UNIVERSITY OF OSLO DEPARTMENT OF ECONOMICS Home exam: ECON 5200 / ECON 9200 Exam period: 5.12.2011, at 09:00 9.12.2011, at 12:30 (unless other individual deadline granted) Grades are given: 3.01.2012. Guidelines: Submit your exam answer electronically to: submissions@econ.uio.no (note that this generate an automated notice that the mail is received) or on paper to the Department office on 12 th floor. Last day for submission: 9.12.2011, at 12:30, (unless other individual deadline granted). Written text should be in pdf format. Please remember to also submit Delcaration which you will find on the course web page. This must be submitted as a separate document. Use your candidate number both as the name of the file you submit, and as the author name in the file. Do NOT use your name! You will find your candidate number on your StudentWeb. If you have problems, please contact Tone Enger at tone.enger@econ.uio.no. Further instructions: The questions are in English, but you can give your answers in English, Norwegian, Swedish or Danish. The home assignment will be marked. Students on masters level are awarded on a descending scale using alphabetic grades from A to E for passes and F for fail. Students on phd-level are awarded either a passing of failing grade. Your answer must fill the formal requirements, found at http://www.sv.uio.no/studier/ressurser/kildebruk/ (Norwegian) or http://www.sv.uio.no/english/studies/resources/sources-and-references/ (English). It is of importance that your paper is submitted by the deadline (see above). Papers submitted after the deadline, will not be corrected.*) All papers must be delivered to the place given above. You must not deliver your paper to the course teacher or send it by e-mail. If you want to hand in your paper before the deadline, please contact the department office on 12 th floor. *) The rules for illness during exam also applies for the home exams. Please see http://www.sv.uio.no/english/studies/admin/exams/postponed-exam/index.html for further details.
General information: Four of the following six problems must be solved. You may discuss the problems with other candidates, but each candidate must write a separate and independent solution. Problem 1 Chapter 3.I of Mas-Colell et al. derives Equivalent Variation and Compensating Variation as measures of individual welfare changes, while chapter 10.E argues that maximizing the Marshallian surplus yields a Pareto efficient outcome. While EC and CV defines consumer surplus as the area under the Hicksian demand curve, the Marshallinan surplus uses the Walrasian demand. (a) How can the results be reconciled? In a review of the literature, John K. Horowitz and Kenneth E. McConnell ( A Review of WTA/WTP Studies, Journal of Environmental Economics and Management 44, 2002, 426-447) finds that the difference between WTA and WTP (corresponding to CV and EV) is very large for some goods, but it depends on the type of good in question. (b) Discuss these findings in the light of the argument in W. Michael Hanemann Willingness to Pay and Willingness To Accept: How much can they differ (The American Economic Review 81, 1991, 635-647). Do you think that Hanemann s paper gives a plausible explanation for the empirical size of these differences, and how they varies with the type of goods? Problem 2 Consider a finite set of possible levels of a public goods with corresponding distributions of cost (g n, t 1n,..., t In ) for n {1, 2,...N} and where t in is the tax paid by individual i, potentially negative. The sum equals the cost, c(g n ), of the public good: i t in = c(g n ). We consider four alternatives: 1
2 a1 g 1 = t i1 = 0, a2 g 2 = ḡ > 0 and t i2 = 1 N c(g n), a3 g 3 = ḡ > 0 and t i3 is set proportional to WTP: t i3 = WTP in j WTP c(g n ), jn a4 g 3 = ḡ > 0 and t i3 is set such that y i t i3 = y j t j3 for any pair of individuals, i, j, where y i is income for individual i. Consider a Cost-Benefit ranking where alternatives are ranked according to WTPin c(g n ), where WTP in is i s individuals willingness to pay for the public good level in alternative n. (Formally: (y i WTP in, g n ) i (y i, 0).) (a) List the possible rankings of the three alternatives that may result from such a Cost-Benefit ranking. (b) Amartya Sen ( Personal Utilities and Public Judgements: Or what is wrong with welfare economics, The Economic Journal 89, 1979, 537-558) argues that Arrow s approach to social choice severely limits the information that is allowed to enter into the ranking. What information is used in the Cost-Benefit ranking above while it is ruled out in Arrow s approach? Explain why the Cost-Benefit ranking is not a counterexample to Arrow s Impossibility Theorem. Problem 3 Proposition 6.B.3 in Mas-Colell et al. shows that ranking according to expected utility follows from axioms, where independence is one of the key axioms. Uri Gneezy and Jan Potters ( An experiment in Risk Taking and Evaluation Periods Quarterly Journal of Economics 112, 1997, 631-645) test an implication of the expected utility hypothesis due to P.A. Samuelson; that if a single gamble is rejected at every relevant wealth position, then accepting the multiple repetitions of the same gamble is inconsistent with maximization of expected utility.
(a) Prove Samuelson s claim and, in particular, explain why the assumption concerning wealth is important. (b) Do Gneezy and Potters results violate the implication of expected utility pointed out by Samuelson? Mathew Rabin and Richard H. Thaler ( Anomalies: Risk Aversion Journal of Economic Perspectives 15, 2001, 219-232) extend Samuelson s theoretical result and concludes negatively on the descriptive validity of expected utility, to the extent that they conclude: Indeed, we aspire to have written one of the last articles debating the descriptive validity of the expected utility hypothesis (p. 229). (c) Do the results cited above provide evidence for violation of the axioms of expected utility, and in particular the independence axiom? 3 Problem 4 (a) Consider Example 27.1 in Osborne & Rubinstein of a second-price auction. Formulate this as a Bayesian game and show that this game has a Nash equilibrium where all players bid their valuations. (b) Consider now the same situation, but with the difference that v i max j N\{i} a j is replaced by v i a i. This turns the game into a firstprice auction. Formulate this as a Bayesian game and characterize a symmetric Nash equilibrium. (c) Use Roger B. Myerson, Optimal Auction Design (Mathematics of Operations Research 6, 1981, 58 73) and John G. Riley and William F. Samuelson Optimal Auctions (American Economic Review 71, 1981, 381 392) to argue that the expected revenue for the auctioneer is the same for both auctions. (d) Provide a general discussion of what assumptions this revenue equalization theorem is based on.
4 Problem 5 Consider two classes of games: (i) The class of undiscounted finitely repeated games and (ii) the class of δ-discounted infinitely repeated games, where 0 < δ < 1. Folk theorems for these classes of games are proven in Jean-Pierre Benoit and Vijay Krishna Finitely Repeated Games (Econometrica 53, 1985, 905 922) and Drew Fudenberg and Eric Maskin The Folk Theorem in Repeated Games with Discounting or with Incomplete Information (Econometrica 54, 1986, 533 554). See also Sections 8.8 and 8.9 of Osborne & Rubinstein and Appendix A of Chapter 12 of Mas-Colell et al. (a) For each of these classes of games, what is meant by the folk theorem? (b) For each of these classes of games, explain what assumptions are needed to establish this result for games with 3 or more players. In particular, what is the role of the dimensionality assumption? (c) Why is the dimensionality assumption not needed for games with only two players? Problem 6 In the hidden action model in Section 14.B of Mas-Colell et al., the agent is risk averse and effort is one-dimensional. (a) Characterize the optimal contract when a good, objective output measure is available. (b) Use the analysis by Holmstrom and Milgrom ( Multitask Principal- Agent Analyses: Incentive Contracts, Asset Ownership, and Job Design, J Law Econ Org 7, 1991, special issue, 24 52) to discuss how your conclusion changes when effort is two-dimensional and a good, objective output measure is available only for one of the dimensions.