The Economics of Faith: Using an Apocalyptic Prophecy to Elicit Religious Beliefs in the Field

Transcription

1 The Economics of Faith: Using an Apocalyptic Prophecy to Elicit Religious Beliefs in the Field Ned Augenblick Haas School of Business Ernesto Dal Bo Haas School of Business Jesse Cunha Naval Post-graduate Academy Justin M. Rao Yahoo! Research

2 beliefs about the supernatural

3 people

4 who believe(d) in a specific date of Christ s return

5 a belief that is tempting to dismiss as crazy

6 question:

7 can we learn anything about humans form beliefs more generally?

8 are they mentally unsound?

9 or maybe

10 these people are normal

11 like you and I

12 but simply hold beliefs we don t agree with?

13 if so then

14 we can learn about how and why people in general:

15 we can learn about how and why people in general: 1) form extreme beliefs

16 we can learn about how and why people in general: 1) form extreme beliefs 2) adjust to evidence against them

17 we argue they are normal

18 all humans are a bit strange

19 beliefs based on insufficient evidence are hard to rationalize

20 (within traditional economics)

21 prefs over outcomes

22 beliefs have instrumental role

23 less traditional theories

24 1) cognitive dissonance costs Akerlof and Dickens (1982)

25 2) anticipatory utility Brunnermeier and Parker (2005)

26 beliefs play a hedonic role

27 rational agents manipulate beliefs (based on costs)

28 demand for beliefs

29 general behavioral evidence

30 stated overconfidence in surveys (Svenson 1983)

31 (attribute and signal space) Santos-Pinto and Sobel (2005) Benoit and Dubra (2011)

32 self-serving bias Babcock and Lowenstein (1993)

33 attribution theory Heider (1958)

34 confirmation bias Lord, Ross and Lepper (1977)

35 valence-dependent processing/search Eil and Rao (2011) Mobius et al. (2011)

36 (could be just a bias towards positive, not demand for beliefs)

37 ideally we d have:

38 1) beliefs formed recently

39 1) beliefs formed recently 2) data cannot be misinterpreted with slight biases or difference in priors

40 1) beliefs formed recently 2) data cannot be misinterpreted with slight biases or difference in priors 3) trigger financial decisions

41 the Family Radio group satisfy these properties

42 we study them in 4 ways

43 1) time-preference (belief) study

44 1) time-preference (belief) study 2) survey of group members

45 1) time-preference (belief) study 2) survey of group members 3) model of beliefs

46 1) time-preference (belief) study 2) survey of group members 3) model of beliefs 4) analysis of online message board

47 <Background>

48 Harold Camping

49 Camping May, years old

50 Successful Civil Engineer UC Berkeley Graduate

51 Founder and head of Family Radio

52 theological basis

53 biblical literalism

54 through various other calculations/numerology date pinned

55

56 through various other calculations/numerology date pinned

57 through various other calculations/numerology date pinned

58 no longer a possibility

59 absolute certainty

60 guaranteed

61 Prediction: On May 21

62

63

64 Predicted May 21

65 then 5 months of

66 Armageddon

67 Oct 21, 2011 Physical End of World

68 (based on literal interpretation of Book of Revelation)

69 other denominations opt for metaphorical interpretation

70 Timeline of the Prediction

71 2004: Camping discusses prediction on FR air

72 2004: Camping discusses prediction on FR air 2004: Long-time financial manager of FR dies

73 2004: Camping discusses prediction on FR air 2004: Long-time financial manager of FR dies 2005 March: Camping publishes Time Has an End formally proving May 21, 2011 Judgment Day

74 2004: Camping discusses prediction on FR air 2004: Long-time financial manager of FR dies 2005 March: Camping publishes Time Has an End formally proving May 21, 2011 Judgment Day : FR liquidates assets (FCC spectrum), donations pick up

75 2004: Camping discusses prediction on FR air 2004: Long-time financial manager of FR dies 2005 March: Camping publishes Time Has an End formally proving May 21, 2011 Judgment Day : FR liquidates assets (FCC spectrum), donations pick up : spending balloons, mainly through advertising, assets swell then drop

76 Family Radio balance sheet (compiled from IRS 990 filings)

77 (guess when prediction happened)

78 May 7-14, 2011: Our experiment

79 May 7-14, 2011: Our experiment May 21, 2011: Nothing out of the ordinary

80 May 7-14, 2011: Our experiment May 21, 2011: Nothing out of the ordinary May 23, 2011: Camping declares spiritual judgment had occurred. End of world would still occur Oct 21, Rapture now added to Oct 21.

81 May 7-14, 2011: Our experiment May 21, 2011: Nothing out of the ordinary May 23, 2011: Camping declares spiritual judgment had occurred. End of world would still occur Oct 21, Rapture now added to Oct 21. June 6, 2011: Camping suffers stroke, makes no further public appearances

82 May 7-14, 2011: Our experiment May 21, 2011: Nothing out of the ordinary May 23, 2011: Camping declares spiritual judgment had occurred. End of world would still occur Oct 21, Rapture now added to Oct 21. June 6, 2011: Camping suffers stroke, makes no further public appearances Oct 21, 2011:?

83 evidentiary basis similar to religious groups in general (more controversial)

84 2006 Gallup poll: 3 in 10 Americans: Bible is literal word of God 4 in 10: God created humans in present form about 10,000 years ago 4 in 10: angles and demons are active in the world 92% believe in God, 79% absolute certainty

85 why study these people?

86 1) Religious beliefs guide decisions

87 2) These beliefs have a similar (or at least not too different) basis as those held by many people

88 3) verifiable predictions based in the supernatural are rare

89 <Experiment and Survey>

90 experiment basic idea: - more you believe in May 21, less valuable money delivered after May 21 is

91 experiment basic idea: - more you believe in May 21, less valuable money delivered after May 21 is - to measure demand for beliefs, we vary stakes, through changing probability decision is executed

92 experiment basic idea: - more you believe in May 21, less valuable money delivered after May 21 is - to measure demand for beliefs, we vary stakes, through changing probability decision is executed - use Seventh-day Adventists as controls

93 experiment design: - used BDM mechanism - check for $5 cashable immediately - check for $X cashable in 4 weeks - $X ranged from $ make choice - decision enacted with either 6% prob or 50% prob

94 experiment features: - totally private - decisions not seen by anyone except person handling payments

95 decision sheet:

96 Control group

97 Seventh-day Adventists - believe in imminent return of Christ

98 Seventh-day Adventists - 9 th largest denomination in U.S.

99 Seventh-day Adventists - approached in same way, after services

100 Summary statistics and participant overview

101

102

103

104 FR seem to believe, up to stakes $500

105 SDA show reasonable discount factor (0.71)

106 inelastic demand for beliefs

107 what did FR members think they would do if prediction failed?

108 Imagine you wake up on May 21 st and there are no visible signs that Judgment Day occurred, what percent chance would you then place on the End of the World occurring on October 21st?

109 Even answering admitting fallible?

110 all but 6 answered - 1 blank - 5 May 21 will happen!

111

112 mean: 80.1% significantly different from May 21 belief: z=2.44

113 shows more doubt but

114 modal subject says they will continue to believe 100% even if prediction fails in key ways

115 but a minority do say they will stop believing

116 What did happen?

117 May 21, 2011

118 May 21, 2011 quietly passes

119 May 23, 2011

120 Spiritual judgment occurred. God has judged us all, silently. Through His mercy has spared us Armageddon. The end of the world and Rapture will come quietly on Oct 21, Camping (in national press conference)

121 How did believers adjust?

122 IRB: could not follow members

123 Next best: textual analysis of online message board

124 Message board overview - open to public to read, must join to post - we recorded all messages 1 month before and 1 month following May 21 - RA coded over 5000 messages for content - Analysis only includes users who posted before and after

125

126 flurry of uncertainty following May 21

127

128 alternate dates

129

130 (none prior to May 21)

131 alternate date: typically predicts Rapture within 1 day of post

132 seem hopeful

133 user-level analysis

134

135 Before: 87% strong believers After: 53%

136 Before: 4.5% non-believers After: 11.2%

137 89% continue to believe at least somewhat

138 what about attrition?

139 if we assume all attriters became non-believers: 28.5% non-believers

140 ex-post rationalization? or

141 commitment to believe

142 mean prospective continued belief: 80% (quite similar to observed)

143 word frequency analysis

144

145

146 Official theory: Pre-May 21: Rapture, earthquakes, fire, etc.

147 Official theory: Pre-May 21: Rapture, earthquakes, fire, etc. Post-May 23: spiritual judgment, no Rapture

148 spiritual judgment not even mentioned before May 21

149 other alternate theories were discussed Ex. Matthew 24:36

150 related work

151 Festinger (1957) UFO cult and cognitive dissonance

152 Hader (1983) Nuclear war cult

153 Often barriers to exit high

154 Family Radio had

155 no church

156 no compound

157 members lived normally in community

158 yet still significant attachment to a falsified theory

159 (consistent with what the told us ahead of time)

160 Model

161 religious beliefs in general seem extreme

162 some believe fully in God

163 others think this is crazy

164 and we all have similar evidence (more controversial)

165 colloquially you believe or you don t

166

167

168 What is the probability you think God of taken more broadly, a spiritual higher power exists?

169

170 <Conclusion>

171 even beliefs formed on insufficient evidence seem tough to move

172 the mind seems to commit in advance to many beliefs

173 could help explain other examples of a failure to update

174 a demand for beliefs model can explain extreme religious beliefs generally

175 the supply and demand for beliefs appears to be a strong determinant of human behavior

176 <The End>