What VIF value indicates multicollinearity concerns?

• A. VIF < 2
• B. VIF = 0
• C. VIF > 5 (some use > 10)
• D. VIF > 15

Answer: (C)

Variance Inflation Factor shows how much the variance of a coefficient is inflated because a predictor is correlated with other predictors. A VIF around 1 means little to no multicollinearity; the larger the VIF, the more troublesome the relationship among predictors. A commonly used rule of thumb is that multicollinearity becomes a concern when VIF exceeds about 5, with some analysts using 10 for a stronger signal. This is why the threshold of greater than 5 is the best fit here—it matches the widespread practical warning level and signals potential issues with coefficient stability and standard errors. Values well below 2 suggest negligible collinearity, and a VIF cannot be exactly zero since some inflation occurs whenever there’s any correlation with other variables. A threshold as high as 15 would indicate very severe multicollinearity, which is less typically used as the standard warning level.