关键词 > ECMT6002/6702
ECMT 6002/6702: Econometric Applications Week 10
发布时间:2023-06-19
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
ECMT 6002/6702: Econometric Applications
1 Practice problems (and empirical application)
We will consider the housing pricing example given in Wooldridge’s textbook and compare the
following two competing housing price models:
(1) yt = β 1 + β2 x2t + β3 x3t + β3 x4t + ut .
(2) yt = β 1 + β2 log x2t + +β3 log x3t + β3 log x4t + ut .
price log lot size log square ft . log # of beds
We will implement the RESET and Davidson-MacKinnon tests to compare the above models.
Instructions:
1. Note that the questions to be given below can be answered without computing the required test statistics, and you are strongly recommended to think about those.
2. Consider the model (1). Implement the RESET test by including and
. Compute the relevant critical values and report the test results at 95% significance level.
- To implement this, you first need to obtain the OLS fitted values t from the model (1). Then consider the relevant RESET regression model with
and
. The computed F and Wald test statistics must be similar to
F = 4.7 and W = 10.0. (1.1)
- Each of the test statistics will be greater than the relevant critical value.
- Q: How can we interpret this result?
3. Implement the same RESET test for the model (2). Report the results.
- The computed F and Wald test statistics must be similar to
F = 17.67 and W = 37.47. (1.2)
- Each of the test statistics will be greater than the relevant critical value.
- Q: How can we interpret this result?
4. Implement the Davidson-MacKinnon test for the model (2) against the model (1) at 5% significance level.
- The coefficient of t must be close to 0.810 and the standard error must be close to 0.373. The null hypothesis of the Davidson-MacKinnon test is rejected.
- Q: What this test result means?
5. Implement the Davidson-MacKinnon test for the model (1) against the model (2) at 5% significance level.
- The coefficient of t must be close to 0.526 and the standard error must be close to 0.310. The null hypothesis of the Davidson-MacKinnon test is not rejected (at 5% significance level).
- Q: What this test result means?
Q: Combining all the above results, which model is better?
6. Check if how the above results change if yt is replaced by log yt .