Jan 2020

Question 1

OLS slope coefficient equals ratio of sample covariance of X and Y to sample variance of X

Answer 1

True

Question 2

OLS intercept is that value of X consistent with Y = 0

Answer 2

False

Question 3

For OLS to be unbiased, all 4 Gauss-Markov assumptions must hold

Answer 3

False

Question 4

The R^2 for a regression will always increase with extra variables

Answer 4

True

Question 5

The omission of a relevant variable will normally lead to bias in the remaining coefficient estimates

Answer 5

True

Question 6

In a log-linear regression equation the elasticity of Y with respect to X is calculated by multiplying the slope coefficient by X-bar/Y-bar

Answer 6

False

Question 7

Serial correlation in the residuals implies bias in coefficient estimates

Answer 7

False

Question 8

Heteroscedasticity in the residuals implies OLS is inefficient

Answer 8

True

Question 9

If Durbin-Watson statistic is < 2 it indicates negative serial correlation

Answer 9

False

Question 10

In bivariate regression the F statistic is equal to the t statistic for the slope coefficient

Answer 10

False

Question 11

Biased estimator always has higher mean-square error than unbiased

Answer 11

False

Question 12

OLS residuals are by construction uncorrelated with exogenous variables of regression equation

Answer 12

True

Question 13

The reason why OLS coefficient estimates usually follow t-distribution rather than normal is that the error variance is usually unknown

Answer 13

True

Question 14

If we wish to test hypothesis that a coefficient is positive then we use a two-tailed test

Answer 14

False

Question 15

The standard error of the regression always lies in the range 0 to 1 and the closer it is to 1 the better the fit of the model

Answer 15

False

Question 16

The F test for joint significance of rhs variables is distributed as F with degrees of freedom T - k - 1 where T is number of observations, and k is number of slope coefficients

Answer 16

True

Question 17

Serial correlation of errors in a regression model would not, in itself, lead us to expect OLS coefficient estimates to be biased

Answer 17

True

Question 18

If DW test stat lies between upper and lower critical bounds then we should reject the null that there is no serial correlation

Answer 18

False

Question 19

Heteroscedasticity is most often found in time series regression models

Answer 19

False

Question 20

Heteroscedasticity can often be dealt with by scaling the data appropriately

Answer 20

True