Testing for structural breaks in GARCH Models I am looking for a package in R that can test for structural Breaks in GARCH models. I have estimated my coefficients with rugarch, and I am highly suspicious that there might be some structural break happening, however I need to test this more formally. 
Any help would be appreciated :) 
 A: How about applying processStream function from library(cpm) to residuals of ugarchfit of library(rugarch) or residuals of dccfit of library(rmgarch) will do?  
This way you will calculate structural break on residuals of univariate garch/dcc mgarch.
For a further read try:  


*

*Testing Structural Breaks in GARCH Models (D.R. Smith, 2008)  

*Real Time Detection of Structural Breaks in GARCH Models (Zhongfang He and John M. Maheu, 2009)

A: What about testing for structural breaks in the data BEFORE estimating a GARCH-model?
A good reference is this paper: 
ANDREOU (2002) DETECTING MULTIPLE BREAKS IN FINANCIAL MARKET VOLATILITY DYNAMICS
The CUMSUM type test of Inclán and Tiao (1994) is suitable for highly dependent data such as returns and implemented in the breakpoints package (the test is called segneigh.var.css there). I have not found out about the Kokoszka and Leipus (1998, 2000) test in R yet. If someone knows about this issue please let me know.
A: What about using the GAUSS code provided by Sansó et. al (2004). This method is a development of the ICSS algorithm method proposed by Incl´an and Tiao (1994). Sanso et. al (2004) is widely used to detect the structural breaks in the unconditional variance, and it has improved results when the residuals are not normally distributed. 
In order to control for the breaks detected by this method, researchers from a dummy variable for every break, that takes 1 from the date of the break and zero everywhere else. If you have many breaks try to exclude the breaks that happen in less than 63 days in between them ( a common practice). 
Potter and Dijk (2004) imposed a minimum distance restriction between
breakpoints for daily data as 63 or 126 business days (three or six months, respectively)
You can find the code here: http://www.uib.cat/depart/deaweb/personal/profesores/personalpages/andreusanso/publ_archivos/icss.zip
References: 
Incl´an, C. and Tiao, G.C. (1994). Use of cumulative sums of squares for retrospective detection of changes in variance. Journal of the American Statistic Association 89, 913-923.
Pooter, M. and Dijk, D.V. (2004) “Testing for Changes in
Volatility in Heteroskedastic Time Series – A Further Examination”
Econometric Institute Report EI 2004-38:1-39.
Sansó, A., V. Aragó & J.Ll. Carrion (2004): Testing for Changes in the Unconditional Variance of Financial Time Series. Revista de Economía Financiera. 4, 32-53. Here there are the
