Research projects

Project CEMAPRE internal

TitleStructural changes in the variance of time series
ParticipantsRui Louro, Nuno Sobreira (Principal Investigator)
SummaryTime series can be affected by structural changes in variance, which if unaccounted for can have
detrimental effects in the analysis and modelling steps. One paper develops an R package named
varbreaks, which contains the most common tests for detecting the presence of structural changes in
variance. Tests for single and mutiple breaks are included within the package, as well as a
procedure to deal with potential breaks in the conditional mean. A detailed description of the
various tests, computational implementation and examples are provided by this paper. Another paper
compares the finite sample properties of the most common tests to detect the presence of structural
changes in the variance of a given time series. We pay particular attention to cases where
structural changes can occur both in the conditional mean and variance processes.