The Knobe Effect in Experimental Economics

Kyoto Sangyo University, Kyoto , Japan, 1st October, 2011

Yan Zhou (Kyoto Sangyo University)Sobei H. Oda (Kyoto Sangyo University)

Construction

1. Philosophical Experiment by Knobe (2003) 2. Economic Experiment of the Knobe Effect by Utikal and

Fischbacher (2009)3. Our Experiment4. summary

Harm Scenario The CEO of a company is sitting in his office when

his Vice President of R&D comes in and says: “We are thinking of starting a new program. It will help us increase profits, but it will also harm the environment.”

The CEO responds that he doesn’t care about harming the environment and just wants to make as much profit as possible. The program is carried out, profits are made and the environment is harmed.

Question: Did the CEO intentionally harm the environment?

HELP ScenarioThe CEO of a company is sitting in his office when his Vice President of R&D comes in and says: “We are thinking of starting a new program. It will help us increase profits, and it will also help the environment.” The CEO responds that he doesn’t care about helping the environment and just wants to make as much profit as possible. The program is carried out, profits are made and the environment is helped.Question: Did the CEO intentionally help the environment?

The Knobe EffectExperimental Results: Those who said the CEO did intentionally is

82 per cent for the harm case and 23 per cent for the help case.

Philosophical Implication: The notion of intentionality can depend on moral judgement; people tend to think that negative side effects are bought about intentionally by the decision-maker while positives ones are not.

Social Implication: You will be criticised for harming others intentionally if you do something good for you but bad for others, while you may not be praised for helping others intentionally if you do something good both for you and others.

PositiveNegative

Intended Not Intended

PunishmentWorthy

Not RewardWorthy

Side Effect

Third Party’s Judgement

International Conference

How and why economists and philosophers do experiments:

dialogue between experimental economics

and experimental philosophyKyoto Sangyo University, Kyoto,

Japan27-28 March 2010

http://www.cc.kyoto-su.ac.jp/project/orc/execo/conf2010/index2010.html

Utikal and Fischbacher’s three-person two-stage game (2009)

1. Player 1 can move from X to Y, which increases his/her profit while generating (positive/negative) side-effect on Player 2. Question to Player 1: Do you move to Y?

2. Player 3 redistributes income between Players 1 and 2. Question to Player 3: How do you want to redistribute income if Player 1 moves (does not move) to Y?

Main effect: Y1-X1>0

Side effect: Y2-X2 >0 Y2-X2 =0 Y2-X2 <0

Scenarios (U-F experiment)+1

-2

+1

0

+1

+2

+1

-2

+1

0

+1

+1

+1

+4

U-F Working Hypothesis : 0 ≤ reward < punishment

6000

3000

4500

4500By chance

5000

5000

5000

1000

Side effect-2000

Side effect+2000

4500 –punishment

4500 +punishment

4500 +reward

4500 - reward

Chance Move Player 1’s Choice Player 3’s Choice

U-F Working Hypothesis : 0 ≤ reward < punishment

6000

3000

4500

4500By chance

5000

5000

5000

1000

Side effect-2000

Side effect+2000

4500 –punishment

4500 +punishment

4500 +reward

4500 - reward

Chance Move Player 1’s Choice Player 3’s Choice

Not observed in their experiment : denial of the Knobe Effect 。

Our experiment

Introduction of the Keynesian beauty contest

Opinion questions: What do you want to do (as a fair non-interested third party)?

Guess questions: What do you think will be the median opinion? (Monetary rewards will be shared among those who will have given the median answer.)

-2000

+1000

-1000

+1000

0

+1000

+1000

+1000

+2000

+1000

Scenarios (Our experiment)

Results of Our ExperimentPlayer 3’s Answers to Guess QuestionsRedistribution of unequally distributed income

6000

3000

4500 + α

4500 - α

4500 - β

4500 + β

55.5%

33.8%

10.7%

4500

4500

<

<

<Player 1’s Income > Player 2’s Income

Player 1’s Income=Player 2’s Income

Player 1’s Income <Player 2’s Income

Player 1’s Income >Player 2’s Income

(39.3%)

(50.4%)

(10.4%)

(Opinion)

Results of Our ExperimentPlayer 3’s Answers to Guess QuestionsRedistribution of unequally distributed income

3000

6000

4500 - α

4500 + α

4500 + β

4500 - β

8.0%

26.6%

65.4%

4500

4500

<

<

<

Player 1’s Income <Player 2’s Income

Player 1’s Income =Player 2’s Income

Player 1’s Income >Player 2’s Income

Player 1’s Income <Player 2’s Income

(Opinion)

(9.8%)

(50.1%)

(40.1%)

Results of Our ExperimentPlayer 3’s Answers to Guess Questions

Redistribution of equally distributed income

4500

4500

4500 + α

4500 - α

4500 - β

4500 + β

10.2%

57.7%

32.1%<

<4500

4500

Player 1’s Income >Player 2’s Income

Player 1’s Income =Player 2’s Income

Player 1’s Income <Player 2’s Income

Player 1’s Income =Player 2’s Income

(Opinion)

(7.9%)

(78.6%)

(13.5%)

Redistribution of unequally distributed incomePlayer 3’s Answers to Opinion & Guess Questions

Player I’s Income > Player J’s Income

Player I’s Income > Player J’s Income

Player I’s Income = Player J’s Income

Player I’s Income < Player J’s Income

Redistribution of equally distributed incomePlayer 3’s Answers to Opinion & Guess Questions

Player I’s Income = Player J’s Income

Player I’s Income < Player J’s Income

Player I’s Income = Player J’s Income

Player I’s Income > Player J’s Income

SummaryA: Inequality Aversion: unequal distribution must be equalised (no matter

what is the reason for the present inequality).

B: Inequality-reversing Aversion. The existing inequality must be respected (no matter what is the reason for the present inequality). Inequality should be reduced, but must not be reversed.

C: The Knobe Effect. Who is responsible for the present distribution matters.

A>B in the answers to Opinion questions, while B<A in the answers to Guess questions.

C is the third principle in the answers to either question, but stronger in the answers to guess questions.

Subjects may have not seriously considered opinion questions, which they knew would not affect their income: they may have saved time to concentrate their attention to guess questions, which they knew would affect their income.

If subjects answered both opinion and guess questions seriously, they followed a simple rule  ( inequality aversion ), suspecting that others would take account of other factors ( inequality-reversing aversion).

Thank you for your attention ！

Player 1’s Answer to Opinion Questions Side Effect Aversion

-2000 -1000 0 1000 20000

25

50

75

100

1 2 3

[6000,3000]

[4500,4500]

[3000,6000]

0

25

50

75

100

total (average [%])

Side effects

[Player 1’s income, Player 2’s income]

Percentage of moving to X(to increase his/her own income)

Player 1’s Answer to Guess Questions Side Effect Aversion

-2000 -1000 0 1000 20000

25

50

75

100

1 2 3

Side effects [Player 1’s income, Player 2’s income]

Percentage of moving to X(to increase his/her own income)

[6000,3000]

[4500,4500]

[3000,6000]

0

25

50

75

100

total (average [%])

Table 7

Table 8

Table 9

Table 10