Sorting Into Incentives for Prosocial Behavior Christian J. Meyer - - PowerPoint PPT Presentation

Sorting Into Incentives for Prosocial Behavior

Christian J. Meyer & Egon Tripodi

European University Institute, Department of Economics

June 29, 2018 2018 ESA World Meeting, Berlin

Introduction Theoretical Framework Experimental Design Results Conclusion

‘The Twenty-eight World Health Assembly, [...] urges Member States to promote the development of national blood services based on voluntary unremunerated donation of blood.”

— World Health Assembly Resolution 28.72 (1975)

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 2

Introduction Theoretical Framework Experimental Design Results Conclusion

‘The Twenty-eight World Health Assembly, [...] urges Member States to promote the development of national blood services based on voluntary unremunerated donation of blood.”

— World Health Assembly Resolution 28.72 (1975)

◮ Concerns for the quantity and quality of blood supply led to

unremunerated donations being the policy objective of the World Health Organization (WHO), the Red Cross, and the Council of Europe

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 2

Introduction Theoretical Framework Experimental Design Results Conclusion

‘The Twenty-eight World Health Assembly, [...] urges Member States to promote the development of national blood services based on voluntary unremunerated donation of blood.”

— World Health Assembly Resolution 28.72 (1975)

◮ Concerns for the quantity and quality of blood supply led to

unremunerated donations being the policy objective of the World Health Organization (WHO), the Red Cross, and the Council of Europe

◮ 83% of 88 million whole blood donations globally are reported as

voluntary and non-remunerated (WHO, 2017)

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 2

Introduction Theoretical Framework Experimental Design Results Conclusion

Blood Collection in Germany

Un Unpaid Pa Paid

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 3

Introduction Theoretical Framework Experimental Design Results Conclusion

Blood Collection in Germany

More info

Un Unpaid Pa Paid

Germany’s market for whole blood has desirable properties

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 3

Introduction Theoretical Framework Experimental Design Results Conclusion

Blood Collection in Germany

More info

Un Unpaid Pa Paid

Germany’s market for whole blood has desirable properties

◮ over 70% of unpaid donations

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 3

Introduction Theoretical Framework Experimental Design Results Conclusion

Blood Collection in Germany

More info

Un Unpaid Pa Paid

Germany’s market for whole blood has desirable properties

◮ over 70% of unpaid donations ◮ largest world-wide supply of blood per capita (WHO 2017)

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 3

Introduction Theoretical Framework Experimental Design Results Conclusion

Blood Collection in Germany

More info

Un Unpaid Pa Paid

Germany’s market for whole blood has desirable properties

◮ over 70% of unpaid donations ◮ largest world-wide supply of blood per capita (WHO 2017) ◮ $110 wholesale price of a blood unit, compares to $190 in Sweden and

Switzerland, and $211 in the US

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 3

Introduction Theoretical Framework Experimental Design Results Conclusion

Blood Collection in Germany

More info

Un Unpaid Pa Paid

Germany’s market for whole blood has desirable properties

◮ over 70% of unpaid donations ◮ largest world-wide supply of blood per capita (WHO 2017) ◮ $110 wholesale price of a blood unit, compares to $190 in Sweden and

Switzerland, and $211 in the US

◮ some of the safest blood supplies: e.g. low incidence of Hep C and HIV

Other dual collection systems Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 3

Introduction Theoretical Framework Experimental Design Results Conclusion

We Study Dual Markets for Charitable Giving

◮ Can a “dual market” increase the supply of charitable donations?

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 4

Introduction Theoretical Framework Experimental Design Results Conclusion

We Study Dual Markets for Charitable Giving

◮ Can a “dual market” increase the supply of charitable donations?

◮ relative to a single market where no one is paid Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 4

Introduction Theoretical Framework Experimental Design Results Conclusion

We Study Dual Markets for Charitable Giving

◮ Can a “dual market” increase the supply of charitable donations?

◮ relative to a single market where no one is paid ◮ relative to a single market where everyone is paid Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 4

Introduction Theoretical Framework Experimental Design Results Conclusion

We Study Dual Markets for Charitable Giving

◮ Can a “dual market” increase the supply of charitable donations?

◮ relative to a single market where no one is paid ◮ relative to a single market where everyone is paid

◮ What economic mechanisms make a “dual” collection system potentially

more efficient?

◮ Starting point: We hypothesize that it enables donors to sort ◮ Our focus: Image concerns Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 4

Introduction Theoretical Framework Experimental Design Results Conclusion

We Study Dual Markets for Charitable Giving

◮ Can a “dual market” increase the supply of charitable donations?

◮ relative to a single market where no one is paid ◮ relative to a single market where everyone is paid

◮ What economic mechanisms make a “dual” collection system potentially

more efficient?

◮ Starting point: We hypothesize that it enables donors to sort ◮ Our focus: Image concerns

◮ We study these questions in a controlled environment in two steps:

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 4

Introduction Theoretical Framework Experimental Design Results Conclusion

We Study Dual Markets for Charitable Giving

◮ Can a “dual market” increase the supply of charitable donations?

◮ relative to a single market where no one is paid ◮ relative to a single market where everyone is paid

◮ What economic mechanisms make a “dual” collection system potentially

more efficient?

◮ Starting point: We hypothesize that it enables donors to sort ◮ Our focus: Image concerns

◮ We study these questions in a controlled environment in two steps:

◮ We use a model of incentivized charitable giving building on

Benabou and Tirole 2006

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 4

Introduction Theoretical Framework Experimental Design Results Conclusion

We Study Dual Markets for Charitable Giving

◮ Can a “dual market” increase the supply of charitable donations?

◮ relative to a single market where no one is paid ◮ relative to a single market where everyone is paid

◮ What economic mechanisms make a “dual” collection system potentially

more efficient?

◮ Starting point: We hypothesize that it enables donors to sort ◮ Our focus: Image concerns

◮ We study these questions in a controlled environment in two steps:

◮ We use a model of incentivized charitable giving building on

Benabou and Tirole 2006

◮ We test these predictions in the laboratory Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 4

Introduction Theoretical Framework Experimental Design Results Conclusion

Our Model Produces Three Testable Predictions

Model overview

Hypothesis 1 (Incentive Effect)

Irrespective of whether actions are visible, making monetary incentives available increases donations.

Hypothesis 2 (Social Image Effect)

Irrespective of whether monetary incentives are available, making actions visible increases donations.

Hypothesis 3 (Sorting)

In a dual market, a positive share of agents chooses to be not paid. This share is larger when actions are taken in public.

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 5

Introduction Theoretical Framework Experimental Design Results Conclusion

Our Model Produces Three Testable Predictions

Model overview

Hypothesis 1 (Incentive Effect)

Irrespective of whether actions are visible, making monetary incentives available increases donations.

Hypothesis 2 (Social Image Effect)

Irrespective of whether monetary incentives are available, making actions visible increases donations.

Hypothesis 3 (Sorting)

In a dual market, a positive share of agents chooses to be not paid. This share is larger when actions are taken in public. Donations across market designs Share donating not paid Private action Not paid < Paid = Dual Market s > 0 Public action Not paid < Paid = Dual Market S > s

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 5

Introduction Theoretical Framework Experimental Design Results Conclusion

Design of Laboratory Experiment

◮ Subjects generate a donation by entering a sequence of keys on

computer keyboard 400 times

Screenshots ◮ Monetary incentives for subjects to donate decrease social value, net

amount is given to charity chosen by subject

◮ 3 rounds of task, decision to participate in each, outside option for skipping Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 6

Introduction Theoretical Framework Experimental Design Results Conclusion

Design of Laboratory Experiment

◮ Subjects generate a donation by entering a sequence of keys on

computer keyboard 400 times

Screenshots ◮ Monetary incentives for subjects to donate decrease social value, net

amount is given to charity chosen by subject

◮ 3 rounds of task, decision to participate in each, outside option for skipping

◮ Full 3 × 2 between-subject design

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 6

Introduction Theoretical Framework Experimental Design Results Conclusion

Design of Laboratory Experiment

◮ Subjects generate a donation by entering a sequence of keys on

computer keyboard 400 times

Screenshots ◮ Monetary incentives for subjects to donate decrease social value, net

amount is given to charity chosen by subject

◮ 3 rounds of task, decision to participate in each, outside option for skipping

◮ Full 3 × 2 between-subject design

◮ 3 treatments represent single and dual market designs ◮ Single market not paid: Donor receives no remuneration

y = {0}

◮ Single market paid:

Donor receives remuneration y = {˜ y}

◮ Dual market:

Donors can sort into one of the two y = {0, ˜ y}

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 6

Introduction Theoretical Framework Experimental Design Results Conclusion

Design of Laboratory Experiment

◮ Subjects generate a donation by entering a sequence of keys on

computer keyboard 400 times

Screenshots ◮ Monetary incentives for subjects to donate decrease social value, net

amount is given to charity chosen by subject

◮ 3 rounds of task, decision to participate in each, outside option for skipping

◮ Full 3 × 2 between-subject design

◮ 3 treatments represent single and dual market designs ◮ Single market not paid: Donor receives no remuneration

y = {0}

◮ Single market paid:

Donor receives remuneration y = {˜ y}

◮ Dual market:

Donors can sort into one of the two y = {0, ˜ y}

◮ 2 treatments to change visibility of actions Illustration ◮ Private ◮ Public: subjects told to publicly announce their actions at the end of the

experiment by standing up in the laboratory

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 6

Introduction Theoretical Framework Experimental Design Results Conclusion

Design of Laboratory Experiment

◮ Subjects generate a donation by entering a sequence of keys on

computer keyboard 400 times

Screenshots ◮ Monetary incentives for subjects to donate decrease social value, net

amount is given to charity chosen by subject

◮ 3 rounds of task, decision to participate in each, outside option for skipping

◮ Full 3 × 2 between-subject design

◮ 3 treatments represent single and dual market designs ◮ Single market not paid: Donor receives no remuneration

y = {0}

◮ Single market paid:

Donor receives remuneration y = {˜ y}

◮ Dual market:

Donors can sort into one of the two y = {0, ˜ y}

◮ 2 treatments to change visibility of actions Illustration ◮ Private ◮ Public: subjects told to publicly announce their actions at the end of the

experiment by standing up in the laboratory

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 6

Introduction Theoretical Framework Experimental Design Results Conclusion

Design of Laboratory Experiment

◮ Subjects generate a donation by entering a sequence of keys on

computer keyboard 400 times

Screenshots ◮ Monetary incentives for subjects to donate decrease social value, net

amount is given to charity chosen by subject

◮ 3 rounds of task, decision to participate in each, outside option for skipping

◮ Full 3 × 2 between-subject design

◮ 3 treatments represent single and dual market designs ◮ Single market not paid: Donor receives no remuneration

y = {0}

◮ Single market paid:

Donor receives remuneration y = {˜ y}

◮ Dual market:

Donors can sort into one of the two y = {0, ˜ y}

◮ 2 treatments to change visibility of actions Illustration ◮ Private ◮ Public: subjects told to publicly announce their actions at the end of the

experiment by standing up in the laboratory

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 6

Introduction Theoretical Framework Experimental Design Results Conclusion

Design of Laboratory Experiment

◮ Subjects generate a donation by entering a sequence of keys on

computer keyboard 400 times

Screenshots ◮ Monetary incentives for subjects to donate decrease social value, net

amount is given to charity chosen by subject

◮ 3 rounds of task, decision to participate in each, outside option for skipping

◮ Full 3 × 2 between-subject design

◮ 3 treatments represent single and dual market designs ◮ Single market not paid: Donor receives no remuneration

y = {0}

◮ Single market paid:

Donor receives remuneration y = {˜ y}

◮ Dual market:

Donors can sort into one of the two y = {0, ˜ y}

◮ 2 treatments to change visibility of actions Illustration ◮ Private ◮ Public: subjects told to publicly announce their actions at the end of the

experiment by standing up in the laboratory

◮ 329 subjects from BonnEconLab, balanced on observables

Procedures Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 6

Introduction Theoretical Framework Experimental Design Results Conclusion

Results (1/3): No Evidence for Crowding Out

Hp 1: Irrespective of whether actions are visible, making monetary incentives available increases donations

1.51 1.85 1.94

1 1.5 2 2.5 Average participation in the donation task

• ver all three rounds

Not paid Not paid Not paid

Public

1.13 1.31 1.52

1 1.5 2 2.5

• ver all three rounds

Not paid Not paid Not paid

Private

Not paid Paid Choose Not paid Paid Choose

Subject participation in the donation task (number of rounds, averaged over all three rounds)

Private Public

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 7

Introduction Theoretical Framework Experimental Design Results Conclusion

Results (1/3): No Evidence for Crowding Out

Hp 1: Irrespective of whether actions are visible, making monetary incentives available increases donations

1.51 1.85 1.94

1 1.5 2 2.5 Average participation in the donation task

• ver all three rounds

Not paid Not paid Not paid

Public

1.13 1.31 1.52

1 1.5 2 2.5

• ver all three rounds

Not paid Not paid Not paid

Private

rank-sum p-val. 0.048 (N = 106) rank-sum p-val. 0.29 (N = 108) rank-sum p-val. 0.016 (N = 113) rank-sum p-val. 0.545 (N = 128)

Subject participation in the donation task (number of rounds, averaged over all three rounds)

Private Public

Not paid Paid Choose Not paid Paid Choose

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 7

Introduction Theoretical Framework Experimental Design Results Conclusion

Results (2/3): Social image effects across all market designs

Hp 2: Irrespective of availability of compensation, making actions visible increases donations.

1.13 1.51 1.31 1.85 1.52 1.94

1 1.5 2 2.5

• ver all three rounds

Not paid Paid Choose

Subject participation in the donation task, by visibility (number of rounds, averaged over all three rounds)

Public Private

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 8

Introduction Theoretical Framework Experimental Design Results Conclusion

Results (2/3): Social image effects across all market designs

Hp 2: Irrespective of availability of compensation, making actions visible increases donations.

1.13 1.51 1.31 1.85 1.52 1.94

1 1.5 2 2.5

• ver all three rounds

Not paid Paid Choose

Subject participation in the donation task, by visibility (number of rounds, averaged over all three rounds)

Public

rank-sum p-val. 0.025 (N = 93) rank-sum p-val. 0.004 (N = 110) rank-sum p-val. 0.025 (N = 126)

Private

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 8

Introduction Theoretical Framework Experimental Design Results Conclusion

Results (3/3): Sorting Into Unpaid

Hp 3: In a dual market, (a) a positive fraction of donors chooses to be not paid and (b) this fraction is bigger when actions are taken in public .1 .2 .3 .4 .5 Round 1 Round 2 Round 3 Public actions Private actions Share of subjects participating not paid

◮ Support for hypothesis 3 (a)

◮ A positive share of donors

always chooses to be not paid.

◮ No definitive support for

hypothesis 3 (b)

◮ No significant image effects

(rank-sum test p-val 0.353, N=126).

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 9

Introduction Theoretical Framework Experimental Design Results Conclusion

Summary and Conclusion

◮ We find empirical support for theoretical predictions ◮ Key insight: Letting people choose incentives to donate lets us leverage

heterogeneous social preferences to the advantage of the collection system

◮ Some altruistic and some image-concerned people choose unpaid donations ◮ Image effects improve efficiency of collection

◮ In the paper we discuss policy implications for the blood market

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 10

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Index of Appendices

• 6. Case Study: Blood Donations
• 7. Appendix: Background
• 8. Appendix: Theory
• 9. Appendix: Experimental Design
• 10. Appendix: Lab Results

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 12

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

The Economic Problem of Eliciting Blood Donations

◮ Supply of human blood is a public good provision problem

◮ About 90 million blood donations collected globally each year ◮ Voluntary provision often not enough ◮ Cannot (yet) be generated artificially ◮ Increasing demand: demography, medical technology Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 13

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

The Economic Problem of Eliciting Blood Donations

◮ Supply of human blood is a public good provision problem

◮ About 90 million blood donations collected globally each year ◮ Voluntary provision often not enough ◮ Cannot (yet) be generated artificially ◮ Increasing demand: demography, medical technology

◮ Can we use monetary compensation to generate donations?

◮ Long literature in social sciences going back to Titmuss (1971) ◮ Heavy-handed regulation since 1980s Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 13

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

The Economic Problem of Eliciting Blood Donations

◮ Supply of human blood is a public good provision problem

◮ About 90 million blood donations collected globally each year ◮ Voluntary provision often not enough ◮ Cannot (yet) be generated artificially ◮ Increasing demand: demography, medical technology

◮ Can we use monetary compensation to generate donations?

◮ Long literature in social sciences going back to Titmuss (1971) ◮ Heavy-handed regulation since 1980s ◮ May attract “risky” donors ◮ Overtaken by screening technology in high-income countries Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 13

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

The Economic Problem of Eliciting Blood Donations

◮ Supply of human blood is a public good provision problem

◮ About 90 million blood donations collected globally each year ◮ Voluntary provision often not enough ◮ Cannot (yet) be generated artificially ◮ Increasing demand: demography, medical technology

◮ Can we use monetary compensation to generate donations?

◮ Long literature in social sciences going back to Titmuss (1971) ◮ Heavy-handed regulation since 1980s ◮ May attract “risky” donors ◮ Overtaken by screening technology in high-income countries ◮ Crowding out of non-remunerated, voluntary donations ◮ Open empirical question Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 13

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Policy Maker’s Objective Is to Increase Donations

◮ Policy makers re-evaluating incentivized or paid donations

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 14

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Policy Maker’s Objective Is to Increase Donations

◮ Policy makers re-evaluating incentivized or paid donations

◮ Policy shift against backdrop of short supply and better screening Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 14

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Policy Maker’s Objective Is to Increase Donations

◮ Policy makers re-evaluating incentivized or paid donations

◮ Policy shift against backdrop of short supply and better screening ◮ Objective is to increase supply, even using costly measures Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 14

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Policy Maker’s Objective Is to Increase Donations

◮ Policy makers re-evaluating incentivized or paid donations

◮ Policy shift against backdrop of short supply and better screening ◮ Objective is to increase supply, even using costly measures

◮ On national wholesale markets for blood, single unpaid provider (Red

Cross) appears to command large price premium

$211 $50 $68

50 100 150 200

Cost of acquistion Cost of inhouse production 2007 US$

logistics and admin mandated tests

US data for 2007, based on Toner et al. (2011) ◮ Scope for costly incentives Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 14

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Germany Stands Out for Its Competitive Market

⊳ Intro

◮ Dual market: Paid and Not paid

Details ◮ Payment of up to US$ 30 ◮ Largely paid: Hospitals and commercial providers (29% of market) ◮ Never paid: Red Cross (71% of market) Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 15

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Germany Stands Out for Its Competitive Market

⊳ Intro

◮ Dual market: Paid and Not paid

Details ◮ Payment of up to US$ 30 ◮ Largely paid: Hospitals and commercial providers (29% of market) ◮ Never paid: Red Cross (71% of market)

◮ Competition in urban centers

◮ 69% of population can reach paid donation within 30 min drive ◮ Similar travel time: 7.2min to not paid, 11.7min to paid Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 15

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Germany Stands Out for Its Competitive Market

⊳ Intro

◮ Dual market: Paid and Not paid

Details ◮ Payment of up to US$ 30 ◮ Largely paid: Hospitals and commercial providers (29% of market) ◮ Never paid: Red Cross (71% of market)

◮ Competition in urban centers

◮ 69% of population can reach paid donation within 30 min drive ◮ Similar travel time: 7.2min to not paid, 11.7min to paid

◮ High safety

◮ Example: Residual risk for Hepatitis C, per million donations (Offergeld et al,

2005; Stramer, 2007)

◮ 0.10 (France), 0.18 (Germany), 0.87 (United States), 2.23 (Spain) Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 15

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Germany Stands Out for Its Competitive Market

⊳ Intro

◮ Dual market: Paid and Not paid

Details ◮ Payment of up to US$ 30 ◮ Largely paid: Hospitals and commercial providers (29% of market) ◮ Never paid: Red Cross (71% of market)

◮ Competition in urban centers

◮ 69% of population can reach paid donation within 30 min drive ◮ Similar travel time: 7.2min to not paid, 11.7min to paid

◮ High safety

◮ Example: Residual risk for Hepatitis C, per million donations (Offergeld et al,

2005; Stramer, 2007)

◮ 0.10 (France), 0.18 (Germany), 0.87 (United States), 2.23 (Spain)

◮ Low cost of a standardized blood unit on wholesale market

◮ US$ 110 in Germany ◮ US$ 190 in Sweden and Switzerland (Trimborn, 2009) ◮ US$ 210 in the United States (Toner et al., 2011) Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 15

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Can the German Case Inform Market Design?

◮ Blood donations remain an important challenge for market design ◮ The German “dual market” seems to have desirable properties

◮ in line with our theoretical and experimental findings ◮ Caveat: We abstract from competition or social welfare Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 16

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Can the German Case Inform Market Design?

◮ Blood donations remain an important challenge for market design ◮ The German “dual market” seems to have desirable properties

◮ in line with our theoretical and experimental findings ◮ Caveat: We abstract from competition or social welfare

◮ What happens when introducing the option to choose in single

unpaid market (e.g. United States)?

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 16

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

German Blood Market: Institutional Environment

Market Share, by Number of Donations (2014)

Red Cross 0.71 Other 0.29 Not paid Largely paid

◮ Institutional environment makes empirical

study difficult

◮ Government data not at. donor level ◮ German Red Cross consists of

quasi-independent bodies...

◮ ... none of which want to cooperate ◮ Hospitals very willing to cooperate, but do

not have centralized data ◮ Even if we had the data, not clear that we

could identify crowding-out

◮ Spatial competition varies by region ⊳ Back

Spatial distribution of collectors, Nov ’16-Jan ’17

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 18

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Global Blood Supply, by Remuneration

◮ 1.5m of 83m (1.8%) donations reported to WHO are paid (2012)

◮ Family/replacement donations explains much of the low share of unpaid ◮ Problems with this data

Percentage of Voluntary Unpaid Blood Donations (2008)

⊳ Back Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 19

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Charity Donations

⊳ Back Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 20

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Benabou & Tirole 2006 Original Model

Utility for donor i is

Ui(d, y) = (1 − p xi) [ ai + y − c ] d + p xi [E (a|d, y )]

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 21

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Benabou & Tirole 2006 Original Model

Decision to Donate

When do agents of different image concerns donate?

◮ Image-indifferent agents (xi = 0) contribute if

U(d = 1, y) = ai + y − c ≥ 0

◮ Image-concerned agents (xi = 1) donate if

U(d = 1, y) = E(a|d = 1, y) ≥ E(a|0)

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 22

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Benabou & Tirole 2006 Original Model

Refusing Remuneration

Suppose we allow agents to refuse remuneration: y = {0, ˜ y}

◮ Strictly dominated strategy for image indifferent people (x = 0) to

choose y = 0 U(d = 1, y) = va + y − c

◮ By choosing ˜

y = 0 image concerned people (˜ x → ∞) reveal that they care about image, but don’t gain reputation U(d = 1, y) = E(va|d = 1, y = ˜ y) ≥ E(va|d = 1, y = 0)

◮ Nobody ever turns money down

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 23

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Towards a Theoretical Framework: Sorting

⊳ Back

◮ Starting point: We hypothesize that the German dual market enables

donors to sort

◮ In the seminal model of Benabou and Tirole 2006, introducing the option

to turn down compensation for giving does not affect equilibrium behavior

◮ Compensation paid from outside the model ◮ Inconsistent with sorting and thus not useful to understand dual markets

◮ We simplify their model to reconcile theory and evidence

◮ Compensation paid by the collector of donations ◮ Homogeneous taste for compensation to donate ◮ Focus on image concerns Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 24

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Our Model of Charitable Giving: Overview

◮ The model economy has unit mass of agents and 1 collector

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 25

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Our Model of Charitable Giving: Overview

◮ The model economy has unit mass of agents and 1 collector ◮ The collector remunerates agents for contributions

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 25

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Our Model of Charitable Giving: Overview

◮ The model economy has unit mass of agents and 1 collector ◮ The collector remunerates agents for contributions

◮ Market design determines availability of remuneration and whether being

remunerated is a choice variable

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 25

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Our Model of Charitable Giving: Overview

◮ The model economy has unit mass of agents and 1 collector ◮ The collector remunerates agents for contributions

◮ Market design determines availability of remuneration and whether being

remunerated is a choice variable

◮ Remuneration amount is set exogenously Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 25

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Our Model of Charitable Giving: Overview

◮ The model economy has unit mass of agents and 1 collector ◮ The collector remunerates agents for contributions

◮ Market design determines availability of remuneration and whether being

remunerated is a choice variable

◮ Remuneration amount is set exogenously

◮ Agents are motivated to make a contribution by

◮ Heterogenous intrinsic incentives (altruism) ◮ Homogenous extrinsic incentives (fringe or monetary benefits) ◮ Heterogenous image concerns (to appear altruistic) Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 25

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Model of Blood Donations

Decision to Donate

When do agents of different image concerns contribute?

◮ Donating in public (p = 1):

◮ Image-indifferent agents (xi = 0) contribute if

Ui(d = 1, y, xi = 0) = aiB + (1 − ai)y − c ≥ 0

◮ Image-concerned agents (xi = 1) contribute if

Ui(d = 1, y, xi = 1) = E(a|d = 1, y) ≥ E(a|d = 0)

◮ Donating in private (p = 0):

◮ All agents behave as if they were image-indifferent Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 26

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Model of Blood Donations

Refusing Remuneration

Suppose we allow agents to refuse remuneration: y = {0, ˜ y} ֒ → equivalent to dual market à la Germany

◮ Donating in public (p = 1):

◮ Among image-indifferent (xi = 0), the most altruistic (ai > 1) always

participate and return the money Ui(d = 1, y, xi = 0) = aiB + (1 − ai)y − c

◮ Image-concerned agents (xi = 1) can now improve their reputation by

choosing y = 0, and will always participate Ui(d = 1, y = 0, xi = 1) = E(a|d = 1, y = 0) > E(a|d = 1, y = ˜ y) > E(a|d = 0)

◮ Donating in private (p = 0):

◮ As before, all agents behave as if they were image-indifferent Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 27

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Experimental Design for MTurk

⊳ Back

Subject pool Elicit altruism Not paid Not paid Not paid

Never paid

Choose Choose

Not paid 1st R

Paid Choose Choose

Paid 1st R

Choose

Round 1

Choose

Round 2

Choose

Round 3 Always choose

Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 28

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Experimental Design

Subject pool Elicit altruism Public action Choose Paid Not paid Private action Choose Paid Not paid

⊳ Back Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 29

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Sample Overview and Experimental Procedures

⊳ Back

◮ 15 lab sessions in May 2017 at the Bonn Econ Lab

◮ 329 subjects ◮ Subjects are on average 22 years old, 60 percent women ◮ Computerized experiment

◮ Random assignment to treatments

◮ Groups are balanced on observables Balance table

◮ Payments

◮ Average payment to subject 11.70 euro ◮ Average donation to charity 4.40 euro ◮ Subjects receive proof of donation to charity Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 30

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Screenshot of Experiment: Explanation of Task

⊳ Back Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 31

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Screenshot of Experiment: Participation Choice

⊳ Back Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 32

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Screenshot of Experiment: Donation Task

⊳ Back Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 33

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Summary Statistics and Balance

⊳ Back Full Sample Private Public p- value Not paid Paid Choose Not paid Paid Choose (1) (2) (3) (4) (5) (6) (7) (8) a) Measured before treatment DG: Tokens kept 15.365 14.891 15.271 15.250 15.021 15.677 15.818 0.848 (0.214) (0.621) (0.558) (0.507) (0.618) (0.501) (0.411) Charity rating 4.602 4.783 4.604 4.583 4.660 4.532 4.515 0.131 (0.043) (0.087) (0.129) (0.072) (0.102) (0.123) (0.100) b) Socioeconomic characteristics, measured after treatment Age 21.544 21.630 21.708 21.717 21.511 21.210 21.545 0.499 (0.091) (0.263) (0.223) (0.213) (0.263) (0.184) (0.207) Female 0.611 0.630 0.521 0.717 0.574 0.613 0.591 0.429 (0.027) (0.072) (0.073) (0.059) (0.073) (0.062) (0.061) College major 4.398 4.239 4.417 4.400 4.383 4.661 4.258 0.814 (0.100) (0.277) (0.258) (0.224) (0.273) (0.236) (0.221) Observations 329 46 48 60 47 62 66 ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 Notes: Standard errors are clustered at session level. p-value in column (8) is for a one-way ANOVA on ranks (Kruskal- Wallis) test comparing the six treatment groups in columns (2) to (7). In the dictator game we gave 20 experimental tokens to participants and asked them how many they would like to keep. Charity rating refers to the rating that subjects gave to the charity that they chose to donate to. We asked subjects to agree to the statement “I like the idea of donating money to [chosen charity]” on a 5-point likert scale where 1 is “strongly disagree” and 5 is “strongly agree”. College major is a categorical variable that summarizes the departmental affiliation of our student subjects. Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 34

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Gender Heterogeneity in Image effects

⊳ Back

Linear predictions from OLS regressions of total number of rounds participated on visibility treatment × gender dummy, plus controls. Separate regressions for each treatment. .5 1 1.5 2 2.5 3 Linear prediction of rounds participated Private Public Visibility Female Male

Not paid

.5 1 1.5 2 2.5 3 Linear prediction of rounds participated Private Public Visibility Female Male

Paid

.5 1 1.5 2 2.5 3 Linear prediction of rounds participated Private Public Visibility Female Male

Choose

Table underlying OLS regressions Table ranksum test results Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 35

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Gender Differences in Costly Signaling

⊳ Back

Rank-sum test results

Treatment Not paid Paid Choose (1) (2) (3) a) Private vs. public, whole sample z-score

• 2.247
• 2.872
• 2.249

p-value 0.025 0.004 0.025 Observations 93 110 126 b) Private vs. public, only female z-score

• 1.581
• 2.321
• 0.981

p-value 0.114 0.020 0.327 Observations 56 63 82 c) Private vs. public, only male z-score

• 1.688
• 1.402
• 2.236

p-value 0.092 0.161 0.025 Observations 36 47 42 Notes: Test statistics are for the total number of rounds participated in the donation task based on a two-sample Wilcoxon rank-sum (Mann-Whitney) test of PRIVATE vs. PUBLIC treatments, adminis- tered separately in each incentive treatment. Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 36

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Parametric Analysis I

OLS for Individual Participation in the Donation Task over Three Rounds

Dependent variable: # of donations over the three rounds (1) (2) (3) (4) Paid 0.182 0.207 0.219 0.225 (0.224) (0.220) (0.216) (0.217) Choose 0.386∗∗ 0.367∗∗ 0.421∗∗∗ 0.416∗∗∗ (0.141) (0.129) (0.113) (0.107) Public 0.380∗∗ 0.393∗∗ 0.393∗∗ 0.396∗∗ (0.155) (0.155) (0.150) (0.150) Paid × Public 0.162 0.129 0.189 0.180 (0.293) (0.291) (0.276) (0.280) Choose × Public 0.043 0.058 0.086 0.089 (0.173) (0.162) (0.151) (0.145) Control variables Female 0.224 0.059 (0.129) (0.113) Dictator Game: Tokens kept

• 0.098∗∗∗
• 0.097∗∗∗

(0.014) (0.013) Observations 329 329 329 329 R2 0.080 0.091 0.222 0.223 ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 Notes: Standard errors are clustered at session level. Not Paid is the base market design

• treatment. Private is the base visibility treatment.

⊳ Back Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 37

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Parametric Analysis II

Average Marginal Effects and Relative Risk Ratios from Random Effects Regressions

Participation Augmented Choice (1) (2) Paid 0.426∗∗ (0.216) Choose 0.654∗∗∗ (0.129) Public 0.719∗∗∗ 1.773 (0.183) (1.257) Observations 987 378 ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 Notes: Standard errors are clustered at session level. Not Paid is the base market design treatment. Private is the base visibility treatment. In (1) the dependent variable indi- cating participation to the donation task is observed in any treatment for all 329 subjects over three rounds. For column (2) the augmented choice dependent variable only applies to the 129 subjects in treatments Choose-Private and Choose- Public over three rounds. Augmented choice takes value "0" if subject skips, "1" if participates unpaid, and "2" if participates paid to the donation task in a given round. Column (1) reports average marginal effects. Column (2) reports the relative risk ratio for outcome "1" unpaid participation and base outcome "2" paid participation. The interaction effect between market design treatments and visibility treatments are taken into ac- count in the average marginal effects presented in (1) and irrelevant for the dependent variable that only applies to the Choose treatments in (2). Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 38

Case Study: Blood Donations Appendix: Background Appendix: Theory Appendix: Experimental Design Appendix: Lab Results

Parametric Analysis III: Gender Differences in Costly Signaling

⊳ Back

OLS for Individual Participation in the Donation Task over Three Rounds

Treatment subsamples Not paid Paid Choose (1) (2) (3) Public 0.578* 0.478 1.077** (-0.264) (-0.291) (-0.295) Female 0.245 0.0561 0.456 (-0.229) (-0.229) (-0.225) Public × Female

• 0.306

0.154

• 0.876*

(-0.302) (-0.291) (-0.405) Control variables Dictator Game: Tokens kept

• 0.0668*
• 0.0874**
• 0.141***

(-0.0226) (-0.0242) (-0.025) Observations 93 110 126 R2 0.157 0.205 0.292 ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 Notes: Standard errors are clustered at session level. Not Paid is the base market design treatment. Private is the base visibility treatment. Meyer and Tripodi (EUI): Sorting Into Incentives for Prosocial Behavior 39