330 likes | 444 Views
Joep Sonnemans. Theo Offerman. Gijs van de Kuilen. Peter P. Wakker. A dapting de Finetti's proper scoring rules for Eliciting Bayesian Beliefs to Prospect Theory. July 6, 2006 IAREP, Paris. Topic: Our chance estimates of France to become world-champion.
E N D
Joep Sonnemans Theo Offerman Gijs van de Kuilen Peter P. Wakker Adapting de Finetti's proper scoring rules for Eliciting Bayesian Beliefs to Prospect Theory July 6, 2006IAREP, Paris Topic:Our chance estimates of France to become world-champion. F: France will win. not-F: Italy.
2 Imagine following bet: You choose 0 r 1, as you like. We call r your reported probability of France. You receive F not-F 1 – (1– r)2 1 – r2 What r should be chosen?
3 First assume EV. After some algebra: p true subjective probability;optimizing p(1 – (1– r)2) + (1–p) (1–r2);1st order condition 2p(1–r) – 2r(1–p) = 0;r = p. Optimal r = your true subjective probability of France winning. Easy in algebraic sense. Conceptually: !!! Wow !!! de Finetti (1962) and Brier (1950) were the first neuro-economists!
4 "Bayesian truth serum" (Prelec, Science, 2005). Superior to elicitations through preferences . Superior to elicitations through indifferences ~ (BDM). Widely used: Hanson (Nature, 2002), Prelec (Science 2005). In accounting (Wright 1988), Bayesian statistics (Savage 1971), business (Stael von Holstein 1972), education (Echternacht 1972), medicine (Spiegelhalter 1986), psychology (Liberman & Tversky 1993; McClelland & Bolger 1994), experimental economics (Nyarko & Schotter 2002). We want to introduce these very nice features into prospect theory.
5 Survey Part I. Deriving r from theories: expected value; expected utility; nonexpected utility for probabilities; nonexpected utility for ambiguity. Part II. Deriving theories from observed r. In particular: Derive beliefs/ambiguity attitudes. Will be surprisingly easy. Proper scoring rules <==> Prospect theory:Mutual benefits. Part III. Implementation in an experiment.
6 Part I. Deriving r from Theories (EV, and then 3 deviations).
7 Let us assume that you strongly believe in France (Zidane …) Your "true" subj. prob.(F) = 0.75. EV: Then your optimal rF = 0.75.
8 1 R(p) rEV rEU 0.69 0.61 rnonEU 0.75 0.50 rnonEUA nonEU EU 0.25 EV 0 1 0 0.25 0.50 0.75 p go to p. 11, Example EU go to p. 19, Example nonEUA go to p. 15, Example nonEU next p. Reported probability R(p) = rF as function of true probability p, under: (a) expected value (EV); (b) expected utility with U(x) = x (EU); (c) nonexpected utility for known probabilities, with U(x) = x0.5 and with w(p) as common; rnonEUA: nonexpected utility for unknown probabilities ("Ambiguity").
9 So far we assumed EV (as does everyone using proper scoring rules, but as no-one does in modern decision theory...) Deviation 1 from EV: EU with U nonlinear Now optimize pU(1 – (1– r)2) + (1 – p)U(1 – r2)
p r = (1–p) p + U´(1–r2) U´(1–r2) U´(1–(1–r)2) U´(1–(1–r)2) r p = (1–r) r + 10 Reversed (and explicit) expression:
go to p. 8, with figure of R(p) 11 How bet on France? [Expected Utility]. EV: rEV = 0.75. Expected utility, U(x) = x: rEU = 0.69. You now bet less on France. Closer to safety (Winkler & Murphy 1970).
12 Deviation 2from EV: nonexpected utility for probabilities (Allais 1953, Machina 1982, Kahneman & Tversky 1979, Quiggin 1982, Schmeidler 1989, Gilboa 1987, Gilboa & Schmeidler 1989, Gul 1991, Levy-Garboua 2001, Luce & Fishburn 1991, Tversky & Kahneman 1992, Birnbaum 2005) Fortwo-gainprospects, virtually all those theories areas follows: For r 0.5, nonEU(r) = w(p)U(1 – (1–r)2) + (1–w(p))U(1–r2). r < 0.5, symmetry; soit! Different treatment of highest and lowest outcome: "rank-dependence."
1 w(p) .51 1/3 1 0 1/3 p 2/3 Figure.The common weighting function w. w(p) = exp(–(–ln(p))) for = 0.65. w(1/3) 1/3; w(2/3) .51 13
w(p) r = (1–w(p)) w(p) + U´(1–r2) U´(1–r2) U´(1–(1–r)2) U´(1–(1–r)2) Reversed (explicit) expression: ) ( r w–1 p = (1–r) r + 14 Now
go to p. 8, with figure of R(p) 15 How bet on France now? [nonEU with probabilities]. EV: rEV = 0.75. EU: rEU = 0.69. Nonexpected utility, U(x) = x, w(p) = exp(–(–ln(p))0.65). rnonEU = 0.61. You bet even less on France. Again closer to safety. Deviations were at level of behavior so far, not of be-liefs. Now for something different; more fundamental.
16 3rd violation of EV: Ambiguity (unknown probabilities; belief/decision-attitude? Yet to be settled). No objective data on probabilities. How deal with unknown probabilities? Have to give up Bayesian beliefs descriptively. According to some even normatively.
Instead of additive beliefs p = P(F), nonadditive beliefs B(F) (Dempster&Shafer, Tversky&Koehler, etc.) 17 All currently existing decision models: For r 0.5, nonEU(r) = w(B(F))U(1–(1–r)2) + (1–w(B(F)))U(1–r2).Don't recognize? Is just '92 prospect theory = Schmeidler (1989)! Write W(F) = w(B(F)).Can always writeB(F) = w–1(W(F)). For binary gambles: Pfanzagl 1959; Luce ('00 Chapter 3); Ghirardato & Marinacci ('01, "biseparable").
w(B(F)) rF = (1–w(B(F))) w(B(F)) + Reversed (explicit) expression: U´(1–r2) U´(1–r2) ) U´(1–(1–r)2) U´(1–(1–r)2) ( w–1 B(F) = r (1–r) r + 18
go to p. 8, with figure of R(p) 19 How bet on France now?[Ambiguity, nonEUA]. rEV = 0.75. rEU = 0.69. rnonEU = 0.61 (under plausible assumptions). Similarly, rnonEUA = 0.52. r's are close to always saying fifty-fifty. "Belief" component B(F) = w–1(W) = 0.62.
20 B(F): ambiguity attitude /=/ beliefs?? Before entering that debate, first: How measure B(F)? Our contribution: through proper scoring rules with "risk correction."
) ( w–1 p = We reconsider reversed explicit expressions: ) ( U´(1–r2) U´(1–r2) w–1 B(F) = U´(1–(1–r)2) U´(1–(1–r)2) r r (1–r) (1–r) r + r + Part II. Deriving Theoretical Things from Empirical Observations of r 21 Corollary.p = B(F) if related to the same r!!
stock 20, CSM certificates dealing in sugar and bakery-ingredients.Reported probability: r = 0.75 91 91 Example (participant 25) 22 For objective probability p=0.70, also reports r = 0.75. Conclusion:B(elief) of ending in bar is 0.70! We simply measure the R(p) curves, and use their inverses: is risk correction.
23 Our proposal takes the best of several worlds! Need not measure U,W, and w. Get "canonical probability" without measuring indifferences (BDM …; Holt 2006). Calibration without needing many repeated observations. Do all that with no more than simple proper-scoring-rule questions.
24 Mutual benefits prospect theory <==> proper scoring rules We bring insights of modern prospect theory to proper scoring rules. Make them empirically more realistic. (EV is not credible in 2006 …) We bring insights of proper scoring rules to prospect theory. Make B very easy to measure and analyze. Directly implementable empirically. We did so in an experiment, and found plausible results.
25 Part III. Experimental Test of Our Correction Method
26 Method Participants. N = 93 students. Procedure. Computarized in lab. Groups of 15/16 each. 4 practice questions.
27 Stimuli 1. First we did proper scoring rule for unknown probabilities. 72 in total. For each stock two small intervals, and, third, their union. Thus, we test for additivity.
28 Stimuli 2. Known probabilities: Two 10-sided dies thrown. Yield random nr. between 01 and 100. Event F: nr. 75 (etc.). Done for all probabilities j/20. Motivating subjects. Real incentives. Two treatments. 1. All-pay. Points paid for all questions. 6 points = €1. Average earning €15.05. 2. One-pay (random-lottery system). One question, randomly selected afterwards, played for real. 1 point = €20. Average earning: €15.30.
29 Results
0.7 0.6 0.5 treatment one 0.4 0.3 0.2 treatment all 0.1 0 31 F( ρ ) 1 0.9 Individual corrections 0.8 ρ -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
32 uncorrected uncorrected Figure 9.1. Empirical density of additivity bias for the two treatments Fig. b. Treatment t=ALL Fig. a. Treatment t=ONE 160 160 140 140 corrected 120 120 100 100 80 80 60 60 corrected 40 40 20 20 0 0 0.2 0.6 0.4 0 0.2 0.4 0.6 0.6 0.4 0.2 0 0.2 0.4 0.6 For each interval [(j2.5)/100, (j+2.5)/100] of length 0.05 around j/100, we counted the number of additivity biases in the interval, aggregated over 32 stocks and 89 individuals, for both treatments. With risk-correction, there were 65 additivity biases between 0.375 and 0.425 in the treatment t=ONE, and without risk-correction there were 95 such; etc.
33 Summary and Conclusion • Modern decision theories: proper scoring rules are heavily biased. • We correct for those biases, with benefits for proper-scoring rule community and for nonEU community. • Experiment: correction improves quality; reduces deviations from ("rational"?) Bayesian beliefs. • Do not remove all deviations from Bayesian beliefs. Beliefs seem to be genuinely nonadditive/nonBayesian/sensitive-to- ambiguity.