Fin Maverick
QuarksCourses
PartnershipsShowdown
Log inSign up
Interview tracksAll
1Investment Banking
Question bankPuzzlesCase studies
2Equity Research
Question bankPuzzlesCase studies
3Venture Capital
Question bankPuzzlesCase studies
4Private Equity
Question bankPuzzlesCase studies
5Hedge Funds
Question bankPuzzlesCase studies
6Quant
Question bankPuzzlesCase studies
7Financial Analysis
Question bankPuzzlesCase studies
8Private Wealth Management
Question bankPuzzlesCase studies
9Debt Capital Markets
Question bankPuzzlesCase studies
10Risk Management
Question bankPuzzlesCase studies
11Derivatives Foundation
Question bankPuzzlesCase studies
12Portfolio Management
Question bankPuzzlesCase studies
13Mutual Fund Mastery
Question bankPuzzlesCase studies
068

Case 068Regression and model reviewWarm up

An analyst adds 12 macro variables to a 3-variable return model estimated on 120 months, and R-squared rises from 8% to 17%. Compute adjusted R-squared before and after and an F-test on the added block. Did the twelve add anything?

1The situation

A Jalvika Capital analyst has a model of a strategy's monthly returns on three factors, estimated over 120 months, with an R-squared of 8%. Hoping to explain more, the analyst adds twelve macro variables: inflation surprises, rate changes, oil, the rupee and so on. R-squared rises to 17%, and the analyst reports that the macro block more than doubles the model's explanatory power.

You are asked to review the claim before it goes into the strategy's monthly report.

2Your task

Compute adjusted R-squared for both models, run the F-test for the twelve added variables, and say whether the macro block has earned its place.

Quick check

R-squared rose from 8% to 17% when twelve variables were added to a model fitted on 120 months. What does that tell you on its own?

Worked solution

Try it on paper, then open one step at a time.

30-second answerThe answer to give first

Adjusted R-squared falls from 5.6% to 5.0%, and the F-test on the twelve added variables gives F = 0.94 against a 5% critical value of about 1.85, so the macro block adds nothing. R-squared rises mechanically with every variable; twelve noise series would have added about 9.5 points on 120 months, and these added 9. The p-value is about 0.51. Keep the three-factor model.

Step 1Why does R-squared rise even when the new variables are noise?

Give a student twelve extra guesses on a multiple-choice test and the score rises whether or not the student knows anything, because some guesses land. Least squares does the same: every added variable is fitted to the sample's quirks, so R-squared can only rise, and with n observations and k variables a noise variable removes about 1/(n - k - 1) of the remaining unexplained variance on average. Here twelve noise series would raise R-squared by roughly 12/116 of the 92% unexplained, 9.5 points. The observed rise is 9.0 points. On its face, the macro block has done exactly what random numbers would have done.

Step 2What does adjusted R-squared say?

Adjusted R-squared charges a price for each variable: 1 - (1 - R-squared) x (n - 1)/(n - k - 1). Before, it is 1 - 0.92 x 119/116 = 5.62%; after, 1 - 0.83 x 119/104 = 5.03%, lower than before, so by this measure the twelve variables cost more than they explain. The fall is small because the penalty is a crude one, but the direction is the finding. A report that quotes R-squared doubling while adjusted R-squared fell is the kind of thing a reviewer catches in one line.

ModelVariablesR-squaredAdjusted R-squaredResidual degrees of freedom
Three factors38.0%5.62%116
Plus twelve macro variables1517.0%5.03%104
Adding twelve variables lifts R-squared from 8.0% to 17.0% but lowers adjusted R-squared from 5.62% to 5.03%, because the fit improved by less than the twelve degrees of freedom it consumed.
Step 3Does the F-test agree?

The F-test asks whether the improvement in fit per added variable is larger than the unexplained variance per remaining degree of freedom. F = ((0.17 - 0.08)/12) / ((1 - 0.17)/104) = 0.00750 / 0.00798 = 0.94; under the null of no effect F averages about 1, the 5% critical value with 12 and 104 degrees of freedom is about 1.85, and the p-value is about 0.51. A result this size or larger happens half the time when the twelve variables are pure noise. For the block to clear the 5% bar, R-squared would have needed to rise by about 16 points, to 24%, not 9.

The relationship
F=(R22−R12)/q(1−R22)/(n−k2−1)=0.09/120.83/104=0.94againstF0.05(12,104)≈1.85F = \frac{(R_2^2 - R_1^2)/q}{(1 - R_2^2)/(n - k_2 - 1)} = \frac{0.09/12}{0.83/104} = 0.94 \quad\text{against}\quad F_{0.05}(12, 104) \approx 1.85
qnumber of variables added, 12
R_1^2, R_2^2R-squared before and after, 0.08 and 0.17
n - k_2 - 1residual degrees of freedom of the larger model, 104
What it says in wordsThe twelve variables improve the fit by 0.94 times what a random variable would, which is no improvement at all; a real block would score well above 1.8.
R-squared rises, adjusted R-squared falls, and the F-test says the twelve add nothing5%10%15%20%8.0%5.6%3 variables17.0%5.0%15 variablesR-squaredadjusted R-squaredadjustedfalls01235% critical value 1.85reject: the blockadds somethingobserved F = 0.94p = 0.51: as likely as not under pure noiseF-test on the 12 added variables(12 and 104 degrees of freedom)Twelve noise variables would raise R-squaredby about 9.5 points; these raised it by 9.0
Adding the twelve macro variables raises R-squared from 8.0% to 17.0% but lowers adjusted R-squared from 5.6% to 5.0%, and the F-statistic of 0.94 sits far below the 5% critical value of 1.85, exactly where twelve noise series would land.
Step 4What should the review say?

Drop the block, and say why in one sentence: on 120 months, twelve extra variables buy nine points of R-squared for free, and these bought nine. If the analyst believes one or two of the macro series matter, the honest route is to pick them before looking at the fit, add them alone, and test them on months the model has not seen. Adding a dozen and keeping whichever looks best is the same mistake with a smaller block. State the limitation: the F-test assumes independent, equal-variance errors, which monthly returns often break, so a robust version should confirm the result; but a block that cannot pass the ordinary test will not pass a stricter one.

Where candidates lose it

Candidates read a doubling of R-squared as a doubling of explanatory power. R-squared cannot fall when variables are added, and on a short sample a dozen variables add many points by chance; the only honest comparisons are adjusted R-squared and the F-test on the block.

The second miss is computing adjusted R-squared, seeing a small fall, and calling it close. The F-test puts a number on the chance that noise produced the whole rise, and that number is about one in two.

What the interviewer asks next

  • How would the answer change with 600 months of data and the same R-squared figures?
  • Why is testing the twelve jointly better than keeping the ones with t-statistics above 2?
  • What out-of-sample check would you run before accepting any macro variable?
← Case 067One-month implied volatility on an index is 30% and your desk forecasts 22% realised. You sell a delta-hedged at-the-money straddle on Rs 10 crore notional. Estimate the expected profit from the gamma-theta relationship, the result if realised volatility is 35% instead, and what can go wrong between hedges.Case 069 →An endowment can mix a risky portfolio with 8% expected excess return and 16% volatility with cash. It wants 10% volatility. What allocation does it hold, and what excess return should it expect?

Company names and figures are illustrative.

Fin Maverick

Finance education that ends in a job, not a certificate that gathers dust. Built for young India.

LEARN
CalculatorsFrameworksComparisonsInterview RoadmapsShowdown
RESOURCES
All CoursesFree CoursesBootcampsNISMQuarksAI for Finance
COMPANY
AboutJob openingPartnership
LEGAL
Privacy PolicyTerms & ConditionsContent LicenseReturn & Refund Policy
© 2026 FIN MAVERICK / BUILT FOR INDIA.DO FINANCE, DO NOT JUST READ ABOUT IT.