2015
DOI: 10.1016/j.jeconom.2013.10.018
|View full text |Cite
|
Sign up to set email alerts
|

Instrumental variable and variable addition based inference in predictive regressions

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

0
36
0

Year Published

2015
2015
2024
2024

Publication Types

Select...
9

Relationship

0
9

Authors

Journals

citations
Cited by 43 publications
(36 citation statements)
references
References 25 publications
0
36
0
Order By: Relevance
“…This is achieved through an IV method developed in Phillips and Magdalinos (2009) in the context of the cointegration literature and which we adapt to our current context (see also Breitung and Demetrescu (2014)). The key idea is to instrument x t with a slightly less persistent version of itself using its own innovations (hence the IVX terminology).…”
Section: Testingmentioning
confidence: 99%
“…This is achieved through an IV method developed in Phillips and Magdalinos (2009) in the context of the cointegration literature and which we adapt to our current context (see also Breitung and Demetrescu (2014)). The key idea is to instrument x t with a slightly less persistent version of itself using its own innovations (hence the IVX terminology).…”
Section: Testingmentioning
confidence: 99%
“…Closely related methods to IVX that use modifed variable addition (VA) regressions have most recently been introduced by Breitung and Demetrescu (2015) where lagged predictors are replaced by persistent time series using the Phillips-Magdalinos methodology of self-generated variables and instrumentation. These VA methods, like IVX, help remove non-pivotal inference problems when there are LUR predictors, and are similarly applicable when there are multiple predictors.…”
Section: Ivx Endogenous Instrumentationmentioning
confidence: 99%
“…Although nearly integrated asymptotics approximates the finite sample behavior of the t-statistic for no predictability considerably better when regressors are persistent, the exact degree of persistence of a given regressor, and thus the correct critical values for a predictability test, are unknown in practice. To overcome these difficulties, a number of alternative (robust) approaches have been proposed in the literature to test predictability without characterizing the stochastic properties of regressors (i.e., whether they are stationary or nearly integrated or unit root); see, for instance, Cavanagh et al (1995), 1 Campbell and Yogo (2006), Jansson and Moreira (2006), Phillips and Lee (2013), Cai and Wang (2014), Breitung and Demetrescu (2015), Kostakis, et al (2015), Demetrescu and Rodrigues (2016), and references therein.…”
Section: Introductionmentioning
confidence: 99%
“…see Demetrescu (2014) and Breitung and Demetrescu (2015) for more details on this model and the variable addition approach. Hence, an interesting question is whether the lagged variables are really econometrically needed in real applications, i.e., how to test the existence of a lagged predicted variable in a predictive regression, which apparently has not been formally addressed in predictive regressions when regressors may be nearly integrated.…”
Section: Introductionmentioning
confidence: 99%