本研究採用Granger and Teräsvirta (1993) 和 Teräsvirta (1994)所提出的平滑移轉模型迴歸檢定方法,自非線性(non-linear)的角度切入探討住宅價格指數對於各變數多寡不同的情境下所可能對房價產生不同的影響衝擊。樣本區間取樣自2013年1月至2020年12月之96筆月資料觀察值。以非線性平滑移轉模型探討總體經濟變數及對住宅價格指數之影響關係。並以建築貸款餘額作為門檻變數,觀察項項解釋變數在門檻值左右或偏離門檻值時,五大行庫房貸利率、上市營建類股股價、貨幣供給量、買賣移轉棟數、建築貸款餘額,會如何對住宅價格指數產生影響。 本研究首先得到研究模型為指數型平滑移轉(ESTAR)函數模型,接著,得到門檻變述建築貸款餘額數值在一兆五千九百億時,呈現顯著。進一步研究分析,發現五大行庫房貸利率及上市營建類股股價兩項變數皆不顯著。另外,貨幣供給量及建築貸款餘額,無論在門檻值左右或偏離門檻值時,都會得到顯著的正向影響。然而,買賣移轉棟數在門檻值左右時,對住宅價格指數影響為正向;但是當偏離門檻值時,則會轉變成負向的衝擊。
This paper applies the non-linear methodology of smooth transition autoregressive model (STAR) elaborated by Granger and Teräsvirta (1993) and Teräsvirta (1994) to investigate the effect of real estate price with several variables including : mortgage rate, construction stock price, M2, housing buy-sell number and construction loan., and taking construction loan as a threshold variable The sample period adopted is from 2013 Jan to 2020 Dec, which has totally 96 observations. As a result, the study gets the threshold variable at 159.023 which equals to 1.59 trillion NTD. Furthermore, it shows mortgage rate, construction loan., M2 has the same impact regardless of the threshold variable approached or nor. Most importantly, the housing buy-sell number indicates a totally opposite influence to the real estate price when it closed and deviated the threshold variable or not. Keywords: Real Estate Price, Monetary Variables, Construction Stock Price, Construction Loan, Smooth Transition Autoregressive Model (STAR)