透過您的圖書館登入
IP:3.146.105.194
  • 學位論文

考量結構轉變的單根檢定之模擬與應用

Structural change and Unit Root Tests: Simulations and Applications

指導教授 : 楊奕農

摘要


在過去已發表的文章中,認為變數的長期趨勢不會受到重大的經濟事件的影響。但在Nelson and Plosser於1982利用Dickey-Fuller單根檢定,發現大多數的美國總體經濟資料屬於隨機趨勢之後,開啟了學者探討總體變數特性的興趣。 本文以ADF單根檢定為基礎,嘗試探討兩種內生決定結構轉變點的方法:以最小t值認定的ZA法和以最大F值決定的Hansen(2001)。利用蒙地卡羅模擬法,模擬在有限樣本下,兩種估計方法於(1)尋找真實結構轉變點(2)在虛無假設為允許有結構轉變之單根數列之下犯型I誤差的機率和(3)檢定力三方面的表現。並將其應用在台灣的實質國民生產毛額、消費者物價指數、工業生產指數、貨幣供給、台幣兌美元匯率等五個變數上,檢視這些變數在考慮結構轉變之後的長期特性。 在蒙地卡羅模擬分析裡, Hansen(2001)犯型I誤差的機率要比ZA來得小;在檢定力方面,隨著樣本數增加,ZA與Hansen(2001)的檢定力均有向上提升的現象,當樣本數較多時, ZA的檢定力優於Hansen(2001)。此外,實際結構轉變點發生的位置並不會影響這兩種估計方法(ZA與Hansen(2001))在型I誤差和檢定力方面的表現。 當我們將這兩種尋找結構轉變的方法,實際應用在台灣的總體變數上,實證結果顯示,在進一步考慮單根檢定式中存在結構轉變,兩種估計方法所得出的結構轉變點並不一致,但若將估計的結構轉變點代入單根檢定式檢定,除了貨幣供給會因為選用的檢定模型不同,而無法判定之外,其餘變數(實質國民生產毛額、消費者物價指數、工業生產指數)不論是在何種檢定式下都不拒絕具有單根的虛無假設。

並列摘要


The traditional view holds that current shocks only have a temporary effect and that the long run movement in the series is unaltered by such shocks. Nelson and Plosser (1982) changed this view and argued, using the Dickey-Fuller unit root test, they found the most of American macroeconomic data series are stochastic trend. The dynamic characteristic about time series has being widely discussed. Because the ADF test and PP test don’t consider the impact of structural change on series, we consider the structure change as endogenous and the breakpoint is estimated not only by Zivot and Andrews (1992) but also by Hansen (2001). We investigate the accuracy of the breakpoint of those two methods and computing their empirical size and empirical power by Monte Carlo simulation and use five Taiwan macroeconomic data series to be our samples. In the Monte Carlo simulation the case of the ZA test has lager size distortions than Hansen(2001), but has a higher empirical power than Hansen(2001). As a result, the empirical size and empirical power of these two methods won’t be affected by the position of breakpoint. And the estimative breakpoints aren’t the same in two methods and except the result of money supply vary because of different models, other series didn’t against the unit-root hypothesis.

參考文獻


李政峰、何祖平(2001),「隨機趨勢抑或確定趨勢?再探台灣國民所得數列」,經濟論文,29:3,341-364。
徐之強(2001),「多次結構變動下趨勢穩定與差分穩定之認定-台灣總體資料實證研究」,經濟論文,29:3,321-339。
Bai, J. (1999), “Likelihood Ratio Tests for Multiple Structural Changes.” Journal of Econometrics, 91, 299-323.
Cheung Y.W. and M.D. Chinn (1997), “Further Investigation of Uncertain Uit Root in GNP.” Journal of Business and Economic Statistics, 15, 68-73.
Campbell, J.Y. and P. Perron (1991), “Pitfalls and opportunities: what macroeconomists should know about unit roots.” NBER Macroeconomics Annual, 141-201.

被引用紀錄


莊鈞婷(2009)。差異性商品之特徵價格與共整合關係之研究—以臺灣黃豆產業為例〔碩士論文,國立臺灣大學〕。華藝線上圖書館。https://doi.org/10.6342/NTU.2009.10430
林孝貞(2016)。中國股票市場效率檢定〔碩士論文,逢甲大學〕。華藝線上圖書館。https://doi.org/10.6341/fcu.M0320404
劉偉航(2009)。名目利率與實質利率之因果關係檢定:臺灣實證研究〔碩士論文,國立臺北大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0023-2508200900275900

延伸閱讀