Testing for Structural Breaks in Dynamic Factor Models

By J. Breitung, Sandra Eickmeier. Social Science Research Network, 2011

229 citations recorded by the source corpus.

In this paper we investigate the consequences of structural breaks in the factor loadings for the specification and estimation of factor models based on principal components and suggest procedures for testing for structural breaks. It is shown that structural breaks severely inflate the number of factors identified by the usual information criteria. The hypothesis of a structural break is tested by using LR, LM and Wald statistics. The LM test (which performs best in our Monte Carlo simulations) is generalized to test for structural breaks in factor models where the break date is unknown and the common factors and idiosyncratic components are serially correlated. The proposed test procedures are applied to datasets from the US and the euro area.

Subfields: Economics and Econometrics, Finance

View Paper: https://www.semanticscholar.org/paper/beb9b978ec922e0b6e1eea188506c6c9a038d147

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Economics and Econometrics

In Finance

Citations in the corpus

Do Euro Area Countries Respond Asymmetrically to the Common Monetary Policy (2014, unknown): “On the one hand, Breitung and Eickmeier (2011) prove that in presence of a structural breaks, factor loadings λi may be inconsistently estimated. On the other hand, Stock and Watson (2002) demonstrate that the space spanned by the common factors ft can still be estimated consistently if there is lim”
GLS Estimation of Dynamic Factor Models (2011, methodology): “For example, recent work by Breitung and Eickmeier (2011) shows that a Chow-type test for structural breaks can be derived using the iterated PC–GLS estimator.”
Testing for Structural Stability of Factor Augmented Forecasting Models (2014, background): “Breitung and Eickmeier (2011) propose tests for the null hypothesis of factor loading coefficient stability.”
TESTS FOR PARAMETER INSTABILITY IN DYNAMIC FACTOR MODELS (2013, methodology): “Breitung and Eickmeier (2011) also suggest a pooled LM test for testing (2.3), but their pooled test requires that the idiosyncratic shocks eit and ejt are independent for all i = j , which is too restrictive compared with the approximate factor structure.”
Testing for Factor Loading Structural Change under Common Breaks (2015, methodology): “…paper casts new light on this nonmonotonic power problem in considering the factor loading structural change tests recently proposed by Breitung and Eickmeier (2011) (hereafter, BE tests) in the context of dynamic factor models.3 Our focus is on the case in which the factor loadings…”
Dynamic Factor Models, Factor-Augmented Vector Autoregressions, and Structural Vector Autoregressions in Macroeconomics (2016, background): “1984q4 (Stock andWatson, 2009; Breitung and Eickmeier, 2011, Section 3). The Quandt likelihood ratio (QLR) version allows for an unknown break date and is the maximum value of the Chow statistic (the sup-Wald statistic) for potential breaks in the central 70% of the sample, see Breitung and Eickmeie”
On Time-Varying Factor Models: Estimation and Testing ∗ (2017, methodology): “’s (2014) sup-LM and sup-Wald tests, and Breitung and Eickmeier’s (2011) have lower power than our test too. In particular, these tests have little or low power in detecting deviations from the null in DGPs 5 and 7 but reasonable power against DGPs 6 and 8. It is easy to explain why some of these ot”
Simultaneous multiple change-point and factor analysis for high-dimensional time series (2016, unknown): “Investigations into the effect of a single break in the loadings or the number of factors on the factor structure, with accompanying change-point tests and estimators for its location, can be found in Breitung and Eickmeier (2011), Chen et al.”
State-Varying Factor Models of Large Dimensions (2018, methodology): “They have been successfully used in finance, for example, Ross (1976), Chamberlain and Rothschild (1983), and Ludvigson and Ng (2009), and in macro-economics, for example, Stock and Watson (2002) and Jurado, Ludvigson, and Ng (2015).”
Macroeconomic forecast accuracy in a data‐rich environment (2019, unknown): “The number (7)See, among others, Breitung and Eickmeier (2011), D’Agostino et al. (2013), Eickmeier et al. (2015), Cheng et al.”
DYNAMIC FACTOR MODELS (2018, methodology): “Breitung and Eickmeier (2009) find evidence of a change in the factor space in the U.”
Identification and estimation of a large factor model with structural instability (2017, unknown): “Plug in X = GΓ ′ + E, we have G̃ − GJ = 1 NT (GΓ ′E′G̃+ EΓG′G̃+ EE′G̃)V −1 NT and g̃t − J gt = V −1 NT ( 1 T ∑T s=1 g̃sγN (s, t) + 1 T ∑T s=1 g̃sζst + 1 T ∑T s=1 g̃sηst + 1 T ∑T s=1 g̃sξst), (15) where ζst = eset N − γN (s, t), ηst = g′ sΓ ′et N and ξst = g′ tΓ ′es N .”
Estimation of large dimensional factor models with an unknown number of breaks (2018, background): “2015), Lam and Yao (2012), Onatski (2012) and Wang and Fan (2017) for recent advancement in estimation methods and inference theory in large-dimensional factor modeling.”
Consistent Factor Estimation in Dynamic Factor Models with Structural Instability (2013, background): “…nearly all of this theory has been developed for the case in which the DFM parameters are stable, in particular, in which there are no changes in the factor loadings (the coefficients on the factors); among the few exceptions are Stock and Watson (2002, 2009) and Breitung and Eickmeier (2011).”
Shrinkage estimation of common breaks in panel data models via adaptive group fused Lasso (2016, unknown): “, Breitung and Eickmeier (2011) and Cheng et al. (2014)) and this further complicates the analysis.”
Detecting big structural breaks in large factor models (2014, methodology): “Further, given the assumptions we make about the et error terms, the case analyzed by Breitung and Eickmeier (2010) where the eit disturbances are generated by individual specific AR(pi) processes is also considered.”