通貨膨脹率預測:考慮結構變動之動態因子模型應用 ; Forecasting inflation:An Application of the Dynamic Factor Model with Structural Instability

NCUIR > School of Management at National Central University > Graduate Institute of Economics > Electronic Thesis & Dissertation > Item 987654321/44218

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/44218

Title:	通貨膨脹率預測:考慮結構變動之動態因子模型應用;Forecasting inflation:An Application of the Dynamic Factor Model with Structural Instability
Authors:	宋政達;Chen-Ta Sung
Contributors:	經濟學研究所
Keywords:	通貨膨脹率;結構性變動;因子模型;factor model;inflation rate;structural change
Date:	2010-06-25
Issue Date:	2010-12-08 14:54:34 (UTC+8)
Publisher:	國立中央大學
Abstract:	本文以 Stock and Watson (2008a) 的模型方法作為基礎架構,建立一個納入結構性變動的動態因子預測模型。本研究對台灣 115 個總體經濟變數分別作 Chow 檢定以及樣本內預測,而樣本外預測則針對通貨膨脹率作未來一季至四季的預測。我們並將變數依其特性區分成生產面和金融面市場變數,以探討不同市場對通貨膨脹率的預測能力。實證結果發現,動態因子預測模型中的參數確實存在結構性變動,而動態因子模型在加入結構性變動後,在結構性變動前與結構性變動後的兩段時間內,樣本內預測的預測能力皆勝過未加入結構性變動的動態因子模型。將變數區分市場後,雖無法提高對生產面變數的預測能力,但卻可以提高對金融面變數的預測能力,並且提高動態因子模型對通貨膨脹率的樣本外預測能力。In this paper,we use a structural change dynamic factor forecasting model proposed by Stock and Watson (2008a).We focus on Taiwan's 115 macroeconomic variables employ in the Chow test and in-sample forecasting.The inflation rate for the next four quarter do out of sample forecasting.We also classify the macroeconomic variables into three kinds of markets (namely, the commodity, finance and labor markets) and discussion of the different markets predict the inflation rate.Empirical results indicate that the dynamic factor forecasting model existence of structural changes in the parameters.We find that DFM with structural changes has batter performance of in-sample forecasting than without structural changes.We also find that discriminate the market variables can improve the dynamic factor model to predict the inflation rate.
Appears in Collections:	[Graduate Institute of Economics] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	1041	View/Open

社群 sharing

Loading...