Incentives and Behavior Prof. Dr. Heiner Schumacher KU Leuven 8. - - PowerPoint PPT Presentation

Incentives and Behavior

• Prof. Dr. Heiner Schumacher

KU Leuven

• 8. Social Preferences
• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

1 / 36

Introduction

A standard assumption in economics is that individuals only care about their own monetary payo¤. However, there is lots of evidence that individuals do not always act sel…shly. In this lecture, we look at the evidence for pro-social motivations and what it implies for decision making in markets and organizations.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

2 / 36

Introduction

Overview Experimental Evidence Behavioral Motives Robustness

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

3 / 36

Experimental Evidence

In the dictator game, there are two parties: the dictator and a receiver. The dictator has a certain endowment, say 10 EUR. She decides how much of this endowment she wants to give to the receiver. The receiver is passive, i.e., she has to accept whatever the dictator gives to her (so if the dictator keeps the 10 EUR for herself, the receiver walks away with 0 EUR).

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

4 / 36

Experimental Evidence

What will a sel…sh dictator do? What economic situation does this game capture?

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

5 / 36

Experimental Evidence

On average, dictators give 20 percent of their endowment. There is substantial heterogeneity. There are many sel…sh dictators (around 30 percent of subjects give nothing). However, there are also many dictators who give 50 percent of their endowment. Almost nobody gives more than 50 percent. These …ndings are robust to anonymity, stakes, cultural background, etc. Behavior in dictator games proves that many individuals care about the welfare of others.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

6 / 36

Experimental Evidence

In the ultimatum game we have the same parties as in the dictator game: a dictator and a receiver. The dictator again has a certain endowment, say 10 EUR. She proposes how to split this endowment between her and the receiver, e.g., 10 x for the dictator and x for the receiver. After observing the dictator’s proposal, the receiver decides whether she accepts or rejects it. If the receiver accepts the o¤er, she gets x, while the dictator gets 10 x. Otherwise, both players get 0.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

7 / 36

Experimental Evidence

What outcomes are Nash outcomes if both players are purely sel…sh? What outcomes can occur in a subgame-perfect Nash equilibrium if both players are purely sel…sh? What economic situation does this game capture?

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

8 / 36

Experimental Evidence

Most dictators o¤er something between 40 and 50 percent of their

• endowment. Few subjects o¤er more than 50 percent.

There are few o¤ers below 20 percent. Low o¤ers (20 percent or less) are rejected half of the time. The probability of rejection decreases in x. Again, these results are relatively robust with respect to scale (in some studies the dictators’ endowment were several monthly salaries). Demographic variables such as age, gender, education typically have no in‡uence. The ultimatum game highlights an extreme con‡ict between sel…shness and fairness. The results show that many individuals care for fairness. Dictators seem to anticipate these fairness considerations.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

9 / 36

Experimental Evidence

In the trust game (or investment game), there are two parties: a …rst-mover and a second-mover. The …rst-mover has an endowment that she can keep it for herself or invest into the second-mover. Let x be the amount she gives to the second-mover. The invested amount is tripled by the experimenter before it reaches the second-mover. The second-mover therefore receives 3x. She then decides how much she gives back to the …rst-mover. Let y be the amount she returns to the …rst-mover. At the end of the game, the …rst mover earns (10 x) + y, while the second-mover earns 10 + 3x y. Economists frequently refer to the amount x as a measure of “trust”, and the amount y as a measure of “trustworthiness” or “reciprocity”.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

10 / 36

Experimental Evidence

What should players do in order to maximize the sum of payo¤s? What outcomes can occur in a subgame-perfect Nash equilibrium if both players are purely sel…sh? What economic situation does this game capture?

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

11 / 36

Experimental Evidence

First-movers invest on average 50 percent into the second-mover. Second-movers return on average 95 percent of the …rst-mover’s investment (i.e., sometimes trust pays o¤, sometimes not). Again, behavior is very heterogeneous. Half of the second-movers return nothing or a very small amount. The trust game is a simple way to measure peoples’ trust. Behavior in the trust game shows that many individuals trust their anonymous

• counterparts. On average, second-movers reciprocate this trust.
• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

12 / 36

Behavioral Motives

Experimental evidence suggests that many individuals do not act completely sel…shly. Which motives drive pro-social behavior? We discuss the most important behavioral motivations: altruism, reciprocity, and image concerns.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

13 / 36

Behavioral Motives

Altruism The behavior in dictator games implies that individuals derive utility from both their own and others’ payo¤. We have to distinguish between pure and impure altruism (Andreoni 1989).1 A purely altruistic individual derives utility from the payo¤ of others. For example, if x is one’s own payo¤, and y is the payo¤ of another person, then a purely altruistic individual has the utility function U(x, y), which is strictly increasing in both components. Individuals who engage in impure altruism receive utility from the act

• f giving (i.e., they experience a “warm glow”). Let g be the amount

that an individual gives to another person. If she is an impure altruist, her utility function is given by U(x, y, g), which is strictly increasing in the last component.

1Andreoni, James (1989): “Giving with Impure Altruism: Applications to Charity and

Ricardian Equivalence,” Journal of Political Economy 97(6), 1447 - 1458.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

14 / 36

Behavioral Motives

The distinction between pure and impure altruism is important for many applications. Consider, for example, the provision of a public good like donations to a charity. The government can …nance the charity from public funds,

• r the charity can collect private donations.

If all individuals are pure altruists, then an increase in public funding fully crowds out private donations (why?). If some individuals are impure altruists, crowding out is not complete, i.e., funding for the charity increases (why?). Empirical data suggests that crowding out is rather low, an indication for impure altruism.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

15 / 36

Behavioral Motives

Reciprocity The behavior in the trust and the gift exchange game suggests that many individuals behave reciprocal manner: they are kind (unkind) to those who previously were kind (unkind) to them. According to sociologists and anthropologists, one of the most widespread norms of human culture is the “rule of reciprocation”. It requires that one person tries to repay, in kind, what another person has provided. We are trained from childhood to comply to the rule or su¤er serious social disapproval. The following experiment shows how the rule of reciprocation enables to mutual cooperation.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

16 / 36

Behavioral Motives

Fehr and Gächter (2000) compare behavior in the public good game with and without the option to punish those who do not contribute to the “public good”.2 The public good game is played in groups of four. Each group member i has an endowment of 20 tokens and can decide what amount gi he wants to contribute to the public good. The payo¤ of group member i is then πi = 20 gi + 0, 4

4

∑

j=1

gj. What strategy pro…le maximizes the group’s total payo¤? What is the optimal strategy for each group member (given that all members are purely sel…sh)?

2Fehr, Ernst, and Simon Gächter (2000): “Cooperation and Punishment in Public

Goods Experiments,” American Economic Review 90(4), 980 - 994.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

17 / 36

Behavioral Motives

All subjects play the public good game with and without punishment

• ption.

If there is no punishment option, the game is over if payo¤s are realized. If there is a punishment option, then after the public good game is played (and all group members are informed about the decisions of their opponents), there is a second stage in which they have the

• pportunity to punish each other.

Speci…cally, group member j can punish group member i by assigning punishment points pi

j to i. Each punishment point assigned to i

reduces her payo¤ πi from the …rst stage by 10 percent (it cannot get negative though). Punishment is costly (see the next slide).

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

18 / 36

Behavioral Motives

The …nal payo¤ of group member i after the punishment stage is π…nal

i

= πi " 1 1 10 min ( ∑

j6=i

pi

j, 10

)# ∑

j6=i

c(pj

i ),

where c(pj

i ) is the cost member i incurs for the punishment of

member j. The cost of punishment are as follows: pj

i

1 2 3 4 5 6 7 8 9 10 c(pj

i )

1 2 4 6 9 12 16 20 25 30 Each condition is repeated 10 times with stranger matching (so that there can be no reputation e¤ects).

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

19 / 36

Behavioral Motives

Does the punishment option change the equilibrium outcome if all players are purely sel…sh? What should happen in the experiment according to the rule of reciprocation?

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

20 / 36

Behavioral Motives

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

21 / 36

Behavioral Motives

The existence of the punishment option increases the average contribution to the public good substantially. In the punishment condition, contribution rates amount to 58 percent of the endowment; in the no-punishment condition, contributions converge to zero. A subject is more heavily punished the more his or her contribution falls below the average contribution of other group members. Contributions above the average are punished much less and do not elicit a systematic punishment response. The punishment option initially causes a relative payo¤ loss. Yet, toward the end there is a relative payo¤ gain. In the …nal period, the relative payo¤ gain is roughly 10 percent.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

22 / 36

Behavioral Motives

Given the strength of the rule of reciprocity, it is no surprise that it can be used to in‡uence individuals’ decisions.3 One pro…table strategy is to give something before asking for a return favor (“foot-in-the-door technique”). An individual will often agree to a request for a substantially larger favor than the one she received in the beginning. Another tactic is the “rejection-then-retreat technique” (or “foot-in-the-face technique”): Start with an extreme request that is sure to be rejected; then retreat to a smaller request (which is the desired one). This request then seems like a concession. The rule of reciprocation implies that it is more likely to be accepted as compared to the case without initial large request.

3Several successful tactics are explained in Cialdini, Robert (2009): In‡uence –

Science and Practice, Pearson, 5th Edition, Chapter 2 (“Reciprocation: The Old Give and Take... and Take”).

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

23 / 36

Robustness

How malleable are social preferences? In the following, we consider two studies which show that pro-social preferences are not as robust as the simple experiments seem to suggest. Speci…cally, some pro-social actions may be due to image concerns. Whether an individual acts pro-socially or not then depends on her

• ptions to maintain a positive image.

We …rst consider the study by Dana et al. (2007).4

4Dana, Jason, Roberto Weber, and Jason Xi Kuang (2007): “Exploiting moral wiggle

room: experiments demonstrating an illusory preference for fairness,” Economic Theory 33, 67 - 80.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

24 / 36

Robustness

The baseline treatment is a binary dictator game. The dictator chooses between two options, A and B: option A implies 6 USD for the dictator and 1 USD for the receiver; option B implies 5 USD for the both the dictator and the receiver. In the hidden information treatment, the dictator is allowed to remain ignorant about the consequences of her action; again, she receives 6 USD for option A and 5 USD for option B. The receiver’s payo¤ from

• ption A and B is determined randomly: it is 1 and 5 USD,

respectively (as in the baseline treatment) or “‡ipped” (5 and 1 USD, respectively) so that choosing A made both parties better o¤. The dictator can reveal the payo¤s by clicking a button (i.e. it is essentially costless for her to have full information about her payo¤s).

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

25 / 36

Robustness

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

26 / 36

Robustness

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

27 / 36

Robustness

What should happen if all subjects have stable distributional preferences? What may happen in the experiment? Why?

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

28 / 36

Robustness

In the baseline treatment, 74% of the subjects chose option B (the fair option). In the hidden information treatment, only 27% of the subjects who faced the same payo¤s as in the baseline treatment chose option B (the di¤erence between the two treatments is statistically signi…cant). Only 56% chose to reveal the true payo¤s. Many subjects seem to have a desire to remain ignorant to the consequences of a self-interested choice. By remaining ignorant about the consequences

• f actions, individuals act self-interestedly, while maintaining the

illusion of fairness. Of those who chose to remain ignorant, 85% chose option A.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

29 / 36

Robustness

In the United States, 90% of individuals donate money to a charity

• rganization each year.

However, it is not clear what is the motivation for charity donations. First, pure or warm glow altruism may cause donations. In this case, donations unambiguously increase the giver’s utility. Second, individuals may give, because the solicitor exerts social pressure to give (recall the behavioral motivation social esteem). In this case, donations may decrease the giver’s utility. DellaVigna et al. (2012) …nd evidence for both motivations.5

5DellaVigna, Stefano, John A. List, and Ulrike Malmendier (2012): “Testing for

Altruism and Social Pressure in Charitable Giving,” Quarterly Journal of Economics 127(1), 1 - 56.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

30 / 36

Robustness

The …eld experiment conducted by DellaVigna et al. (2012) is about a door-to-door fundraising drive for two charities: a local children’s hospital, and an out-of-state charity organization (which is relatively unknown). There are around 8000 households in the sample. In the baseline treatment, individuals are not informed about the solicitor. In the …rst alternative treatment, a ‡yer on the doorknob noti…es households one day in advance about the one-hour time interval in which a solicitor will arrive at their homes the next day. In the second alternative treatment, the ‡yer also includes a box to be checked if the household does not want to be disturbed.

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

31 / 36

Robustness

If altruism the main behavioral motivation for giving, the ‡yer should increase both the presence at home and giving (explain why!). If social pressure is the main driver of giving, the ‡yer should lower both the frequency of opening the door and the frequency of giving (explain why!).

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

32 / 36

Robustness

Figure: Frequency of answering the door

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

33 / 36

Robustness

Figure: Frequency of unconditional giving

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

34 / 36

Robustness

Figure: Frequency of giving conditional on answering the door

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

35 / 36

Robustness

The ‡yer decreases the probability of opening the door by 9%; for the ‡yer with opt-out option the e¤ect is even 23%. The ‡yer does not decrease giving (How can we interpret this e¤ect?). However, the ‡yer with opt-out option decreases giving signi…cantly, by 28%. The decrease in giving in the opt-out treatment is driven by small donations up to 10 USD. What do we learn from these observations? What could be the implications of these …ndings for charity organizations?

• Prof. Dr. Heiner Schumacher (KU Leuven)

Incentives and Behavior

• 8. Social Preferences

36 / 36