| Abstract: | 由於自我實現是種訊息"傳"與"染"所造成的結果。因此,我們希望分析的模型能夠掌握到這個"傳"與"染"的途徑。在總體經濟學上曾經風騷一時的"代表性個人" (representative agents)模型,因是一人模型,在所有現象均同步發生的前提下,並不宜用於捕捉"傳染"的過程。雖然,有關"傳染"對經濟行為的影響在Shiller(1980)曾有精闢的描述,但是他並沒有提供我們一可以操作的數學模型。而在傳染模型上面,從40年代voe Neumann所思考的一部"生命自我複製機" (self-reproducing automata)到60年代Conway的"生命之局"(game of life)可以說已開啟了我們一個可以運用的數學模型。而這一發展到了Wolfram的細胞互動(cellular automata)模型提出後更趨成熟。晚近,細胞互動模型的應用也進一步的延伸到經濟學的領域來,如Keenan與O`brien(1993)便是一例。 本文便是利用Wolfram的細胞互動模型來建立一個可以觀察訊息傳與染的透明過程。第二節是探討這個問題的佈局與數學化。並從該過程中來思考自我實現預期下的多元均衡為什麼能形成一種誘惑的來源,從而探討誠實是否是最好的策略。在第二節中,我們提出對問題的第一種佈局,並根據這佈局架構起本文所使用的細胞互動模型。第三節到第六節則是對第二節所建立的模型進行模擬。第三節的模擬旨在透過不同的參數比較來說明媒體的一個特質,即它不僅只是消極的報導群眾的意向,甚至還能積極發揮整體群眾意向至共識的功能。第四節則模擬不實報導如何能透過這個功能而產生有利於獨占媒體的一種共識,從而揭露不實報導誘因的來源。第五節則用模擬來了解這種誘因強度在群眾是理性的前提下,會受到什麼影響?群眾的理性是由貝氏學習來代表。模擬顯示,"在一般情況下,誠實並非是最好的策略"。第六節則在考慮了獨占媒體的有限理性後,重做第五節的模擬,發現"條件性的誠實才是最好的策略"。第七節則是結論。
What is the nature of the existence of stochastic multiple equilibria and coordination failures in a decentralized economy? "Can and should" the government do anything about the stochastic multiple equilibria? More precisely, does the government have to take the risk of losing its credibility while attempting to fine-tune the economy? This paper extends the idea of path-dependent processes into a model based on cellular automata (or interacting heterogenous agents) to illustrate the nature of stochastic multiple equilibria from the perspective of animal spirits which are characterized by the coexistence of Bayesian learning and mimetic contagion in each cellular automaton. The asymptotic distribution of multiple equilibria which was obtained through simulation delivers the message that the effectiveness of government intervention is conditioned by the size of shock, a result which has been largely neglected in the rule-vs.-discretion debate. When the size of the shock is "moderate", the government can in fact gain rather than lose its credibility by fine-tuning the economy. |
