Risk Aversion, Prospect Theory, and Strategic Risk in Law - - PowerPoint PPT Presentation

Goal and Background Experimental Design Results

Risk Aversion, Prospect Theory, and Strategic Risk in Law Enforcement: Evidence From an Antitrust Experiment

• M. Bigoni1

S.O. Fridolfsson2

• C. Le Coq3
• G. Spagnolo 345

1Universit`

a di Padova

2IfN, Stockholm 3SITE, Stockholm 4Universit`

a di Tor Vergata, Roma

5CEPR

IAREP-SABE 2008

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 1 / 31

Goal and Background Experimental Design Results

Outline

1

Goal and Background

2

Experimental Design

3

Results

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 2 / 31

Goal and Background Experimental Design Results

Main question: how does deterrence work?

focus on organized crime

Premise:

• rganized crime as the equilibrium outcome of a repeated

strategic interaction: must rely on self enforcing contracts different policies for law enforcement affect the sustainability of this equilibrium in different ways Questions: through which cognitive and behavioral “channels” does deterrence work? do agents react to law enforcement measures in a fully rational way? if not, how do behavioral departures from the standard rationality paradigm interact with the different law enforcement policies?

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 3 / 31

Goal and Background Experimental Design Results

Optimal law enforcement

Risk neutrality and full rationality (Becker, 1968) ⇒ the expected fine determines deterrence.

increase the fine and decrease the probability of detection ⇒ minimize the cost of prosecution

The idea has been recently under debate... Risk aversion (Polinsky Shavell 2000 ):

• ptimal law enforcement is affected by the offenders’ risk attitude.

Organized crime/collusion

Incentive compatibility in “team crimes”: Stigler (1964)

Antitrust policy adopted ⇒ affects IC constraint. Rey (2003), Motta Polo (2003) ; Spagnolo (2000, 2004) ; Harrington (2008)

Higer strategic risk reduces trust, hence collusion: Spagnolo (2004), Blonski et al. (2004, 2007) )

higher strategic risk if leniency is granted to the whistleblowers; higher absolute fines increase strategic risk under leniency.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 4 / 31

Goal and Background Experimental Design Results

Optimal law enforcement

Risk neutrality and full rationality (Becker, 1968) ⇒ the expected fine determines deterrence.

increase the fine and decrease the probability of detection ⇒ minimize the cost of prosecution

The idea has been recently under debate... Risk aversion (Polinsky Shavell 2000 ):

• ptimal law enforcement is affected by the offenders’ risk attitude.

Organized crime/collusion

Incentive compatibility in “team crimes”: Stigler (1964)

Antitrust policy adopted ⇒ affects IC constraint. Rey (2003), Motta Polo (2003) ; Spagnolo (2000, 2004) ; Harrington (2008)

Higer strategic risk reduces trust, hence collusion: Spagnolo (2004), Blonski et al. (2004, 2007) )

higher strategic risk if leniency is granted to the whistleblowers; higher absolute fines increase strategic risk under leniency.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 4 / 31

Goal and Background Experimental Design Results

Behavioral Law and Economics

Law enforcement creates incentives to obey the rules understanding how agents react to incentives is essential to make them more effective: Frey and Jegen (2001); Fehr and Falk (2002); Gneezy and Rustichini (2004) A recent stream of works has applied behavioral insights from psychology to law and economics (Jolls, Sustein and Thaler 1998 ; Jolls 2007 ): Bounded rationality (endowment effect, loss-aversion,

• veroptimism, availability heuristic, self-serving bias, ...)

Bounded willpower (myopia, hyperbolic discounting...) Bounded self-interest (other regarding preferences, fairness...) affect the way people react to legal prescriptions ⇒ optimal law enforcement.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 5 / 31

Goal and Background Experimental Design Results

Behavioral Law and Economics

Law enforcement creates incentives to obey the rules understanding how agents react to incentives is essential to make them more effective: Frey and Jegen (2001); Fehr and Falk (2002); Gneezy and Rustichini (2004) A recent stream of works has applied behavioral insights from psychology to law and economics (Jolls, Sustein and Thaler 1998 ; Jolls 2007 ): Bounded rationality (endowment effect, loss-aversion,

• veroptimism, availability heuristic, self-serving bias, ...)

Bounded willpower (myopia, hyperbolic discounting...) Bounded self-interest (other regarding preferences, fairness...) affect the way people react to legal prescriptions ⇒ optimal law enforcement.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 5 / 31

Goal and Background Experimental Design Results

Factors in risk perception

Disregarding/overweighting small probabilities (Kahneman Tversky 1979 ): “Highly unlikely events are either ignored or overweighted, and the difference between high probability and certainty is either neglected or exaggerated.” Availability heuristic and salience of punishments (Tversky Kahneman 1982; Akerlof 1991; Slovic et al. 2004 ; Keller et al. 2006) punishment experienced in past periods may be associated with negative affect that might increase the level of risks perceived stronger affect as the fine gets larger or as detection becomes more frequent.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 6 / 31

Goal and Background Experimental Design Results

Factors in risk perception

Disregarding/overweighting small probabilities (Kahneman Tversky 1979 ): “Highly unlikely events are either ignored or overweighted, and the difference between high probability and certainty is either neglected or exaggerated.” Availability heuristic and salience of punishments (Tversky Kahneman 1982; Akerlof 1991; Slovic et al. 2004 ; Keller et al. 2006) punishment experienced in past periods may be associated with negative affect that might increase the level of risks perceived stronger affect as the fine gets larger or as detection becomes more frequent.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 6 / 31

Goal and Background Experimental Design Results

Behavioral Interaction between Risk and Trust

Organized crime requires trust among members of the criminal

• rganization.

Experimental evidence (trust game) on the relation between risk attitudes and trust is mixed: risk averse players less trustful: Karlan (2005),Sapienza et al. (2007),Schechter (2007) no significant relation: Eckel and Wilson (2004), Ashraf et al. (2006) Bohnet and Zeckhauser (2004): individuals more willing to take risks when the outcome is due to chance than when it depends

• n whether another player proves trustworthy ⇒ betrayal

aversion. Organized crime is a more complex game: “external” risk of being detected and sanctioned “internal” risk of being betrayed

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 7 / 31

Goal and Background Experimental Design Results

Behavioral Interaction between Risk and Trust

Organized crime requires trust among members of the criminal

• rganization.

Experimental evidence (trust game) on the relation between risk attitudes and trust is mixed: risk averse players less trustful: Karlan (2005),Sapienza et al. (2007),Schechter (2007) no significant relation: Eckel and Wilson (2004), Ashraf et al. (2006) Bohnet and Zeckhauser (2004): individuals more willing to take risks when the outcome is due to chance than when it depends

• n whether another player proves trustworthy ⇒ betrayal

aversion. Organized crime is a more complex game: “external” risk of being detected and sanctioned “internal” risk of being betrayed

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 7 / 31

Goal and Background Experimental Design Results

Behavioral Interaction between Risk and Trust

Organized crime requires trust among members of the criminal

• rganization.

Experimental evidence (trust game) on the relation between risk attitudes and trust is mixed: risk averse players less trustful: Karlan (2005),Sapienza et al. (2007),Schechter (2007) no significant relation: Eckel and Wilson (2004), Ashraf et al. (2006) Bohnet and Zeckhauser (2004): individuals more willing to take risks when the outcome is due to chance than when it depends

• n whether another player proves trustworthy ⇒ betrayal

aversion. Organized crime is a more complex game: “external” risk of being detected and sanctioned “internal” risk of being betrayed

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 7 / 31

Goal and Background Experimental Design Results

What we did

Ran a series on explicit collusion in oligopoly

Results also relevant for other forms of corporate crime, corruption, financial fraud, etc.

Simulated a repeated oligopoly in the lab, and embedded it in different law enforcement environments:

Absence of enforcement: collusion is allowed “traditional” antitrust law enforcement policies Leniency programs

Looked at how deterrence varies under these alternative policies depending also on:

subjects degree of risk aversion size of fines and probability of detection players’ experiences in previous rounds

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 8 / 31

Goal and Background Experimental Design Results

What we did

Ran a series on explicit collusion in oligopoly

Results also relevant for other forms of corporate crime, corruption, financial fraud, etc.

Simulated a repeated oligopoly in the lab, and embedded it in different law enforcement environments:

Absence of enforcement: collusion is allowed “traditional” antitrust law enforcement policies Leniency programs

Looked at how deterrence varies under these alternative policies depending also on:

subjects degree of risk aversion size of fines and probability of detection players’ experiences in previous rounds

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 8 / 31

Goal and Background Experimental Design Results

What we did

Ran a series on explicit collusion in oligopoly

Results also relevant for other forms of corporate crime, corruption, financial fraud, etc.

Simulated a repeated oligopoly in the lab, and embedded it in different law enforcement environments:

Absence of enforcement: collusion is allowed “traditional” antitrust law enforcement policies Leniency programs

Looked at how deterrence varies under these alternative policies depending also on:

subjects degree of risk aversion size of fines and probability of detection players’ experiences in previous rounds

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 8 / 31

Goal and Background Experimental Design Results

What we found

Our results suggest that:

strategic risk (weaker trust, worse coordination) is the main determinant of deterrence when leniency is granted to the whistleblowers. risk aversion enhances deterrence via the “external” risk of detection, but not via the “internal” risk of betrayal. ⇒ Risk attitude not related with trust and strategic risk availability heuristic is very important: harshness of sanctions experienced in past periods increases deterrence significantly. very small probabilities not disregarded (i.e. not rounded to zero).

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 9 / 31

Goal and Background Experimental Design Results

Experimental design - 1

In(de)finitely Repeated Differentiated Bertrand duopoly

15% of probability of being re-matched at the end of each period

Possibility to discuss lowest acceptable price (to form a cartel) before the price game termination rule: at least 20 periods and then 15% of probability

• f continuation

Treatment variables

Probability of detection and size of fine when joining a cartel Possibility to report the cartel before (secretly) and after the chosen prices become public information Presence of leniency for whistleblowers.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 10 / 31

Goal and Background Experimental Design Results

Experimental design - 1

In(de)finitely Repeated Differentiated Bertrand duopoly

15% of probability of being re-matched at the end of each period

Possibility to discuss lowest acceptable price (to form a cartel) before the price game termination rule: at least 20 periods and then 15% of probability

• f continuation

Treatment variables

Probability of detection and size of fine when joining a cartel Possibility to report the cartel before (secretly) and after the chosen prices become public information Presence of leniency for whistleblowers.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 10 / 31

Goal and Background Experimental Design Results

Experimental design - 1

In(de)finitely Repeated Differentiated Bertrand duopoly

15% of probability of being re-matched at the end of each period

Possibility to discuss lowest acceptable price (to form a cartel) before the price game termination rule: at least 20 periods and then 15% of probability

• f continuation

Treatment variables

Probability of detection and size of fine when joining a cartel Possibility to report the cartel before (secretly) and after the chosen prices become public information Presence of leniency for whistleblowers.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 10 / 31

Goal and Background Experimental Design Results

Experimental design - 1

In(de)finitely Repeated Differentiated Bertrand duopoly

15% of probability of being re-matched at the end of each period

Possibility to discuss lowest acceptable price (to form a cartel) before the price game termination rule: at least 20 periods and then 15% of probability

• f continuation

Treatment variables

Probability of detection and size of fine when joining a cartel Possibility to report the cartel before (secretly) and after the chosen prices become public information Presence of leniency for whistleblowers.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 10 / 31

Goal and Background Experimental Design Results

Experimental design - 1

In(de)finitely Repeated Differentiated Bertrand duopoly

15% of probability of being re-matched at the end of each period

Possibility to discuss lowest acceptable price (to form a cartel) before the price game termination rule: at least 20 periods and then 15% of probability

• f continuation

Treatment variables

Probability of detection and size of fine when joining a cartel Possibility to report the cartel before (secretly) and after the chosen prices become public information Presence of leniency for whistleblowers.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 10 / 31

Goal and Background Experimental Design Results

Experimental design - 1

In(de)finitely Repeated Differentiated Bertrand duopoly

15% of probability of being re-matched at the end of each period

Possibility to discuss lowest acceptable price (to form a cartel) before the price game termination rule: at least 20 periods and then 15% of probability

• f continuation

Treatment variables

Probability of detection and size of fine when joining a cartel Possibility to report the cartel before (secretly) and after the chosen prices become public information Presence of leniency for whistleblowers.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 10 / 31

Goal and Background Experimental Design Results

Experimental Design -2

stage game

1

Communication decision: simultaneous

2

Communication: exchange price signals for 30 secs.

3

Pricing: simultaneous

4

First possibility of reporting: before knowing competitor’s price

5

Information about prices and 2nd possibility of reporting

6

Detection

7

Summary Treatment Communication: steps 4, 5, 6 missing.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 11 / 31

Goal and Background Experimental Design Results

Payoff table and myopic best replies

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 12 / 31

Goal and Background Experimental Design Results

Payoff table and myopic best replies

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 12 / 31

Goal and Background Experimental Design Results

Payoff table and myopic best replies

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 12 / 31

Goal and Background Experimental Design Results

Experimental Design - 3

Investment game: endowment: 25 e risky investment: return = 2.5 with 50% prob., return= 0 with 50% prob.

Treatments

Antitrust fine (F) probability report report’s Policy

• f detection (α)

effects ANTITRUST 200 0.10 Yes pay the full fine 1000 0.02 300 0.2 1000 LENIENCY 200 0.10 Yes no fine (half the fine if both report) 1000 0.02 300 0.2 1000 COMMUNICATION No –

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 13 / 31

Goal and Background Experimental Design Results

Experimental Design - 3

Investment game: endowment: 25 e risky investment: return = 2.5 with 50% prob., return= 0 with 50% prob.

Treatments

Antitrust fine (F) probability report report’s Policy

• f detection (α)

effects ANTITRUST 200 0.10 Yes pay the full fine 1000 0.02 300 0.2 1000 LENIENCY 200 0.10 Yes no fine (half the fine if both report) 1000 0.02 300 0.2 1000 COMMUNICATION No –

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 13 / 31

Goal and Background Experimental Design Results

Results 1 - Institutions and Deterrence

Result (I)

The size of the actual fine affects deterrence regardless of the presence of leniency programs. The size of the expected fine only matters under ANTITRUST.

table

Possible to achieve higher deterrence while decreasing prosecution costs by increasing the size of the fine and reducing the effort spent in investigation. Mitigates a current concern: the many leniency applications keep the agency busy with prosecution, to the detriment of investigation ⇒ lower α.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 14 / 31

Goal and Background Experimental Design Results

Results 1 - Institutions and Deterrence

Result (I)

The size of the actual fine affects deterrence regardless of the presence of leniency programs. The size of the expected fine only matters under ANTITRUST.

table

Possible to achieve higher deterrence while decreasing prosecution costs by increasing the size of the fine and reducing the effort spent in investigation. Mitigates a current concern: the many leniency applications keep the agency busy with prosecution, to the detriment of investigation ⇒ lower α.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 14 / 31

Goal and Background Experimental Design Results

Results 2

Result (II)

Players’ degree of risk aversion is negatively correlated with their willingness to establish collusion under ANTITRUST treatments, but not under COMMUNICATION and LENIENCY treatments.

table

Risk attitude seemingly unrelated to trust in this setting.

Result (III)

The total size of the punishment faced by a subject negatively affects his willingness to collude both under ANTITRUST and under LENIENCY. The frequency of detection by the Antitrust Authority in a player’s past history of play does not.

table Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 15 / 31

Goal and Background Experimental Design Results

Results 2

Result (II)

Players’ degree of risk aversion is negatively correlated with their willingness to establish collusion under ANTITRUST treatments, but not under COMMUNICATION and LENIENCY treatments.

table

Risk attitude seemingly unrelated to trust in this setting.

Result (III)

The total size of the punishment faced by a subject negatively affects his willingness to collude both under ANTITRUST and under LENIENCY. The frequency of detection by the Antitrust Authority in a player’s past history of play does not.

table Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 15 / 31

Goal and Background Experimental Design Results

Results 3: Conviction effects

plot

Conviction has a symmetric outcome, if it hurts both players in the same way (e.g. cartel detected by the Authority, or reporting as punishment under Leniency) asymmetric outcome, if one of the cartel’s members gets hurt more than the other (e.g. only one player deviates and simultaneously reports)

Result (IV)

Subjects’ willingness to form a cartel is reduced immediately after conviction, and this effect is significantly stronger when conviction has asymmetric effects. Symmetric outcomes are much more frequent under Antitrust treatments (94.12%) than under Leniency (37.90%). possibly explains why deterrence is higher under leniency.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 16 / 31

Goal and Background Experimental Design Results

Results 3: Conviction effects

plot

Conviction has a symmetric outcome, if it hurts both players in the same way (e.g. cartel detected by the Authority, or reporting as punishment under Leniency) asymmetric outcome, if one of the cartel’s members gets hurt more than the other (e.g. only one player deviates and simultaneously reports)

Result (IV)

Subjects’ willingness to form a cartel is reduced immediately after conviction, and this effect is significantly stronger when conviction has asymmetric effects. Symmetric outcomes are much more frequent under Antitrust treatments (94.12%) than under Leniency (37.90%). possibly explains why deterrence is higher under leniency.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 16 / 31

Goal and Background Experimental Design Results

Results 3: Conviction effects

plot

Conviction has a symmetric outcome, if it hurts both players in the same way (e.g. cartel detected by the Authority, or reporting as punishment under Leniency) asymmetric outcome, if one of the cartel’s members gets hurt more than the other (e.g. only one player deviates and simultaneously reports)

Result (IV)

Subjects’ willingness to form a cartel is reduced immediately after conviction, and this effect is significantly stronger when conviction has asymmetric effects. Symmetric outcomes are much more frequent under Antitrust treatments (94.12%) than under Leniency (37.90%). possibly explains why deterrence is higher under leniency.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 16 / 31

Goal and Background Experimental Design Results

Summary

Strategic Risk:

under LENIENCY (where betrayal is cheaper), only the actual (not the expected) fine matters deterrence driven by the risk of being betrayed by the partner cartelist, more than by the risk of being detected

Risk and Trust:

risk averse players collude less under ANTITRUST when the decision to collude is a matter of trust, as under COMMUNICATION and LENIENCY, subjects’ attitude towards risk unimportant

Availability Heuristic: The sum of the fines paid by a subject in previous periods has a significant, negative effect on his willingness to communicate: ⇒ people’s perception of a risk is based on its vividness and emotional impact rather than on its actual probability. Perception of Small Probabilities: small prob. of detection not disregarded under ANTITRUST, where deterrence is driven mainly by the risk of being sanctioned.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 17 / 31

Goal and Background Experimental Design Results

Summary

Strategic Risk:

under LENIENCY (where betrayal is cheaper), only the actual (not the expected) fine matters deterrence driven by the risk of being betrayed by the partner cartelist, more than by the risk of being detected

Risk and Trust:

risk averse players collude less under ANTITRUST when the decision to collude is a matter of trust, as under COMMUNICATION and LENIENCY, subjects’ attitude towards risk unimportant

Availability Heuristic: The sum of the fines paid by a subject in previous periods has a significant, negative effect on his willingness to communicate: ⇒ people’s perception of a risk is based on its vividness and emotional impact rather than on its actual probability. Perception of Small Probabilities: small prob. of detection not disregarded under ANTITRUST, where deterrence is driven mainly by the risk of being sanctioned.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 17 / 31

Goal and Background Experimental Design Results

Summary

Strategic Risk:

under LENIENCY (where betrayal is cheaper), only the actual (not the expected) fine matters deterrence driven by the risk of being betrayed by the partner cartelist, more than by the risk of being detected

Risk and Trust:

risk averse players collude less under ANTITRUST when the decision to collude is a matter of trust, as under COMMUNICATION and LENIENCY, subjects’ attitude towards risk unimportant

Availability Heuristic: The sum of the fines paid by a subject in previous periods has a significant, negative effect on his willingness to communicate: ⇒ people’s perception of a risk is based on its vividness and emotional impact rather than on its actual probability. Perception of Small Probabilities: small prob. of detection not disregarded under ANTITRUST, where deterrence is driven mainly by the risk of being sanctioned.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 17 / 31

Goal and Background Experimental Design Results

Summary

Strategic Risk:

under LENIENCY (where betrayal is cheaper), only the actual (not the expected) fine matters deterrence driven by the risk of being betrayed by the partner cartelist, more than by the risk of being detected

Risk and Trust:

risk averse players collude less under ANTITRUST when the decision to collude is a matter of trust, as under COMMUNICATION and LENIENCY, subjects’ attitude towards risk unimportant

Availability Heuristic: The sum of the fines paid by a subject in previous periods has a significant, negative effect on his willingness to communicate: ⇒ people’s perception of a risk is based on its vividness and emotional impact rather than on its actual probability. Perception of Small Probabilities: small prob. of detection not disregarded under ANTITRUST, where deterrence is driven mainly by the risk of being sanctioned.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 17 / 31

Strategic Risk Risk Aversion Results References

Appendix

1

Strategic Risk

2

Risk Aversion

3

Results Institutions and Deterrence Experience and Learning Post-conviction Behavior

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 18 / 31

Strategic Risk Risk Aversion Results References

Strategic risk

Consider an infinitely repeated PD game, whose stage game is: C D C r s D t p The standard threshold value δ is identified by the constraint: r 1 − δ ≥ t + δ p 1 − δ ⇔ δ r − p 1 − δ

LR inc. to coop.

≥ t − r

• SR inc. to def.

Strategic risk approach suggests that the short run disincentive to cooperate (p − s) matters, so the new threshold value δ∗ > δ is determined by: δ r − p 1 − δ

LR inc. to coop.

≥ t − r + p − s

• total SR inc. to def.

Return Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 19 / 31

Strategic Risk Risk Aversion Results References

Strategic risk

Consider an infinitely repeated PD game, whose stage game is: C D C r s D t p The standard threshold value δ is identified by the constraint: r 1 − δ ≥ t + δ p 1 − δ ⇔ δ r − p 1 − δ

LR inc. to coop.

≥ t − r

• SR inc. to def.

Strategic risk approach suggests that the short run disincentive to cooperate (p − s) matters, so the new threshold value δ∗ > δ is determined by: δ r − p 1 − δ

LR inc. to coop.

≥ t − r + p − s

• total SR inc. to def.

Return Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 19 / 31

Strategic Risk Risk Aversion Results References

A measure for risk aversion

Frequency distribution of the risky investment

Positive (21.16%), significant (s.l. 0.001) correlation between gender and investment: male players invest more.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 20 / 31

Strategic Risk Risk Aversion Results References Institutions and Deterrence Experience and Learning Post-conviction Behavior

Institutions and Deterrence

A generalized linear mixed model (three-levels logit model) Dependent variable: communication decision in period t(when it is risky). The covariates are: Leniency: a dummy variable, indicating the presence of leniency programs. Fine: the actual fine F. Exp.Fine: the expected fine, αF. RiskAttitude: measured as the ratio between the amount of money put into the risky asset in the “investment game” and the total sum available (25). Interaction terms: to check wether the impact of size of the fine, expected fine and degree of risk aversion vary depending on the institutional framework.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 21 / 31

Strategic Risk Risk Aversion Results References Institutions and Deterrence Experience and Learning Post-conviction Behavior

Institutions and deterrence

regression

Coeff.

• Std. Err.

Constant 1.352∗ 0.701 Leniency

• 2.568∗∗

1.014 Fine

• 0.157∗∗

0.063 LenXfine 0.059 0.083 Exp.Fine

• 3.848∗∗∗

1.083 LenXexpfine 4.525∗∗∗ 1.473 RiskAttitude 0.457 0.907 AntiXra 1.077 0.795 LenXra

• 0.626

1.039 LogLikelihood

• 2686.947

#obs. 5398

∗p < 0.1, ∗∗p < 0.05, ∗∗∗p < 0.01

Return Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 22 / 31

Strategic Risk Risk Aversion Results References Institutions and Deterrence Experience and Learning Post-conviction Behavior

Institutions and deterrence

regression

Coeff.

• Std. Err.

Constant 1.352∗ 0.701 Leniency

• 2.568∗∗

1.014 Fine

• 0.157∗∗

0.063 LenXfine 0.059 0.083 Exp.Fine

• 3.848∗∗∗

1.083 LenXexpfine 4.525∗∗∗ 1.473 RiskAttitude 0.457 0.907 AntiXra 1.077 0.795 LenXra

• 0.626

1.039 LogLikelihood

• 2686.947

#obs. 5398

∗p < 0.1, ∗∗p < 0.05, ∗∗∗p < 0.01

Return Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 23 / 31

Strategic Risk Risk Aversion Results References Institutions and Deterrence Experience and Learning Post-conviction Behavior

Experience and Learning

A generalized linear mixed model (four-levels logit model) Dependent variable: communication decision in period t(when it is risky). The covariates are: Match: number of matches played since the beginning of the game. PeriodInMatch: number of periods elapsed since the beginning of the current match. Distrust: number of periods in which the player has been cheated upon / total number of periods in which the player has been part of a price setting agreement. RiskAttitude: measured as the ratio between the amount of money put into the risky asset in the “investment game” and the total sum available (25). PaidFine: sum of the fines paid up to the current period. Exper α frequency of detection experienced.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 24 / 31

Strategic Risk Risk Aversion Results References Institutions and Deterrence Experience and Learning Post-conviction Behavior

Institutions and deterrence

Communication Antitrust Leniency Match

• 0.0642

0.259∗∗∗

• 0.163

(0.137) (0.0869) (0.192) PeriodInMatch

• 0.112∗∗∗

0.0395

• 0.0348

(0.0334) (0.0528) (0.0381) Distrust 1.532 0.426 0.519 (1.182) (1.023) (0.453) RiskAttitude

• 0.860

2.004∗ 0.965 (0.828) (1.082) (0.665) PaidFine

• 1.803∗∗
• 0.857∗∗∗

(0.747) (0.315) Exper α

• 1.212

0.705 (1.063) (0.580) Constant 2.745∗∗∗

• 1.288∗∗∗
• 1.053∗∗∗

(0.792) (0.230) (0.282) LogLikelihood

• 354.455
• 909.794
• 1378.769

#obs. 736 1906 2756

Return Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 25 / 31

Strategic Risk Risk Aversion Results References Institutions and Deterrence Experience and Learning Post-conviction Behavior

Symmetry of conviction outcomes

Return

Symmetry has a significant effect on post-conviction communication decision even if we control for: the size of the fine period in game and period in match cumulated earning

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 26 / 31

Strategic Risk Risk Aversion Results References Institutions and Deterrence Experience and Learning Post-conviction Behavior

Symmetry of conviction outcomes

Return

Symmetry has a significant effect on post-conviction communication decision even if we control for: the size of the fine period in game and period in match cumulated earning

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 26 / 31

Strategic Risk Risk Aversion Results References

Bibliography I

AKERLOF, G. (1991): “Procrastination and Obedience,” The American Economic Review, 81, 1–19. ASHRAF, N., I. BOHNET, AND N. PIANKOV (2006): “Decomposing trust and trustworthiness,” Experimental Economics, 9, 193–208. BECKER, G. (1968): “Crime and Punishment: An Economic Approach,” The Journal of Political Economy, 76, 169–217. BLONSKI, M., P. OCKENFELS, AND G. SPAGNOLO (2007): “Co-operation in Infinitely Repeated Games: Extending Theory and Experimental Evidence,” . BLONSKI, M. AND G. SPAGNOLO (2003): “Prisoners’ Other Dilemma,” http://www.wiwi.uni-frankfurt.de/Professoren/ blonski/papers/PodRES.pdf. BOHNET, I. AND R. ZECKHAUSER (2004): “Trust, risk and betrayal,” Journal of Economic Behavior & Organization, 55, 467–484. ECKEL, C. C. AND R. K. WILSON (2004): “Is trust a risky decision?” Journal of Economic Behavior & Organization, 55, 447–465.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 27 / 31

Strategic Risk Risk Aversion Results References

Bibliography II

FEHR, E. AND A. FALK (2002): “Psychological foundations of incentives,” European Economic Review, 46, 687–724. FREY, B. AND R. JEGEN (2001): “Motivation Crowding Theory,” Journal of Economic Surveys, 15, 589–611. GNEEZY, U. AND A. RUSTICHINI (2004): “Incentives, Punishment and Behavior,” in Advances in Behavioral Economics, ed. by

• C. Camerer, 572–89.

HARRINGTON, J. E. (2008): “Optimal Corporate Leniency Programs,” Journal of Industrial Economics, 56, 215–246. JOLLS, C. (2007): “Behavioral Law and Economics,” NBER Working Paper. JOLLS, C., C. SUNSTEIN, AND R. THALER (1998): “A Behavioral Approach to Law and Economics,” Stanford Law Review, 50, 1471–1550. KAHNEMAN, D. AND A. TVERSKY (1979): “Prospect Theory: An Analysis of Decision under Risk,” Econometrica, 47, 263–292.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 28 / 31

Strategic Risk Risk Aversion Results References

Bibliography III

KARLAN, D. S. (2005): “Using Experimental Economics to Measure Social Capital and Predict Financial Decisions,” The American Economic Review, 95, 1688–1699. KELLER, C., M. SIEGRIST, AND H. GUTSCHER (2006): “The Role of the Affect and Availability Heuristics in Risk Communication,” Risk Analysis, 26, 631–639. MOTTA, M. AND M. POLO (2003): “Leniency Programs and Cartel Prosecution,” International Journal of Industrial Organization, 21, 347–379. POLINSKY, A. AND S. SHAVELL (2000): “The Economic Theory of Public Enforcement of Law,” Journal of Economic Literature, 38, 45–76. REY, P. (2003): “Toward a Theory of Competition Policy,” in Advances in Economics and Econometrics: Theory and Applications: Eighth World Congress, ed. by M. Dewatripont, L. Hansen, and

• S. Turnovsky, Cambridge University Press, chap. 3.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 29 / 31

Strategic Risk Risk Aversion Results References

Bibliography IV

SAPIENZA, P., A. TOLDRA, AND L. ZINGALES (2007): “Understanding Trust,” . SCHECHTER, L. (2007): “Traditional trust measurement and the risk confound: An experiment in rural Paraguay,” Journal of Economic Behavior & Organization, 62, 272–292. SLOVIC, P., M. FINUCANE, E. PETERS, AND D. MACGREGOR (2004): “Risk as Analysis and Risk as Feelings: Some Thoughts about Affect, Reason, Risk, and Rationality,” Risk Analysis, 24, 311–322. SPAGNOLO, G. (2000): “Optimal Leniency Programs,” Fondazione Eni Enrico Mattei http://www.feem.it/NR/rdonlyres/ F1EEE777-C479-4827-A600-8AA93E9AAE47/868/4200. pdf. ——— (2004): “Divide et Impera: Optimal Leniency Programs,” SSRN eLibrary. STIGLER, G. J. (1964): “A Theory of Oligopoly,” The Journal of Political Economy, 72, 44–61.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 30 / 31

Strategic Risk Risk Aversion Results References

Bibliography V

TVERSKY, A. AND D. KAHNEMAN (1992): “Advances in prospect theory: Cumulative representation of uncertainty,” Journal of Risk and Uncertainty, 5, 297–323.

Presenter: M. Bigoni (Univ. Padova) Deterrence of Organized Crime IAREP-SABE 2008 31 / 31