都市住宅市場預警系統動態模擬模式之研究

The Early Warning System Dynamics Simulation Model of Urban Housing Market

劉正智(Cheng-Chih Liu)

指導教授 ： 何友鋒

朝陽科技大學/設計學院/建築及都市設計研究所/博士(2010年)

摘要

都市住宅市場具有優化資源配置、傳遞資訊、財富再分配、提高勞動效率、優化市民消費結構等功能，有著支柱、帶動與財富作用等非常深遠的經濟和社會意義。90年代以後，相對於社經條件轉變，住宅供需急轉直下，榮景不再，都市住宅市場呈現了「三高--高房價、高自有率、高空屋率」的極端演變，在淺盤經濟情境中呈現高度波動，龐大空屋資源閒置造成資金積壓，影響社會資源利用與金融市場穩定，存在房市泡沫危機。
都市住宅市場是一多重要因不斷交互衝擊與傳導結果的完整體現，受到都市人口增減、市場供需時效、經濟環境榮枯、金融政策寬縮之加乘制約，存在時滯性、動態性、因果性、回饋性、敏感性等鮮明自我動態特徵。世界性房市泡沫事件及美國次級房貸風暴，皆明證都市住宅市場運作良窳繫決國家金融體制及整體經濟發展；然國內有關都市住宅市場預警系統未臻完備，面對詭譎多變的都市住宅市場，無法實施外因變化與趨勢觀測，無以落實即時預警與策略模擬，造成住宅市場投資與交易風險。
本研究以都市視野進行住宅市場結構剖析、定量測算和策略模擬，採行多理論協同、多模型集成與個案實證研究，從都市住宅市場預測及預警兩階段構思完整預警系統。藉由住宅市場均衡與非均衡理論、週期波動理論與傳導衝擊理論，探討都市住宅市場運行機制；以系統工程學為核心，應用模糊德爾菲專家諮詢及敏感度分析組構市場預測預警指標體系；結合系統動態學、實數型遺傳演算類神經網路建構都市住宅市場預測預警系統動態模擬模式；透過景氣對策信號機制評價都市住宅市場警情；經由情境模擬進行住宅市場政策尋優。研究發現多方法協同可以系統化應用於都市住宅市場動態研究；複合情境模擬常較單一情境模擬具有較佳政策評價；經濟面排警措施較能本質性改善住宅市場警情；金融面排警措施雖有短線積極效果，卻有引鴆止渴的強烈後遺症，宜審慎施為。
對於都市住宅市場排警之系統性思維，除建議設置「都市住宅市場預警系統動態模擬實驗室」外，對於周延建築物使用修繕策略、健全住宅不動產租賃市場等配套措施亦應兼籌並顧，高效率調控都市住宅市場，提高都市競爭力，實現都市優質環境。

關鍵字

類神經網路 ； 都市住宅市場 ； 系統動態學 ； 模糊德爾菲 ； 系統工程

並列摘要

The urban housing market can optimize resource allocation, transmit information, re-distribute wealth, enhance labor efficiency and optimize citizen’s consuming structure. Besides, it has profound economic and social meanings with support, driving and wealth-creation function. After 1990, the socioeconomic conditions changed. The relationship between the supply and the demand of housing had a sharp change. The prosperity existed no more. There was an extreme change that three “highs” appeared in the urban housing market- high house price, high home-ownership rate and high vacancy rate. A high degree of fluctuation appeared in shallow economic foundation. A large number of vacant houses and the deserted resources overstocked the capital. It hindered the application of social resources and affected the stability of financial market. Thus, a potential bubble crisis of housing market existed.
Urban housing market completely reflected continuous mutual impacts and conductions among many factors, such as the increment and the decrement of urban population, the timing of market supply and demand, the prosperity and the decline of economic environment, the width reduction of the financial policy on the loan giving, show delay, dynamic, causal, feedback, sensitivity and other distinct self-dynamic-characters. The bubble events of worldwide housing markets and the storm of USA subprime mortgage all proved that urban housing market influence national financial system and the integrated economic development. However, our early warning system of urban housing market was not perfect, and unable to apply the theory of external-factor change and the trend observation to deal with the ever changing urban housing market. Thus, it was unable to practically apply the early warning and the strategic simulations, and resulted in risks on investment and trade of housing market.
This study, urban vision for the residential market structure analysis, quantitative measurement and strategy simulation, adopt synergy of multiple theories, multi-model integration and case study. The operation mechanism was probed by means of equilibrium theory and non-equilibrium theory of housing market, the periodic fluctuation theory and the conduction impact theory. A market forecast and early warning index system was constructed based on the systematic engineering as a core fundamental, and by applying the specialists’ advice of Fuzzy Delphi theory and the analysis of sensitivity. A dynamic model of urban housing market forecast and early warning system was constructed by integrating systematic dynamics and artificial neural networks. By means of the mechanism of monitoring indicator, the warning status of urban housing market could be judged. The forecast and the early warning of housing market and the policy optimization could be made through scenario simulation. It was found that the synergy of multiple methods could be systematized, and applied to the dynamic study of urban housing market.Complex scenario simulations always had a better policy evaluation than it of single scenario simulation. On the economic issues, the early warning measures could instinctively improve the warning status of urban housing market. On the financial issues, the early warning measures had a good effect on short term, but it has a strong sequela. It was necessary to be careful before using it.
In consideration of early warning system of urban housing market, it is suggested to set up a dynamic simulation lab of urban housing market early warning system. It is suggested to building repair strategies and sound residential real estate rental market. Thus, our government can efficiently regulate and control the urban housing market, enhance the urban competition force, and realize an optimization environment for a big city.

並列關鍵字

Systems Engineering ； Fuzzy Delphi ； Artificial Neural Network ； Urban housing market ； System Dynamics

參考文獻

李政道，2000，"台灣不動產景氣與政府政策之研究"，國立成功大學政治經濟研究所碩士論文。

曹昌勝，2007，"房地產泡沫化模型之探討"，臺灣大學國際企業學研究所碩論。

林素著、林祖嘉, 2001, "台灣地區住宅供給彈性之估計"，《住宅學會2001年年會》，pp 1-12。

謝長宏，1999，《系統概論》，華泰文化事業股份有限公司，台北。

何友鋒、王樹嵩，2004，"新竹科學城永續發展策略模擬與分析"，《設計學報》，第九卷第四期，中華民國設計學會，pp 57-76。