MTest: a bootstrap test for multicollinearity

A nonparametric test based on bootstrap for detecting multicollinearity is proposed: MTest. This gives statistical support to two of the most famous methods in applied work: Klein’s rule and Variance Inflation Factor (VIF essential multicollinearity). As part procedure, MTest generates a distribution coefficient determination which: i) lets researcher assess by setting significance "alfa", or more precisely, an achieved level (ASL) given threshold, ii) using pairwise Kolmogorov-Smirnov (KS) test, establishes guide educated removal variables that are causing multicollinearity. In order show benefits MTest, procedure computationally implemented function linear regression models. tested numerical experiments match expected results. Finally, this paper makes application real data known have problems successfully detects with ASL.