Abstract:
This dissertation consists of two parts: experimental study and auction design. This experimental study was conducted during August-September 2011 at the Faculty of Economics, Chulalongkorn University, Bangkok, Thailand. It set an economic experiment which consisted of four sessions and was voluntarily participated by seventy-nine undergraduate students. There are two main studies from the experiment: the relationship between perceived intention and positive reciprocity and prediction performance of a reciprocity model proposed by Dufwenberg and Kirchsteiger (2004), or "DK model." Both results are related to the positive-reciprocity behavior -- where a receiver who was given kindness by a giver kindly returns. In the first study, it presents theoretical relationship between cost of giving and positive reciprocity. The analysis shows positive relationships between cost of giving and reciprocity. Then, it presents experimental results which tested the relationship. The results confirm this relationship. In the second study, we test the DK model’s performance in predicting positive-reciprocity decisions. A new approach to measure the model's performance was introduced. The results show that the DK model has good performance. Furthermore, we compare the DK model's performance with two alternative prediction methods: DG method and Personal-info method. The results also show that the DK model still performed the best among them. For the auction design, we theoretically propose auctions which are optimal to be applied to an object with countervailing-positive externalities. The newly proposed auction is called "take-or-give auction with second-price payment." Unlike the basic auction which lets bidders compete for obtaining the object, the proposed auction lets bidders compete for their desired allocation. It solves the free-rider problem and allocates the object efficiently. Moreover, to increase the expected revenue the study proposes some extended versions of the take-or-give auction. By introducing a set of revenue-enhancing rules which include entry fee, no sale condition and pooling rule, the auction optimally maximizes the revenue. It is the revenue-maximizing auction for an object with countervailing-positive externalities.
