Kern, Johannes (2014). Four Essays on Decision and Game Theory. PhD thesis, Universität zu Köln.
|
PDF
Dissertation_Kern.pdf Download (1MB) |
Abstract
This dissertation consists of four research papers, covering topics from decision and game theory. Chapter 1 presents a game-theoretically well-founded framework for modeling repeated games in continuous time and proves that in the resulting extensive form every strategy profile induces one and only one outcome. We further show that strategies in this framework are equivalent to a class of strategies on which exogenous restrictions have been imposed hereby providing a clarification as to which restrictions on strategies can be allowed in the sense that the resulting strategies can be derived from a well-defined extensive form. Chapter 2 comments on the definition of Extensive Form in Alós-Ferrer and Ritzberger (2008) and shows that one of the properties there needs to be adjusted. It provides counterexamples showing that with the original version of this property some results do not hold as stated and presents a corrected formulation of the property as well as the corrected statement of the results. It further provides proofs for these results under the new formulation. In Chapter 3 we introduce a class of games with a cyclical structure we coin “circulant games” in which both players payoffs are given by circulant matrices. Given that these payoffs are ordered, we are able to determine the exact number of (pure and mixed) Nash equilibria. Our results further allow us to describe the support of each Nash equilibrium strategy. Chapter 4 investigates the preference reversals phenomenon. We present a new, simple process-based model that explains the preference reversal phenomenon and makes novel predictions about the associated decision times in the choice phase. The phenomenon is jointly caused by noisy lottery evaluations and an overpricing phenomenon associated with the compatibility hypothesis. A laboratory experiment confirmed the model’s predictions for both choice data and decision times. Choices associated with reversals take significantly longer than non-reversals, and non-reversal choices take longer whenever long-shot lotteries are selected. A second experiment showed that the overpricing phenomenon can be shut down, greatly reducing reversals, by using ranking-based, ordinally-framed evaluation tasks. This experiment also disentangled the two determinants of the preference reversal phenomenon since noisy evaluations still deliver testable predictions on decision times even in the absence of the overpricing phenomenon.
Item Type: | Thesis (PhD thesis) | ||||||||
Creators: |
|
||||||||
URN: | urn:nbn:de:hbz:38-56848 | ||||||||
Date: | 2014 | ||||||||
Language: | English | ||||||||
Faculty: | Faculty of Management, Economy and Social Sciences | ||||||||
Divisions: | Weitere Institute, Arbeits- und Forschungsgruppen > Staatswissenschaftliches Seminar | ||||||||
Subjects: | Economics | ||||||||
Uncontrolled Keywords: |
|
||||||||
Date of oral exam: | 10 July 2014 | ||||||||
Referee: |
|
||||||||
Refereed: | Yes | ||||||||
URI: | http://kups.ub.uni-koeln.de/id/eprint/5684 |
Downloads
Downloads per month over past year
Export
Actions (login required)
View Item |