Skip to content

FRM Part I · FRM Exam Part I · Nonstationary Time Series

An analyst models monthly returns of an equity index and suspects that the relationship between returns and a risk factor changed after a regulatory reform at a known date. She estimates the model separately on the pre-reform and post-reform samples and tests whether the coefficients are equal across the two samples. Which procedure is she using?

She is using a Chow test. It splits the sample at a known break date, estimates the model on each part, and tests whether the coefficients are equal. Unit-root, autocorrelation and normality tests address different questions and do not test parameter stability across a known date.

  1. AA Chow test for a structural break at a known dateCorrect
  2. BA Dickey-Fuller test for a unit root
  3. CA Ljung-Box test for residual autocorrelation
  4. DA Jarque-Bera test for normality of residuals

Explanation

A Chow test compares coefficients estimated on sub-samples split at a pre-specified break date, testing the null of parameter stability. The Dickey-Fuller test examines unit roots, Ljung-Box examines autocorrelation, and Jarque-Bera examines normality, so none of these tests parameter equality across a known date.

Did you get it right without looking?

One question tells you little. A timed set on Nonstationary Time Series shows your real accuracy, how long you take and where you lose marks.

Practise Nonstationary Time SeriesFree diagnostic mock

More Nonstationary Time Series questions