本研究建立資產價格危機預警模型進行Probit迴歸模型實證分析,主要變數為私部門信用放款、國內消費、國內投資、有效匯率、長天期利率、消費者物價指數。而資本帳乃代表國際間的資本移動情形,不同因子的流入與流出將影響該國發展,與資產價格的大幅波動是否有密切關係為本研究主要討論之重點。預警模型因子篩選是模型預測效率與否的關鍵,在此篇研究中,資本流入將分成三種形式來進行預警模型的分析,以資本帳餘額、資本帳資產/負債項、資本帳投資項目三種形式建立模型檢定何種變數較具備預測的效率。而此篇研究結果發現私部門信用放款、國內投資、消費者物價指數及外國直接投資為較佳的資產價格預警模型因子。資本帳因子須藉由更細分項目的方式才能提高因子的預測能力與效率。
This study mainly uses the asset price crisis warning system to do the probit regression empirical analysis. The model includes the variables such as current account, capital account, domestic claims on private sector, domestic consumption, domestic investment, effective exchange rate, long-term interest rate and CPI to the warning model. The capital flows shows the international capital flow situation. Different capital flows will affect the development of countries and often have deep relations with the asset price boom or bust. So it is highly concerned in this paper. The key to successfully predict the event is properly select the factor in the model. In this paper we divide the capital inflow variables into 3 different types to do the warning system model analysis: total capital account, capital account asset/liability and capital account investment item. The result shows that domestic claims on private sector, domestic investment, CPI and foreign direct investment (FDI) are highly relationship with asset price boom or bust. Current account can’t properly reflect the asset price boom or bust. Capital account need to breakdown by investment to meet better efficiently prediction.