1
行政院國家科學委員會補助專題研究計畫成果報告
以連續性賽局分析不動產市場交易的模型
─並以實驗室研究實證
計畫類別：x 個別型計畫 □整合型計畫 計畫編號：NSC 89－2416－H－011－013
執行期間： 88 年 08 月 01 日至 89 年 07 月 31 日
計畫主持人：張光第
執行單位：國立臺灣科技大學企業管理系 中華民國八十九年十月三十一日
2
行政院國家科學委員會專題研究計畫成果報告
以連續性賽局分析不動產市場交易的模型
─並以實驗室研究實證
Applying Dynamic Game Theor y to Analyze the Equilibr ium of Real Estate Tr ansaction: With an Empir ical Test of Exper imental Method
計畫編號：NSC 89－2416－H－011－013 執行期限：88 年 08 月 01 日至 89 年 07 月 31 日
主持人：張光第 國立臺灣科技大學企業管理系助理教授
中文摘要
本文主旨在探討連續性賽局的穩定性 及效率性，並運用實驗法瞭解不動產交易 的均衡狀態，故所採取的賽局模型係根據 John Bryant's (1983) 的 Keynesian coordination game(凱因斯合作協調賽局)及 國科會專題研究計畫編號：NSC 88－2416
－H－011－005「以實驗法研究購屋者之需 求因素」為基本架構。本研究著重不動產 交易決策過程的「連續性」，參賽者其制 訂第一期決策的準則，與其維持或改變往 後幾期決策的原因。本研究設計十種不同 的實驗情境，藉以觀察參賽者面臨各種不 同情境時，制訂決策是否有所不同。經由 此次實驗，我們得到下列結論： (一) 連 續性賽局情境的變動影響決策: 此次實驗 的機會成本與沈沒成本在連續性之改變 下，造成平均報酬的動態差異，當獲利程 度不變時，標準差-亦即風險程度-是會改 變的；(二)制裁性策略與協調失敗: 整個 實驗過程中，理性決策者獲得相當無效率 的報酬，但制裁性決策者卻得到較高的(效 率的)報酬。這樣的無效率起因於選擇最適 報酬均衡的預期風險太高，導致協調失 敗；(三) 制裁性策略無法提供最適均衡:
根據演繹法所描述，理性參賽者應該追求 絕對利益最大。事實上，部分參賽者採制 裁性或報復性策略時，實驗過程產生非預 期結果，並產生無效率的均衡。但實驗也 證明，演繹法帶給理性參賽者一項安全 解，以免損失持續擴大。
關鍵字: 連續性賽局、不動產交易、實驗 法、演繹法、制裁性決策者、策略不確定 性、協調失敗
ABSTRACT
This research is to utilize continuous game theory and experimental method to analyze the stability and efficiency of real estate transactions. Therefore, the model used in the research is derived from John Bryant's (1983) Keynesian coordination game and continuing the methodology of NSC project numbered 88-2416-H-011-005. The emphasis of the research is on "continuance"
of all real estate transactions. Participants make their first decision based on pre-existed rules, and their following decisions are based on pre-existed results. Ten scenarios are designed to test participants in order to observe their reactions to the changes of parameters and to analyze the differences of their decision making. The experimental results reveal the following three findings: (1) The ten continuous scenarios do affect the process of participants’ decision making:
when the opportunity cost and the sunk cost are changed, the average dynamic payoffs are affected, and when payoffs are fixed, the standard deviation, i.e., the risk, is also changed. (2) The counter-active strategy is associated with the coordination failure:
rational participants are supposed to seek maximizing their benefits, but the inefficient equilibria found in the experiments are associated with irrational or counter-active participants. In fact, when some of the participants make counter-active strategies, the experiments resulted in unexpected outcomes, and also resulted in the coordination failure. (3) The counter-active strategy could not provide optimal solutions:
3
The Deductive Selection Principle (DSP) according to the DSP, rational participants are supposed to look for the absolute maximum of payoffs. However, the results of experiments show that the equilibria with some counter-active participants are inefficient, but they are proved that the DSP at least provides the rational participants with a safety first solution, which leaves them with losing less.
Keywor ds: Continuous Game theor y, Exper imental Method, Deductive Selection Pr inciple, Str ategy Uncer tainty, Real Estate Tr ansaction, Counter -Active Str ategy, and Coor dination Failure
