寬恕政策實施之三階段賽局有效性分析

本研究主要針對廠商的聯合行為及寬恕政策實施有效性，進行二階段與三階段賽局分析，並根據分析結果提出政策建議。藉由將廠商是否因實施寬恕政策而選擇聯合加入第一階段，得出在三階段模型下，執法機構調查成功機率與開始調查前第一家告發廠商之罰款減免比例越高時，政策成效越佳且具高度穩定性;而調查開始後之罰款減免不影響政策實施之成效。另外，比起僅第一家告發廠商可獲得罰款減免，若第二家廠商亦可獲得罰款減免時，其寬恕政策申請之可能性相對較高。再者，當二階段基本模型與三階段模型罰款減免之和相同時，兩模型之結果相等。最後，當將二模型應用至美國、台灣、歐盟及日本四個國家或經濟體時，若以政策有效性以及申請之可能性為判斷依據，我們得出日本制定的政策優於或無異於歐盟的政策，而歐盟政策又優於美國和台灣的政策。

關鍵字

寬恕政策；反托拉斯；聯合行為；子賽局完全均衡

並列摘要

This paper mainly focuses on the effectiveness of the leniency policy which makes cartel members spontaneously report to the enforcement. Through extending the models of Silbye (2012) and Wang et al. (2016) by adding the first stage, either the firms collude or not, to the framework, this paper explores two theoretical models, a two-stage and a three-stage subgames, to investigate leniency policy. The result shows that during the post-investigation, the higher the rate of successful investigation and fine reduction of the first granted firm are, the more effective and stable the policy is. In addition, the fine immunity or reduction rate of per-investigation has no effect on the effectiveness of the policy. Moreover, when the sum of the fine reduction rates of the two-stage model is the same as the three-stage model, the result will be the same. Finally, we apply these two models to the policy of United States, Taiwan, European Union, and Japan. If we take the effectiveness and the possibility of the policies into consideration, we find that Japan’s leniency policy is at least as good as E.U.’s, which is better than U.S. and Taiwan’s policy.