Structural break

from Wikipedia, the free encyclopedia
Illustration of a structural break

In statistics and econometrics  , a structural break in statistical time series is a situation in which the regression parameters are not constant over the entire time series. This means that either the level parameter or the incline parameter or both parameters change.

A general (multiple) linear regression or econometric model of shape

can also be converted into a so-called structural break model by introducing a dummy variable . The time series is divided into two phases at the structural break and so each phase has its own parameters.

The Chow test is a way of testing for structural breaks.

literature

  • Peter Hackl : Introduction to Econometrics . Pearson, 2008, ISBN 9783827373380 , pp. 145ff (Chapter 9.3.1 Chow test and structural breaks)
  • Bernd Müller: Moneten with math (portrait Claudia Kirch). Bild der Wissenschaft Plus (special edition with KIT ), pp. 48–50 (online copies: [1] [2] )

Web links