This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Click to edit Master subtitle style Page ‹#› Lecture #5: Decision Making Under Risk and Uncertainty (Part 1) Decision Making under Risk and Uncertainty (Part 1 of 4) 1 1 Lecture #5: Decision Making Under Risk and Uncertainty (Part 1) "The evils of uncertainty must count for something". Alfred Marshall (1920), in Principles of Economics . Page ‹#› Lecture #5: Decision Making Under Risk and Uncertainty (Part 1) What we learned from the last 2 lectures Expected values, variances, standard deviations, covariances, correlations Discrete and continuous probability distributions Central Limit Theorem Normal Distribution 2 2 Lecture #5: Decision Making Under Risk and Uncertainty (Part 1) Page ‹#› Lecture #5: Decision Making Under Risk and Uncertainty (Part 1) Risk Pooling Making decisions under certainty Making decisions under uncertainty: the expected value model The St. Petersburg Paradox Expected utility model 3 3 Lecture #5: Decision Making Under Risk and Uncertainty (Part 1) Today’s Game Plan Page ‹#› Lecture #5: Decision Making Under Risk and Uncertainty (Part 1) Risk Pooling Let E ( L i ) = expected loss for insured i and 1 ( ) ( ) n T i i E L E L = = ∑ = total expected loss of the risk pool. Then E ( L p ) = w E L i i i n ( ) = ∑ 1 (1) = average loss per policy, where w i = E ( L i )/ E ( L T ). Similarly, σ ρ σ σ Lp i j n i n j ij i j w w 2 1 1 = = = ∑ ∑ (2) = average risk per policy, where ρ ij = σ i / σ i σ j = correlation between losses on policy i and policy j . 4 4 Lecture #5: Decision Making Under Risk Page ‹#› Lecture #5: Decision Making Under Risk and Uncertainty (Part 1) Risk Pooling • Let losses be identically distributed; i.e., E ( L i ) = μ, σ ι 2 = σ 2 , and w i = 1/ n for all insureds, while { 2 2 if if . ij i j i j i j σ ρσ ρ σ σ = ≠ = Therefore, (1) and (2) are rewritten E ( L p ) = w E L i i i n ( ) = ∑ 1 = (1/ n ) n μ = μ, and (1a) σ ρ σ L ij j n i n p n 2 2 2 1 1 1 = = = ∑ ∑ 2 2 1 . n n n σ ρσ = + (2a) 5 5 Lecture #5: Decision Making Under Risk Page ‹#› Lecture #5: Decision Making Under Risk and Uncertainty (Part 1) Risk Pooling • Since 2 2 2 1 p L n n n σ σ ρσ = + , 2 2 lim ; p L n σ ρσ →∞ = i.e., only covariance risk remains. • Now suppose losses are iid . Then 2 lim 0. p L n σ →∞ = Thus the average loss becomes more predictable as the number of risks pooled becomes large. o By pooling many independent risks, insurers can treat uncertain losses as almost known. o Risk pooling effectively defrays risk by exploiting the law of large numbers. 6 6 Lecture #5: Decision Making Under Risk Page ‹#›...
View
Full Document
 Spring '09
 Derivatives, Decision Making, Utility, Options, St. Petersburg paradox, Blanche

Click to edit the document details