Wavelet methods have been successfully applied in various issues. They will surely replace gradually the classic methods of cointegration and causality.
The interest of wavalet is that they do not require stationarity and the analysis is done with nonlinear variables.Indeed, there is evidence showing that several economic variables display consistent nonlinear dependencies. But, one common feature to the papers using wavelet analysis is that they rely on univariate and bivariate frameworks. This is equivalent to a simple correlation analysis.
Do we consider the results of these assessments as biased?
No, we don't! Why ? We apply multi-variate wavelets in technical systems in many fields. And please keep in mind that wavelet react usually not like simple correlation filtes because of their special tine-frequeny-behaviour which is completely diffrent from classical windowed correlation filters.
Thank you for your comment.
I am not an expert on the subject. But from my limited experience, the results are radically different when we take into account or not control variables in the estimates;
No, they are not. Univariate analysis looks at one variable, Bivariate analysis looks at two variables and their relationship.
- Badges
- Science topic
- Similar topics
- Comment