FRM Part I · FRM Exam Part I · Regression Diagnostics
A risk analyst's regression of portfolio returns on two highly correlated factors (sample correlation 0.97) is used only to forecast returns, using factor values whose correlation pattern is expected to persist in the future. Which statement is most accurate?
Forecasts may still be reliable, but interpreting individual coefficients is not. Multicollinearity leaves OLS unbiased and consistent and only inflates coefficient variances. If the factors' correlation pattern persists in the forecast data, predictions remain sound, while separating each factor's effect is imprecise.
- AOLS estimators are biased, so forecasts will be systematically wrong
- BForecasts may still be reliable, but individual coefficient interpretation is unreliableCorrect
- CForecasts are invalid because R-squared is overstated by collinearity
- DDropping either factor is required for the OLS estimator to be consistent
Explanation
Perfect multicollinearity aside, high correlation leaves OLS unbiased and consistent; it only inflates coefficient variances. If the correlation structure persists out of sample, the combined fit and forecasts remain reliable, though separating each factor's effect is imprecise. Dropping a variable could introduce omitted variable bias.
Did you get it right without looking?
One question tells you little. A timed set on Regression Diagnostics shows your real accuracy, how long you take and where you lose marks.
More Regression Diagnostics questions
- An analyst regresses monthly excess returns of a fund on market excess returns and finds that the residual variance is clearly larger in mon…
- A regression has an explanatory variable whose VIF is 16. By what factor is the standard error of that variable's coefficient inflated relat…
- A plot of residuals against fitted values from a regression on monthly data shows a funnel shape that widens at higher fitted values. Which …
- A regression of quarterly sales growth on an interest rate variable produces residuals with a first-order sample autocorrelation of 0.30 ove…
- In a multiple regression with two explanatory variables, the F-statistic for the joint significance of both slopes is very high, but neither…
- A risk analyst adds three irrelevant variables, uncorrelated with the dependent variable in the population, to a correct regression model th…