非中立框架下同時標售類似標的之全付拍賣實驗

Abstract

本文在補習考大學的非中立框架下進行兩個相似標的之最高價拍賣和全付拍賣，比較它們如何影響台灣人對高等教育補貼的態度。每場實驗都有六位買家競標，每回合有兩位買家得獎，最高標買家獲得首獎、另一位獲得次獎。六位買家每回合各自獨立抽取首獎的價值，而次獎的價值為首獎的 80%。實驗結果顯示，全付拍賣的平均出價接近風險中立貝氏納許均衡（RNBNE），但在最高價拍賣中，許多受試者出價恰為首獎的 80%、以確保得標時不會虧損，因此平均出價高於均衡預測。這使得最高價拍賣比全付拍賣產生更高的收益，也更有效率。最後，我們發現在兩種拍賣形式中，買家出價在前十回合朝均衡收斂，但之後就停止並趨於穩定。

We compare first-price and all-pay auctions with two units sold under the non-neutral framing of college admission, and their effect on people's attitudes toward higher education subsidy in Taiwan. In each experimental session, 6 bidders draw private values independently (for each period) and compete to win one of two prizes, with the second prize valued at 80% of the first prize. We find average bids in the all-pay auction close to the risk neutral Bayesian Nash equilibrium (RNBNE), while that in the first-price auction exceed the RNBNE since many subjects bid 80% of the first prize (guaranteeing not making a loss when winning). This results in the first-price auction yielding higher revenue compared to the all-pay auction and being more efficient. Lastly, we find bidding converges toward RNBNE, but stops and stabilizes after the first ten rounds in both auction formats.