOCIETY 255, 261
(1990).
18
Id.
19
Id.

7
 increasing penalties which would lower the expected gains from committing the crime and
thereby, the frequency of the crime.

These two approaches lead to absolutely opposite policy prescriptions, as it a well-known principle
of criminal law in India that if the party committing the crime has done so with no intention, or
mens rea, and his crime is a product of his past experiences, he cannot be held guilty of the crime
in question.20 While the economic approach evaluates crime as individual choices and decisions
made by the potential criminal, game theory involves another actor, the police, as seen in the
previous chapter. Here, it shall be examined whether penalties increase or reduce the frequency of
the crime, based on the police’s propensity to enforce the law in the existence of such penalties,
through the Inspection Game.

Four assumptions have been made for the purpose of this game:

 Assumption 1: Criminals violate the law if the police do not enforce it.
 Assumption 2: Criminals abide by the law if the police enforce it.
 Assumption 3: Police do not enforce the law if there are no criminals around.
 Assumption 4: Police enforce the law if there are criminals around.

It would be easy to explain criminal behaviour based simply on these assumptions. Increasing
punishment would mean that the payoff for the criminal would decrease, keeping the level of police
control and enforcement constant, which is why crime would decrease. But does the level of police
control remain the same? Based on the assumptions, once criminal behaviour decreases, the level
of control would also decrease, which would be followed by a subsequent increase in crime,
leading to a continuous cycle. There is a better explanation through an application of game theory
and the concept of Nash equilibrium, pure strategy and mixed strategy.

Nash equilibrium is a situation wherein neither party has an incentive to modify his strategy based
on a change in the strategy of his opponent. In a condition of Nash equilibrium, both parties have
strategies that are optimal responses to each other.21 Pure strategy is a move that a player will
follow in every possible attainable situation in a game. Such moves are not random, or drawn from

20
WILLARD GAYLIN, THE KILLING OF BONNIE GARLAND 215 (1982).
21
Barick Chung, Notes on Boolean Functions and Nash Equilibrium, SSRN ELECTRONIC JOURNAL (2012),
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1997067 (last visited: Sep 25, 2015).

8
a distribution.22 For example, in every situation, the police officer will either have to enforce the
law or not enforce it. Similarly, the criminal will either commit the crime, or not commit the crime.
These are examples of pure strategies. A mixed strategy on the other hand, as the name suggests
is a mixture of pure strategies.23 Applying these concepts to the assumptions above, it is found
that no matter what the strategy of one of the parties is, there will be incentive for the other party
to modify his strategy, that is, there is no Nash equilibrium in pure strategies.24

Criminals perform a cost-benefit analysis for a certain crime with a given punishment level and
detection probability. Punishment is only threatening if law is enforced and criminals are actually
detected; however, enforcement and detection takes effort.25 Rational control agents will try to
optimize their level of control according to a specific crime level and vice versa. Game theory can
provide predictions on the interaction between crime, punishment and control rates. George
Tsebelis in 1990 pioneered the game theoretic approach of crime and punishment, the premise of
which rests on the assumption explored in the previous chapter; that criminals and law enforcement
agencies have conflicting interests and incentive structures. While criminals like to commit crimes,
control agents like to catch them. If criminals knew that they are not watched, they would prefer
to commit a crime. However, if they committed a crime and control agents knew it, control agents
would prefer to take some efforts to detect them (i.e. to improve their careers). As a consequence,
criminals would resign from crime, which makes control again unattractive.

Players want to avoid predictability in such a situation, which is why they may keep changing their
behaviour and there would no solution in pure strategies. Game theory predicts that criminals and
control agents choose a certain probability for their action, and this choice of probability is called
a mixed strategy. The chosen probability shall leave the opponent indifferent.26 Thus, criminals
calibrate their probability to commit a crime so that control agents are indifferent between control
and no control. The other way round, control agents choose their probability of control so that
criminals are indifferent between committing a crime or not. An example of a mixed strategy which
leaves the opponent ‘indifferent’ could be the watch example from the last chapter, where the

22
Id.
23
Id.
24
Id.
25
TSEBELIS, supra note 17 at 260.
26
H. Rauhut, Higher Punishment, Less Control? Experimental Evidence On the Inspection Game, 21 RATIONALITY
AND SOCIETY 359, 370 (2009).

9
police officer gives a method to his unpredictability that the criminal cannot gauge. Since the
criminal doesn’t know when the police officer may catch him, he will not factor in the police
officer’s move while making a decision to commit or not to commit a crime, and hence will be
‘indifferent’ between these two options. It is also found that criminals are sensitive for payoffs of
control agents and control agents are sensitive for payoffs of criminals. Consequently, higher
punishment does not deter crime but control, since the payoff for the criminal decreases.

The above proposition may be elaborated upon as follows.27 Individuals may decide to commit a
crime with payoff y and punishment costs p if caught. If they do not commit a crime, their payoffs
remain unchanged. Inspectors can decide to inspect criminals. They have to invest inspection costs
k to detect the action of the criminal. If an inspector catches an individual for having committed a
crime, the inspector receives the reward r. If not successful, inspection costs are lost. If they do
not inspect, they remain at their income level. For the criminal, it is assumed that undetected crime
is attractive, and higher punishment is a threat, so that p > y > 0. This makes the crime unattractive.
For inspectors, it is assumed that inspectors gain from successful inspection so that r > k > 0. From
these equations and the four assumptions above, it is seen that there is no Nash equilibrium in pure
strategies, and a mixed strategy has to be adopted. The optimal choice is to choose a mixture of
strategies that leaves the opponent indifferent. If alter is not indifferent, she will take advantage
and exploit ego on ego’s costs and vice versa. Hence, both actors should calibrate their chosen
probability with the utility function of their opponent in such a way that she is indifferent.

Once the Nash equilibrium for these mixed strategies was mathematically computed, scholars
found that less attractive crimes (where punishment is greater than payoff) would cause less
inspection, not less crimes. Conclusively, higher punishment would have no impact on crime;
higher punishment would only cause less inspection.28 One theory suggests that criminals continue
to commit crimes despite increased punishments because most of them are not the dispassionate
rational actors who populate standard economic models, rather they are stand blinded by the
temptation of immediate reward and largely untroubled by the possibility of delayed or uncertain
punishment.29 The evidence suggests that when hardened criminals are reasonably sure that they

27
Ibid at 359.
28
RAUHUT, supra note 26 at 360.
29
MARK KLEIMAN, WHEN BRUTE FORCE FAILS 51-60 (2009).

10
will be caught and punished swiftly, even mild sanctions deter them, but not even the prospect of
severe punishment is effective if offenders think they can get away with their crimes.30

Efficacy of Punishment in Securing Convictions

When discussing punishment, it is also necessary to discuss the famous ‘Prisoner’s Dilemma’
Game and link it to incentives. Can punishment be used as an incentive? Carrying on the example
from the previous chapter, and assuming that there were two co-conspirators who were successful
in murdering the minister. Both of them were arrested and detained for interrogation but the police
has insufficient evidence to convict either of them for murder. When the police investigate a crime,
they usually suffer from an information asymmetry, since the criminals obviously know more
about the crime and the method of commission than the police. If these conspirators co-operate
with each other, it becomes difficult for the police to obtain evidence and secure a conviction.
Applying the traditional version of the game, each conspirator can either confess thereby
implicating the other, or keep mum. No matter what the other suspect does, each can improve his
own position by confessing; in case the other confesses, then one should do the same to avoid a
harsher punishment, and if the other stays silent, one could obtain leniency under state laws by
confessing.31 Either way, confession is the dominant strategy for each.32 But when both confess,
the outcome is worse for both than when both keep silent.33

Leniency policies are vital for this game to function and show results, by getting one of the two or
more conspirators to cave. By offering reduced punishments, leniency systems give individual
conspirators an incentive to ‘betray’ the conspiracy, which provides enforcers with the necessary
information to prove the conspiracy and punish the participants.34 It may be conclusively stated
that in such a scenario, maintaining high punishments would prove detrimental to the interests of
the state, and the penalty has to be lowered in order to secure cooperation from one or more of the

30
Ibid at 245.
31
A leniency policy is an agreement that a member(s) of the conspiracy who assists law enforcement in proving a
conspiracy will be given a reduced punishment, or perhaps no punishment at all.
32
Avinash Dixit, Prisoners’ Dilemma, THE CONCISE ENCYCLOPEDIA OF ECONOMICS, LIBRARY OF ECONOMICS AND
LIBERTY (2012), http://www.econlib.org/library/Enc/PrisonersDilemma.html (last visited Sep 26, 2015).
33
Id.
34
Shreedhar Sasikumar, All in the Game – How the Game Theory Framework Helps Design and Understand Leniency
Laws, The myLaw.net Blog, http://blog.mylaw.net/all-in-the-game-how-the-game-theory-framework-helps-design-
and-understand-leniency-laws/ (last visited Sep 26, 2015).

11
conspirators. This may be illustrated with an example. Assuming that the murder of the minister
is punishable with life imprisonment, but the police only has enough evidence to have the criminals
convicted for a crime punishable with 10 years’ imprisonment.

1. NO LENIENCY IN PENAL POLICY

B is Silent B Talks

A is Silent A: 10 Years, B: 10 Years A: Life, B: Life

A Talks A: Life, B: Life A: Life, B: Life

TABLE 1
2. LENIENCY IN PENAL POLICY

B is Silent B Talks

A is Silent A: 10 Years, B: 10 Years A: Life, B: 0

A Talks A: 0, B: Life A: 6 Years , B: 6 Years

TABLE 2

This shows that leniency policies can be an effective weapon for the police to discover, break and
prove conspiracies. It may also be said that leniency policies act as a deterrent for criminal
conspiracies, since if each conspirator knows that their partner has an incentive to betray them in
the future, they are less likely to enter into the conspiracy at the very beginning.35 This is especially
true when leniency is offered only to the first conspirator to come forward, creating a potential
‘race to betrayal’.36 But there is another side to the coin, where leniency can provide incentives for
false testimony or even encourage the formation of conspiracy groups. Further, leniency must be
adjusted against the actual magnitude of punishment, and must be preceded by effective
enforcement, detection and investigation mechanisms.

35
T. Kippenberger, The Prisoner's Dilemma, 2 THE ANTIDOTE 8 (1997).
36
Id.

12
 CONCLUSION AND SUGGESTIONS

The hypothesis proposed by the researcher viz. “Unpredictability is an effective strategy to make
the police’s strategy superior to the criminal, and increase efficiency of the police system,” has
been proven to be true through the games applied in Part A. Findings from the research paper
reveal that, based on game theory, the key to effective crime-patrolling lies in unpredictability of
the law enforcement agencies’ actions, coupled with means and methods to devise such random
and unpredictable action. Part B of the paper reveals that increasing penalties and punishments
does not help to deter crime but only results in lax law enforcement. As seen through the Prisoner’s
Dilemma game illustrated in Part B, there is also a need to reduce potential punishment for
approvers and those cooperating with the law enforcement agencies if an increased number of
convictions are to be secured. This proves the second part of the hypothesis.

Unpredictability having been established as the key, randomized checkpoints can be established
to catch criminals. Milind Tambe has talked about how randomised checkpoints can be set up in
cities. "Humans tend to fall into predictable patterns and checking in a randomised fashion can be
applied to any situation, like drug trafficking and this has great potential for controlling corruption
and crime," he says.37

As stringent punishments are largely ineffective, the researcher further proposes a solution based
on Mark Klein’s model to curb crime efficiently.38 In his book, “When Brute Force Fails,” he
argues that instead of making punishments more severe, the authorities should increase the odds
that lawbreakers will be apprehended and punished quickly.39 This is an idea based on game theory
and economics of signaling behaviour. One way to make apprehension and punishment more likely
is to spend more on law enforcement. But it is not necessary to increase the budget, if enforcement
is viewed as a dynamic game in which strategically chosen deterrence policies become self-

37
Mishita Mehra, India Has Started Taking Theoretical Techniques like Game Theory Seriously While Framing
Policies, THE ECONOMIC TIMES, September 25, 2011, http://articles.economictimes.indiatimes.com/2011-09-
25/news/30198485_1_game-theory-security-strategy-mumbai-airport (last visited Sep 20, 2015).
38
Robert Frank, A Smarter (And Cost-Efficient) Way to Fight Crime, N.Y. T IMES, October 04, 2009,
http://www.nytimes.com/2009/10/04/business/economy/04view.html?_r=0 (last visited Sep 25, 2015).
39
Id.

13
reinforcing.40 If offense rates fall enough, a tipping point is reached, when even modest
enforcement resources can hold offenders in check.41

Applying this technique to violent organized crime in Mumbai, where such gangs are too numerous
for police to watch them all closely. Members can violate the law with impunity if they are
uncertain of being caught and punished. In such situations, applying the above solution, the police
can gain leverage by publicizing an enforcement priority list. Supposing that all drug violence in
a city is committed by members of one of three hypothetical gangs X, Y and Z, and that the
authorities have enough staffing to arrest and prosecute offenders in only one gang at any one time.
Adopting Klein’s solution, the police should publicly announce that their first priority henceforth
will be offenders in one specific gang, like X. This step would persuade members of that gang that
further offenses will result in swift and sure punishment, which is enough to deter them. The police
can then shift their focus to Y, who are likely to react in the same manner. When focus shifts to Z,
X and Y are not likely in indulge in crime again is that they always remain atop the enforcement
priority list. If they start offending again, police attention will again quickly focus entirely on them.
Since humans tend to fall into predictable patterns, random and unpredictable checkpoints can also
be very effective in curbing crime.

When crime rates are high, spreading existing law enforcement resources equally among all
potential offenders guarantees that any specific offense is unlikely to be punished. That leads to a
vicious cycle of further increases in crime and further strains on law enforcement. Klein’s proposal
to publicly announce a strategic priority enforcement list is likely to create positive feedback.

40
KLEINER, supra note 29 at 137.
41
Ibid at 141.

14
 BIBLIOGRAPHY

BOOKS:

1. Crank, J. (1998). Understanding police culture. Cincinnati, OH: Anderson Pub.

2. Eterno, J. and Silverman, E. (2012). The Crime numbers game. Hoboken: CRC Press.
3. Gaylin, W. (1982). Killing of Bonnie Garland, the. New York: Penguin Books.

4. Kleiman, M. (2009). When brute force fails. Princeton: Princeton University Press.

5. Romp, G. (1997). Game theory. Oxford [England]: Oxford University Press.

JOURNAL ARTICLES:

1. Chung, B. (n.d.). Notes on Boolean Functions and Nash Equilibrium. SSRN Electronic
Journal. Available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1997067
[Accessed 25 Sep. 2015].
2. Dixit, A. (2012). Prisoners’ Dilemma. [online] The Concise Encyclopedia of Economics,
Library of Economics and Liberty. Available at:
http://www.econlib.org/library/Enc/PrisonersDilemma.html [Accessed 26 Sep. 2015].

3. Game Theory and Applications. (2006). International Game Theory Review, 08(02),
pp.327-327.

4. Kippenberger, T. (1997). The Prisoner's Dilemma. The Antidote, 2(4), pp.8-10.

5. Piliavin, I., Gartner, R., Thornton, C. and Matsueda, R. (1986). Crime, Deterrence, and
Rational Choice. American Sociological Review, 51(1), p.101.
6. Rauhut, H. (2009). Higher Punishment, Less Control?: Experimental Evidence On the
Inspection Game.Rationality and Society, 21(3), pp.359-392.
7. Sasikumar, S. (2010). All in the game – how the game theory framework helps design and
understand leniency laws | the myLaw.net blog. [online] Blog.mylaw.net. Available at:
http://blog.mylaw.net/all-in-the-game-how-the-game-theory-framework-helps-design-
and-understand-leniency-laws/ [Accessed 26 Sep. 2015].
8. Smith, R. (1962). Random Patrol. An Application of Game Theory to Police Problems. The
Journal of Criminal Law, Criminology, and Police Science, 53(2), p.258.

15
 9. Tsebelis, G. (1990). Penalty has no Impact on Crime: A Game-Theoretic
Analysis. Rationality and Society, 2(3), pp.255-286.

NEWSPAPER ARTICLES:

1. Frank, R. (2009). A Smarter (And Cost-Efficient) Way To Fight Crime. The New York
Times. [online] Available at:
http://www.nytimes.com/2009/10/04/business/economy/04view.html?_r=0 [Accessed 25
Sep. 2015].

2. Mehra, M. (2011). India has started taking theoretical techniques like game theory
seriously while framing policies. The Economic Times. [online] Available at:
http://articles.economictimes.indiatimes.com/2011-09-25/news/30198485_1_game-
theory-security-strategy-mumbai-airport [Accessed 20 Sep. 2015].

REPORTS:

1. National Crime Records Bureau, (2012). Snapshots 1953-2006. [online] National Crime
Records Bureau. Available at: http://ncrb.nic.in/ciiprevious/Data/CD-CII2006/cii-
2006/Snapshots%201953-2006.pdf [Accessed 20 Sep. 2015].

16