Correlation and the time interval over which the variables are measured

Haim Levy*, Gideon Schwarz

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

22 Scopus citations

Abstract

When two random variables are multiplicative over time, their correlation coefficient is not invariant under changes of the differencing interval even when each of the random variables is a product of i.i.d. variables over time. It is shown that unless Y = kX, k > 0, the coefficient of determination (ρ2) decreases monotonically as the differencing interval increases, approaching zero in the limit. In sampling for empirical studies, the differencing interval is often selected arbitrarily. Such a choice may dramatically affect the sample correlation coefficient, as well as its statistical significance.

Original languageEnglish
Pages (from-to)341-350
Number of pages10
JournalJournal of Econometrics
Volume76
Issue number1-2
DOIs
StatePublished - 1997

Keywords

  • Correlation coefficient
  • Differencing interval
  • Investment horizon
  • Stochastic processes

Fingerprint

Dive into the research topics of 'Correlation and the time interval over which the variables are measured'. Together they form a unique fingerprint.

Cite this