Trust, Reciprocity and Rules. Thomas Rietz, Eric Schniter, Roman Sheremeta, Timothy Shields

Transcription

1 Trust, Reciprocity and Rules Thomas Rietz, Eric Schniter, Roman Sheremeta, Timothy Shields

2 The chief lesson I have learned in a long life is that the only way to make a man trustworthy is to trust him; and the surest way to make him untrustworthy is to distrust him and show your distrust. Henry Lewis Stimson US Secretary of War US Secretary of State Governor General of the Philippines

3 What is Trust and Reciprocity (Trustworthiness)? Trust: Making oneself vulnerable to the actions of others Demonstrated by willfully ceding resources or control to another with the expectation that the other intends to reciprocate and not be opportunistic (i.e., be trustworthy) The amount of trust that can be demonstrated depends on the amount of resources that can be ceded and the risk undertaken Reciprocity/Trustworthiness Not taking advantage of another s vulnerability even when doing so is in one s own immediate self interest Demonstrated by not succumbing to opportunism so as to restitute the resources or control that another has ceded by extending trust The amount of trustworthiness that can be demonstrated depends on the amount of opportunism available

4 Motivation: Trust Relationships are Everywhere Trust and reciprocation is: Evolutionarily stable (Axelrod, 1984, The Evolution of Cooperation) Necessary for the development of societies (Dunbar, Knight and Power, 1999, The Evolution of Culture: A Historical and Scientific Overview) Necessary for the development of exchange economies (Kaplan, et al., 2012, Proceedings of the Royal Society: Biological Sciences) Complete contracting is: Neither possible nor desirable (Tirole, 1999, Econometrica)

5 Motivation: Use of Minimum Standard Rules Human subjects abuses IRB Rules Unfair labor practices Minimum Wage Unsafe driving Speed Limit Financial malfeasance Min. Reporting Stds. Question: Even if minimum standards eliminate the worst behavior, what is their effect on other behavior? Reinforcement of good behavior, making it even better? Elimination of behavior that otherwise would have exceeded the minimum standard?

6 Motivation: Institutional Engineering Sometimes Rules, Incentives and/or Enforcement Backfire Paying people to give blood reduces donations Titmuss, Oakley and Ashton, 1997, Social Policy Mellstrom and Johannesson, 2008, Journal of the European Economic Association Fines for late pickups increase late pickup rates at daycare facilities Gneezy and Rustichini, 2000, The Journal of Legal Studies Imposing monitoring or close supervision by authority decreases work effort (1984 in action!) Frey, 1993, Economic Inquiry Dickinson and Villeval, 2008, Games and Economic Behavior Griffith, 1993, Journal of Applied Social Psychology

7 OUR GOAL: Study the interaction of trust, reciprocity and rules to understand The effects of rules What drives trust and reciprocity in anonymous settings

8 The Trust/Investment Game s 0 Tripled by Experimenter (Known by all) 3s Investor Trustee Endowment = $10 Sends: s $20 Payoff = $10 s + r r 0 Endowment = $0 Receives: 3s Returns: r Payoff = 3s - r

9 Mean Behavior in the Trust/Investment Game In addition, higher s results in higher r Questions: Why invest (s>0), i.e., why trust? Why return (r>0), i.e., why reciprocate? Why do trust and reciprocity exist?

10 Why Trust and Reciprocity in the Game? Behavioral Game Theory Focal point Cooperative outcome Economics It shouldn t exist Should respond to risk/return tradeoffs Trust and Reciprocity Evolutionary Psychology It s hardwired Information returns Sociology Concrete normative behavior Part of larger social construct

11 Our Four Treatments s 0 Tripled by Experimenter (Known by all) 3s Investor Trustee Treatment Rules Zero (Baseline) None, i.e. r 0 Ten r 10% 3s Twenty r 20% 3s Thirty r 30% 3s

12 Hypotheses: Pure Rule Why expect a pure rule effect? By design, rules truncate lower tail of distribution This is the only DIRECT effect Rules 1. Truncate the lower tail of the return distribution mean, but not median, return rate downside risk for investors investment

13 Hypotheses: Reinforcement Why expect reinforcement? Rules reinforce social norms of trust and reciprocity and/or cooperation Rules lower risk to otherwise evolutionarily beneficial behavior Rules: 1. Truncate the lower tail of the return distribution 2. Shift the rest of the distribution up mean and median return rate downside risk and return for investors investment

14 Hypotheses: Replacement Why expect replacement? Rules substitute for social norms of trust and reciprocity and/or cooperation Rules create a new focal point Rules lower information returns to evolutionarily beneficial behavior Rules: 1. Truncate the lower tail of the return distribution 2. Shift the rest of the distribution down mean and median return rate monetary and informational returns for investors investment

15 A Note on the Hypotheses Only the replacement hypothesis is consistent with: A drop in median return rates with low rules, followed by increases in men and median return with sufficiently high rules, i.e., a u shaped return rate response function A drop in investment with low rules, followed by an increase with sufficiently high rules, i.e., a u shaped investment response function The replacement hypothesis implies that the ability to elicit more trustworthy behavior through higher trust is mitigated by rules

16 Hypotheses % Cut Replacement Reinforcement 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100% Return Rate

17 Results: Investment Levels Light bars: ROI < 0 Dark Bars: ROI > 0

18 Results: Return Rates Light bars: < $5 invested Dark bars: > $5 invested

19 Patterns only consistent with the replacement hypothesis. Results

20 Discussion Trust interactions involve a social dilemma: Nothing ventured nothing gained Trusting agents tradeoff self-interest and safety for potential benefits including Potential monetary returns Informational returns Trusted agents Reward trust Invest in a general stock of their own social and informational capital through reciprocity Rules Limit downside risk but inhibit informational returns Net effect: rules destroy trust

21 Example: The Monderman Principal No arrows No stop lights No street signs Low, if any, curbs Few lane markings Tacit negotiation, not rules, determines traffic flow Drachten, The Netherlands Population ~45,000 bestanden/evaluation%20laweiplein.pdf Reduced accident rate from >8/year to 1/year (since 2003) Reduced average crossing time 40%

22 Exogenous Rules Treatment Self Imposed Rules Exogenous rules limit the ability to display trust and trustworthiness Endogenously imposed rules may cue distrust or trustworthiness directly 1. Zero (Baseline): Trustee returns any portion of amount received 2. Ten (Baseline): Trustee returns at least 10% of amount received 3. Investor Decides: Investor selects Zero or Ten setting before investment 4. Trustee Decides: Trustee selects selects Zero or Ten setting before investment

23 PREVIEW OF ENDOGENOUS VERSUS EXOGENOUS IMPOSED RULES WITHIN THE TRUST GAME Co-Authors: Joyce Berg, Tom Rietz, Eric Schniter and Roman Sheremeta

24 Hypotheses in Endogenously Selected Rules What if the (Ten setting) rule is selected? The investor s choice of the rule cues distrust and leads to lower investment and return. The trustee s choice cues trustworthiness to the investor, leading to higher investment and return.

25 Endogenous Rules (preliminary) Results Investment Return Rate

26 Where do We Stand? Trust may work reasonably well in some situations If so, let sleeping dogs lie When trust fails, rules may help, but may hurt The tradeoff between direct and indirect effects needs to be considered There may be better ways to elicit trustworthy behavior

27 THANK YOU VERY MUCH

28 A Model of Trust/Trustworthiness

29 Investment and Trust Levels Investment Amount 2 Obs. = 89 = $ 4.50 $21.67Rule+ $133.33Rule + ε 2 Invested PseudoR = 14.46% Investment u shaped consistent with expecting rules to reduce trustee returns Trust (1.32) *** 3.41 (20.99) 1.03 (65.91) ** 2.09 Value At = Risk ( VAR) $4.50 $18.67Rule+ $23.33Rule (0.74) *** (11.75) 1.59 (36.91) 0.63 Trust unaffected by rules 2 + ε PseudoR Obs. = 89 2 = 23.75%

30 Return Rates Return Rates Return 2 Obs. = 82 = 33.33% % Rule % Rule + ε 2 Rate PseudoR = 7.28% (6.24%) *** 5.34 Return rates u or j shaped consistent with replacement hypothesis Interaction 5.56% % Investment (1.06%) Return ( %) *** 5.25 = Rate % Rule 13.97% Interaction+ ε (41.88%) ** 2.54 (97.42%) ** 2.48 (5.45%) ** 2.56 (303.72%) *** 2.88 Obs. = 82 2 PseudoR = 28.60% Investment increases return rates Rules mitigate this effect

31 Trust (VAR), Returns and Trustworthiness Return Rates 5.56% % Trust (6.27%) Return *** = Rate % Rule 7.94% Interaction+ ε (25.90%) *** 4.72 (0.97%) ( %) Obs. = 82 PseudoR = 27.90% Interaction no longer significant 2 Trust and Trustworthiness Discrentionary ReturnRate = ( Trustworthiness) 5.56% % Trust %) % Rule 7.94% Interaction+ ε (25.90%) 0.86 ( * (0.97%) ** 5.76 ( %) Trust drives trustworthiness PseudoR Obs. = 82 2 = 28.83%