Variance Inflation Factor (VIF)

What Is a Variance Inflation Factor (VIF)?

Variance inflation factor (VIF) is a measure of the amount of multicollinearity in a set of different regression variables. Numerically, the VIF for a regression model variable is equivalent to the ratio of the overall model variance to the variance of a model that incorporates just that single independent variable. This ratio is calculated for every independent variable. A high VIF shows that the associated independent variable is highly collinear with different variables in the model.

Understanding a Variance Inflation Factor (VIF)

A variance inflation factor is a device to assist with distinguishing the degree of multicollinearity. A different regression is utilized when a person needs to test the effect of numerous variables on a specific outcome. The dependent variable is the outcome that is being followed up on by the independent variables — the contributions to the model. Multicollinearity exists when there is a linear relationship, or correlation, between at least one of the independent variables or data sources.

Multicollinearity makes a problem in the different regression in light of the fact that the data sources are impacting one another. Hence, they are not really independent, and it is challenging to test how much the combination of the independent variables influences the dependent variable, or outcome, inside the regression model.

In statistical terms, a numerous regression model where there is high multicollinearity will make it more hard to estimate the relationship between every one of the independent variables and the dependent variable. Small changes in the data utilized or in the structure of the model equation can deliver large and flighty changes in the estimated coefficients on the independent variables.

To guarantee the model is appropriately indicated and working accurately, there are tests that can be run for multicollinearity. Variance inflation factor is one such measuring apparatus. Utilizing variance inflation factors assists with recognizing the seriousness of any multicollinearity issues so the model can be adjusted. Variance inflation factor measures how much the behavior (variance) of an independent variable is affected, or swelled, by its association/correlation with the other independent variables.

Variance inflation factors permit a quick measure of how much a variable is adding to the standard error in the regression. At the point when huge multicollinearity issues exist, the variance inflation factor will be exceptionally large for the variables in question. After these variables are recognized, several methodologies can be utilized to dispose of or join collinear variables, settling the multicollinearity issue.

Multicollinearity

While multicollinearity doesn't reduce a model's overall predictive power, it can create estimates of the regression coefficients that are not statistically critical. As it were, it tends to be considered a sort of twofold including in the model.

At the point when at least two independent variables are closely related or measure practically exactly the same thing, then, at that point, the underlying effect that they measure is being represented two times (or more) across the variables. It becomes troublesome or difficult to say which variable is truly affecting the independent variable. This is a problem on the grounds that the goal of numerous econometric models is to test precisely this kind of statistical relationship between the independent variables and the dependent variable.

For instance, assume that an economist needs to test whether there is a statistically critical relationship between the unemployment rate (independent variable) and the inflation rate (dependent variable). Counting extra independent variables that are connected with the unemployment rate, such another initial jobless claims, would probably bring multicollinearity into the model.

The overall model could show strong, statistically adequate informative power, yet not be able to distinguish assuming that the effect is for the most part due to the unemployment rate or to the new initial jobless claims. This is the thing the VIF would recognize, and it would propose potentially dumping one of the variables out of the model or finding a workable method for uniting them to capture their joint effect contingent upon what specific hypothesis the scientist is keen on testing.

Highlights

• A variance inflation factor (VIF) gives a measure of multicollinearity among the independent variables in a different regression model.
• Distinguishing multicollinearity is important on the grounds that while multicollinearity doesn't reduce the illustrative power of the model, it decreases the statistical significance of the independent variables.
• A large variance inflation factor (VIF) on an independent variable demonstrates a highly collinear relationship to different variables that ought to be thought of or adjusted for in the structure of the model and selection of independent variables.