Variance Inflation Factor (VIF)

Why Trust Techopedia

What is a Variance Inflation Factor (VIF)?

The variance inflation factor, or VIF,  is a statistic that helps determine the correlation between two factors. For investors, this can determine the correlation between market or company factors, like interest rates or company earnings, that might affect a company’s stock price.

Key Takeaways

• VIF helps determine if two investment factors are correlated.
• Investors can use VIF to make more informed stock price predictions.
• A high VIF means that two factors are highly correlated and should be taken into account when predicting the price of an asset.
• A low VIF means that two factors aren’t correlated and can both be independently used to predict the price of an asset.
• VIF can be useful as one part of a comprehensive financial or investment model.

How VIF Works

VIF can help you determine the correlation between two different investing factors. If two independent variables have multicollinearity it means they are actually correlated to one another, and in turn, can cause a financial model to be incorrect.

Multicollinearity can be a big issue when putting together investing or financial models because it’s important to use independent variables when trying to predict asset prices. Plus, you may have variables that you think are independent, but when further examined, are unexpectedly influencing one another.

VIF is used to determine whether variables are multicollinear and, as a result, whether a variable should be removed from a financial model.

VIF Formula and Calculation

Here is the formula for how to calculate variance inflation factor:

In this formula, Ri2 is a number that represents how well the ith predictor can be predicted by all of the other predictors in the model. This number ranges from 0 to 1, with 0 meaning the predictor is not explained by the other predictors and 1 meaning it is completely explained by them.

What VIF tells you is how correlated two different factors are in their relationship to a stock’s price.

Here is what the results of a VIF calculation can tell you:

 Result Correlation VIF = 1 Factors aren’t related VIF 1-5 Some correlation VIF > 5 High correlation VIF > 10 Very high correlation

As you can see, the higher the VIF, the more correlated multiple variables are to one another. Therefore, if you see a high VIF in your analysis, it means that you will likely have to swap one or more of your variables to ensure the quality of your investment model.

VIF Example

Imagine you are looking at a company’s earnings, sales, and industry growth to predict its future stock price. You can use VIF to determine if these factors are correlated to one another as they relate to predicting the stock price.

IF

Ri2 and VIF are high

THEN

The company’s sales can be predicted by its earnings and industry growth

In short, VIF helps you determine whether the company’s sales, earnings, and earnings growth are too closely related to one another to be considered independent investment factors. This can determine whether you should swap variables in and out of your investment model.

VIF Pros and Cons

Pros
• Helps identify the correlation between different factors
• Can help investors make more well-informed decisions
• Allows investors to refine their models to improve investment strategies
Cons
• Correlation in VIF does not necessarily indicate causation
• Does not address other issues with investment models
• Interpreting VIF can be subjective and requires interpretation

The Bottom Line

VIF can be a useful tool for investors who are putting together their own investment models or analyzing existing models. The variance inflation factor definition can help you determine whether to exclude certain variables from your investing model.

However, VIF is only one piece of a comprehensive financial model and should be used alongside other investment analysis before making investment decisions.