2024EXAM Flashcards
(13 cards)
How do you calculate portfolio standard deviation in Excel
SQRT(MMULT(TRANSPOSE(weights),MMULT(VCOV,weights)))
How do you compute the correlation between two assets?
Correlation(i,j) = Cov(i,j) / [SD(i) * SD(j)]
What does the Lagrange multiplier on standard deviation represent in optimization?
The rate at which portfolio return increases for each unit increase in standard deviation (risk).
What is the Shrinkage method used for in financial modeling?
It adjusts the sample VCOV matrix by:
Keeping variances (diagonal) the same
Replacing covariances (off-diagonal) with a weighted average of the sample and a diagonal matrix
Formula: Shrinked VCOV = λ(Sample VCOV) + (1–λ)(Diagonal matrix)
What are the two steps in testing the CAPM using regressions?
First-pass: Estimate β for each asset by regressing returns on market index returns.
Second-pass: Regress average returns of each asset on their β to test:
Intercept = risk-free rate
Slope = market risk premium
What is the slack of the feasibility constraint if the sum of weights = 1?
Zero – the constraint is binding.
Name two good practices in financial statement modeling (besides color-coding).
Use consistent formulas across rows or columns (reduces errors)
Avoid hard-coded numbers within formulas
Excel formula to calculate portfolio mean return using weights and expected returns?
=SUMPRODUCT(weights, returns) or =MMULT(TRANSPOSE(weights), returns)
How does the Constant Correlation method modify the VCOV matrix?
Keeps variances unchanged
Sets all covariances using the formula:
Cov𝑖𝑗 = 𝜌⋅𝜎𝑖⋅𝜎𝑗
where ρ is the average correlation.
What is the formula for 1-day Delta-Normal Value-at-Risk (DNVaR) at a given confidence level?
Portfolio Value
× Portfolio Standard Deviation
× Square Root of the Time Horizon (in days)
× Absolute Value of the Z-score corresponding to the confidence level
What happens to VaR if assets are independent?
Correlation = 0 ⇒ lower portfolio standard deviation ⇒ lower VaR
Difference between Absolute VaR and Relative VaR?
Absolute VaR: loss vs. initial portfolio value
Relative VaR: loss vs. expected future value
How should a financial analyst interpret regression output?
Betas: sensitivity to market/petroleum
t-stats: significance of variables
R²: explanatory power of the model