Most of the research papers who have conducted the Autoregressive Distributed Lag (ARDL) regression method and are published in top quality journals do not show the multicollinearity tests in their papers. However, they always conduct unit root tests before going for the ARDL results.
Is it because checking for multicollinearity is not required when unit root test is conducted for ARDL regression model? Or should we always show a multicollinearity matrix before employing the ARDL model?