Skip to content

FRM Part I · FRM Exam Part I · Regression with Multiple Explanatory Variables

A risk analyst's regression of fund returns on two highly correlated factors (correlation 0.95) gives unstable coefficients. The analyst's goal is solely to forecast fund returns using the same relationship in the same data environment, not to interpret individual factor effects. Which statement is most accurate?

Multicollinearity matters little for pure forecasting. OLS estimates stay unbiased and the fitted relationship still predicts well, provided the correlation pattern between the regressors persists. The damage is to the precision and interpretation of individual coefficients, not to overall predictive fit.

  1. AMulticollinearity is of limited concern, as the fitted values and forecasts remain reliable if the correlation structure persistsCorrect
  2. BThe OLS coefficient estimates are biased, so forecasts will be systematically wrong
  3. CForecasts are invalid because OLS is no longer the best linear unbiased estimator
  4. DThe model must be corrected using generalized least squares before any forecasting

Explanation

Perfect multicollinearity aside, high correlation leaves OLS unbiased and still BLUE; it only inflates the variance of individual coefficients. If the correlation pattern persists, predictions remain reliable. Bias and GLS are not implied by multicollinearity.

Did you get it right without looking?

One question tells you little. A timed set on Regression with Multiple Explanatory Variables shows your real accuracy, how long you take and where you lose marks.

Practise Regression with Multiple Explanatory VariablesFree diagnostic mock

More Regression with Multiple Explanatory Variables questions