August 2025 Flashcards
(124 cards)
What is the BP test to detect heteroskedasticity
BP = R^2 (of Regression not adjusted R) x n with k df
What happens during serial correlation?
There are invalid coefficients and deflated standard errors and there are Type I errors
What do we use to detect multicollinearity?
There are conflicting t- and F-statistics and high VIF
SPE Rule for IFRS
Under IFRS, SPEs must be consolidated if they are conducted for the benefit of the sponsoring entity. Further, under IFRS, SPEs cannot be classified as qualifying.
If equity method is chosen over Acquisition what will be higher?
Net Profit Margin, ROE, ROA
If equity method is chosen over Acquisition what will be lower?
Sales, Expenses, Assets and Liabilities, SH Equity while NI is the same
What is the df for Regression and errors
Regression - K
Error - n-k-1
What is MSR
Mean Squared Regression - RSS/k
What is MSE
Mean Squared Error - SSE/(n-k-1)
What is SST
RSS + SSE
What is R2
RSS/SST
What is R2adj
1 - [ (n-1)/(n-k-1) x (1-R2)]
What is the probability of a log equation?
P = 1/(1 + exp(-(b0+b1X1….+)))
What is the Underwriting
Expense Ratio?
Underwriting expense/ net premiums written
What is the combined ratio?
Loss Expense & Loss Adjustment Expense Ratio + Underwriting
Expense Ratio
When you buy and sell protection with a CDS what does that mean about the underlying?
Buying protection - Underlying is worsening
Selling protection - Underlying is getting better
what is the mean reverting level?
b0/(1 − b1)
Assumptions of the multiple regression model
residuals are normally distributed and not correlated
The variance of the error terms is constant and no exact linear relationship among Xs
What is the difference between AIC and BIC?
AIC is used if the goal is a better forecast, while BIC is used if the goal is a better goodness of fit
How to test if multiple coefficients are significant
Use the F-stat where F = (SSEr - SSEu)/q SSEU/(n - k -1) where q is number of excluded variables in the restricted model and k is the number of independent variables
What happens when there is an omission of an important variable?
They are biased and inconsistent regression parameters, and may lead to correlation or heteroskedasticity in the residual
What happens when there is an inappropriate transformation?
It may lead to heteroskedasticity in the residual
What happens when inappropriate variable scaling?
It leads to multicollinearity or heteroskedasticity in the residuals
What happens when data is improprely pooled?
Lead to heteroskedaticity or serial correlation in the residuals