08 29172 Econometrics 2021
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
A32198
08 29172
Econometrics
January Examinations 2021
Section A
1. Consider the following regression model
Yt = β0 + β1 Ut + β2 Vt + β3 Wt + β4Xt + ∈t , (1)
where U, V, W, X and Y are economic variables observed from t = 1, . . . , 75, β0 , . . . , β4 are the model parameters and ∈t is the random disturbance term satisfying the classical assumptions. Ordinary Least Squares (OLS) is used to estimate the parameters, producing the following estimated model:
t = 1.115 + 0.790Ut − 0.327Vt + 0.763Wt + 0.456Xt
(0.405) (0.178) (0.088) (0.274) (0.017)
where standard errors are given in parentheses, the R2 = 0.941, the Durbin- Watson statistic is DW = 1.907 and the residual sum of squares is RSS = 0.0757. In answering this question, use the 5% level of significance for any hypothesis tests that you are asked to perform, state clearly the null and al- ternative hypotheses that you are testing, the test statistics that you are using and interpret the decisions that you make.
(a) [10%] Describe the concepts of unbiasedness and efficiency. State the
conditions required of regression (1) in order that the OLS estimators of the model parameters possess these properties.
(b) [15%] Perform the following tests on the parameters of regression (1): (i) test whether the parameters β1 , β2 , β3 and β4 are individually statistically significant; (ii) test the overall significance of the regression model; (iii) test whether β4 is statistically equal to 0.5 against whether it is less than
0.5.
(c) [15%] Suppose you wish to test whether the economic variables U and W have the same impact on Y or if they have different impacts on Y . Express this in terms of an appropriate null and alternative hypothesis and show that if the impacts were the same then the regression model would become:
Yt = β0 + β1 Zt + β2 Vt + β4Xt + ∈t , (2)
where Zt = (Ut + Wt ). Perform the test, using the information in the fol- lowing OLS estimated regression:
t = 1.225 + 0.782Zt − 0.403Vt + 0.412Xt
(0.361) (0.147) (0.151) (0.081)
where the RSS = 0.0781 and the DW = 2.043.
(d) [10%] What are the consequences of autocorrelated errors on OLS esti- mators and discuss a way in which this can be resolved? For the model that you have chosen as a result of the test in part (c), perform a test for autocorrelation of the error term.
Section B
2. Answer the following questions:
(a) [8%] Describe the best subset selection, forward subset selection and backward subset selection methods for identifying a subset of the p pre- dictors that you believe to be related to the response, in a linear regression model. Comment on the strengths and weaknesses of the three methods.
(b) [8%] Describe the K-nearest neighbours (KNN) regression. Compare it with the multiple linear regression.
(c) [9%] (i) Discuss the problem of endogeneity and provide an example in economics. (ii) Describe the method of instrumental variables for dealing with endogeneity. (iii) Propose a possible instrumental variable for your example and discuss its validity.
3. Consider the simple linear regression model:
yi = β0 + β1 xi + ui , i = 1, ..., n,
where the xi are fixed (non-random) and the errors ui , for i = 1, ..., n are inde- pendent normal random variables with mean 0 and variance σ 2 .
(a) [5%] Derive the maximum likelihood estimators βˆ0(ML) , βˆ1(ML) and
2,ML of β0 , β1 , and σ 2 .
(b) [5%] Derive the variances Var ╱βˆ0(ML)←and Var ╱βˆ1(ML)← and prove that βˆ0(ML) and βˆ1(ML) are consistent estimators of β0 and β1 .
Suppose now that the variable x is dummy variable. Assume that n0 of the observations have xi = 0 and n1 of the observations have xi = 1, where n0 + n1 = n, the total number of observations. Furthermore, define:
y¯0 =
yi ,
and y¯1 =
yi ;
(3)
that is, y¯0 is the average of yi over all observations that have xi = 0 and y¯1 is the average of yi over all observations that have xi = 1. Show that the following are true:
(c) [10%] βˆ1(ML) = y¯1 − y¯0 .
(d) [5%] βˆ0(ML) = y¯0 .
2022-08-23