本文關(guān)鍵詞： 混頻數(shù)據(jù) 非線性格蘭杰非因果性檢驗 向量自回歸模型 典型相關(guān)分析　出處：《數(shù)量經(jīng)濟技術(shù)經(jīng)濟研究》2017年05期 　論文類型：期刊論文

【摘要】：研究目標:提出一種針對混頻數(shù)據(jù)非線性格蘭杰因果關(guān)系檢驗的方法。研究方法:在混頻向量自回歸模型(MFVAR)的基礎(chǔ)上提出構(gòu)造Wald統(tǒng)計量,通過模擬實驗和實證研究考察統(tǒng)計量的性質(zhì)。研究發(fā)現(xiàn):模擬實驗結(jié)果表明該檢驗相對于其他混頻檢驗和低頻檢驗具有更優(yōu)的性質(zhì),并且對檢驗式誤設(shè)具有穩(wěn)健性。進一步針對中國經(jīng)濟增長與消費者信心因果關(guān)系的實證研究證明,混頻檢驗與低頻檢驗得出的結(jié)論有很大差異,混頻檢驗得出的結(jié)論更符合經(jīng)濟理論。研究創(chuàng)新:利用自助法修正統(tǒng)計量在有限樣本下產(chǎn)生的水平扭曲,利用典型相關(guān)分析實現(xiàn)了數(shù)據(jù)降維。研究價值:在檢驗不同頻率變量之間的非線性格蘭杰因果關(guān)系時避免了信息損失和虛增。
[Abstract]:Research objective: to propose a method to test the nonlinear Granger causality of mixing data. Research methods: on the basis of mixing vector autoregressive model (MFVARA), we propose the construction of Wald statistics. The properties of statistics are investigated through simulation experiments and empirical studies. The results of simulation experiments show that this test has better properties than other mixing tests and low frequency tests. Furthermore, the empirical study on the causality between economic growth and consumer confidence in China proves that the conclusions of mixed frequency test and low frequency test are very different. The results of mixing test are more in line with economic theory. Research innovation: using self-help method to correct the horizontal distortion of statistics in a limited sample. Data dimension reduction is realized by canonical correlation analysis. Research value: information loss and false increase are avoided in testing nonlinear Granger causality between different frequency variables.
【作者單位】： 天津財經(jīng)大學統(tǒng)計系;天津財經(jīng)大學金融系;
【分類號】：F124.1;F126.1;F224.0

【相似文獻】

中國博士學位論文全文數(shù)據(jù)庫 前3條

1 蔡宇;混頻數(shù)據(jù)的計量經(jīng)濟學方法：理論與應用[D];山東大學;2016年

2 鄭玉航;基于混頻數(shù)據(jù)模型的中國經(jīng)濟周期特征研究[D];湖南大學;2015年

3 于揚;混頻數(shù)據(jù)回歸模型的建模理論、分析技術(shù)研究[D];東北財經(jīng)大學;2016年

，

本文編號：1517166