美國公債殖利率與市場指數關聯性實證研究

An Empirical Study of Relationship between US Treasury Yield and Market Indices

延任(Jen Yen)

指導教授 ： 林建甫

國立臺灣大學/社會科學院/經濟學系/碩士(2022年)

https://doi.org/10.6342/NTU202204100

摘要

本研究欲探究日頻率尺度下各市場指數對長天期美國公債殖利率之影響，選取之市場變數包括「標普500指數」、「WTI原油價格」、「貿易加權美元指數」與「美國三個月期公債殖利率」等四個變數對於「美國十年期公債殖利率」之關係，樣本研究資料期間自2021年11月3日至2022年7月29日，共182筆日資料。
分析方式透過單根檢定、共整合分析、向量自我迴歸模型、Granger因果關係檢定、衝擊反應分析及預測誤差變異數分解等計量方法進行研究，實證結果如下：
(1) 變數均為非定態(差分後定態)，變數間不存在共整合關係，即不存在長期均衡。
(2) 向量自我迴歸模型(VAR)顯示美國十年期公債殖利率變動率與WTI原油價格變動率、貿易加權美元指數變動率呈現正相關。
(3) 考慮各變數間之關係下，WTI原油價格變動率將「Granger領先」美國十年期公債殖利率的變動率；「美國十年期公債殖利率變動率」、「貿易加權美元指數變動率」的變動會「Granger影響」WTI原油價格變動率，綜合兩者可發現WTI原油價格的變動率與美國十年期公債殖利率的變動率具相互回饋關係。衝擊反數及預測誤差變異數分解亦顯示美國十年期公債殖利率的變動率有部分可為WTI原油價格的變動率所解釋。

關鍵字

美國十年期公債殖利率 ； WTI原油價格 ； 向量自我迴歸模型 ； Granger因果關係檢定 ； 衝擊反應 ； 預測誤差變異數分解

並列摘要

This empirical research intended to study the relationship between the 10-year U.S. Treasury yield and four market indices including S&P 500 Index, West Texas Intermediate crude oil price, Nominal Broad U.S. Dollar Index, and the 3-month U.S. Treasury yield. By using unit root test, cointegration test, vector autoregression model, Granger causality test, impulse response function and variance decomposition of forecast errors. The data period is from November 3, 2021 to July 29, 2022, with a total of 182 daily observations per variable. The results of the analysis are shown as follows:
(1) By using unit root test, we find that all variables are non-stationary at level but stationary in first difference. Also by using cointegration test, we did not find any cointegration between each variable.
(2) By using vector autoregression model(VAR), the change in 10-year U.S. treasury yield is positively related to the growth rate of WTI crude oil price and also positively related to the growth rate of U.S. Dollar Index.
(3) Considering the interaction between each variable in the model, the change of WTI crude oil price will "Granger-causes" the change of 10-year U.S. treasury yield. And the changer of 10-year U.S. treasury yield and the change of nominal broad U.S. dollar index will "Granger-causes" the change of WTI crude oil price. We can see the change of WTI crude oil price and the change of 10-year U.S. treasury yield shows feedback between the two variables. We can also find that some of the variance from the change in 10-year U.S. treasury yield can be explained by the change of WTI crude oil price.

並列關鍵字

10-year U.S. treasury yield ； WTI crude oil price ； vector autoregression model ； Granger causality test ； impulse response function ； variance decomposition of forecast errors

參考文獻

田瑞駒,黃宇軒. (2022). “以七因子模型分析疫情對臺灣及美國股價報酬率之影響.” 華人經濟研究, 20(1), 33-52. (Tien,Jui-Chu and Huang,Yu-Xuan. (2022).”Implement Seven-factor model to analyze the Impact of epidemic on Taiwan and United States's stock rate of return.” Journal of Chinese Economic Research, 20(1), 33-52.).

陳旭昇. (2013). 時間序列分析: 總體經濟與財務金融之應用. 臺灣東華.

葉國俊. (2015). “公債殖利率能否作為解釋國家主權債務違約的良好指標-歐元區, OECD 與重要新興經濟體的實證研究.” 經濟論文, 43(4), 417-446. (Hsiehlin, Fang-yu and Yeh, Kuo-chun. (2022). “Government Bond Yields and The Sovereign Debt Crises: An Empirical Analysis For The Euro Area, Oecd And The Emerging Markets.” Academia Economic Papers, 43(4), 417-446.)

蕭宇翔,繆維正. (2021). “以高頻物價數據進行通膨預測.” 經濟論文叢刊, 49(3), 371-414. (Hsiao,Yu-Hsiang and Miao,Wei-Cheng. (2021). “Inflation Nowcasting Using High Frequency Price Data.” Taiwan Economic Review, 49(3), 371-414.)

Akaike, H. (1973). “Maximum likelihood identification of Gaussian autoregressive moving average models.” Biometrika, 60(2), 255-265.