Giving to the Less Fortunate: A Real-Effort Experiment

1
Giving to the Less Fortunate A Real-Effort
Experiment

• Nisvan Erkal, Lata Gangadharan
• and Nikos Nikiforakis
• University of Melbourne
• ESA, Rome, June 2007

2
Research Goals

• Goal 1 How does real effort affect giving
  behaviour?
• Goal 2 To what extent does inequality aversion
  depend on the source of the inequality?
• Goal 3 If people do not know the determinants of
  income, are they less likely to give (in general
  to anybody)?

3
Design Stage 1

• Two-stage game.
• Groups of four.
• Stage 1 Participants encode words for 20
  minutes.
• Simple task that does not require talent or
  intelligence.

4
(No Transcript)
5
Design Luck

• In addition to effort, also luck may affect
  income at the end of Stage 1 in some treatments.
• Virtual coin toss if Tails, reduce the number
  of words encoded by 30.

6
Design Stage 2

• Stage 2 Participants are ranked based on points
• Then, they (simultaneously) decide how much money
  to transfer to the others in their group.
• The transfers can range from 0 to all of the
  income they made in Stage 1.
• For each participant, we randomly choose to
  implement one of the suggested transfers.
• One-shot.

7
Design - Earnings

• Earnings are determined in a competitive way.
• 1st ranked 60
• 2nd ranked 45
• 3rd ranked 30
• 4th ranked 15

8
Design - Treatments

• Treatment 1 (Full information, no luck)
• Treatment 2 (Full information, luck)
• Treatment 3 (Incomplete information, luck)
• 156 subjects, all Australian, students of
  University of Melbourne

9
Who receives money?

• Result 1 Monotonic relation between receivers
  rank and likelihood of receiving a transfer

10
Who sends the money?

• Result 2 2nd ranked subjects are more likely to
  transfer than those ranked 1st.

11
Impact of Luck on Receiver

• In T2 (full information, luck), those with bad
  luck received more.

12
Impact of Luck on Sender

• In T2, those with good luck sent more.

13
Impact of Incomplete Information

• On Transfer Behavior

14
Impact of Incomplete Information

• On senders behavior

15
Number of Donors

• The number of donors was higher in T3.

16
Result 5 - Effort

• Number of words encoded was significantly higher
  in T2. (OLS, control for gender, p-value lt 0.01)
• No significant differences between T1T3 or T2T3

17

• The End

18
(No Transcript)