ExamPrep Flashcards
(26 cards)
What does Gauss-Markov Assumption A(1) imply?
Expected value of error term is zero
What does Gauss-Markov Assumption A(2) imply?
Error term and regressors are independent
What does Gauss-Markov Assumption A(3) imply?
Heteroskedasticity
What does Gauss-Markov Assumption A(4) imply?
Autocorrelation
What does Gauss-Markov Assumption A(5) imply?
Normality (Normal Independent Distribution), a combination/replication (substitute) of Gauss-Markov A1 + A3 + A4.
What does Gauss-Markov Assumption A(7) imply?
Explanatory variables and error term uncorrelated (If correlated, then endogeneity i.e. endogenous regressor and the OLS is biased and inconsistent)
What is important to make an EGLS regression?
Log transformation
If a variable satisfies both exogeneity and relevance, what is it said to be?
An instrumental variable
When is something exogenous (1) and relevant (2)?
(1) uncorrelated with the equations error term
(2) correlated with the regressors that they are supposed to be instrumenting.
When is something exogenous (1) and relevant (2)?
(1) uncorrelated with the equations error term
(2) correlated with the regressors that they are supposed to be instrumenting.
Why do we want observations close to the fitted line (small residuals)?
The smaller, the better a test. OLS target = Reduce sum of squared residuals
Why do we want observations close to the fitted line (small residuals)?
The smaller, the better a test. OLS target = Reduce sum of squared residuals
When do we have an error term and when do we have a residual? What is characterized by these two?
Error term occur in the original model, whereas the residual in the estimated model. The first is unobservable, whereas the latter is constructed.
What does the p-value tell something about?
Denotes the marginal significance level for which the null is rejected.
What is required for OLS to be a good estimator?
Gauss-Markov
What is a regression?
A relationship between one or more variables
How do we judge the quality of a test?
The explanatory power / R^2
How to we tell if a variable is NID?
Kurtosis
Skewness
Jarque-Bera
What is a type I and II error ?
Type I: We reject null, while it is true (Prob equal to size of alpha)
Type II: We do not reject null, while it is false. Prob of not making a type II is called “Power of the test” - we prefer it to be high.
What is a type I and II error ?
Type I: We reject null, while it is true (Prob equal to size of alpha)
Type II: We do not reject null, while it is false. Prob of not making a type II is called “Power of the test” - we prefer it to be high.
How do we multiply two matrices?
N x K and K x N = The ‘N’ have to be identical?
What is N x K ?
N = rows and K = Columns
What do we get if we use log on our coefficients in terms of interpretation of these?
We get percentages
What is omitted varibles?
Relevant varibles not included
