Coursework, Finance and Accounting

Coursework, Finance and Accounting
Econometrics

Project description
Econometrics assignment

Topics include :

1. Properties of single variable OLS

2. Multivariate OLS

3. Hypothesis testing and Causality

4. Model Specification

5. Transformations of variables

6. Dummy variables and interactions

Notes: The critical values for t-tests and f-tests can be found from for instance this site:
http://www.statsoft.com/textbook/distribution-tables/ , or from most Econometrics books.
You have data for 15-year old British students, and their results on a reading test. ‘read’ is the score
in the reading test, ‘noroom’ is a dummy variable indicating that the student does not have a room of
her own at home, and ‘noeng’ is another dummy variable indicating that English is not spoken at
home. Variable ‘fiction’ is an integer from 1 to 5, indicating the student’s response to a question “Do
you like to read fiction?”, where response “1” indicates “not at all” and “5” indicates “Very much”,
with values 2-4 being intermediate. You estimate two models as follows. The estimated standard
errors are in parenthesis.
(Model 1) read = 499.3 – 20.8*noroom – 45.3*noeng N = 11984, R2 = 0.020
(0.9) (2.9) (3.4)
(Model 2) read = 419.7 – 18.6*noroom – 50.0*noeng + 29.4*fiction N = 11984, R2 = 0.204
(1.7) (2.6) (3.1) (0.6)
1. Using results of Model 2, interpret the coefficient for noroom, and test for the null hypothesis that
the parameter is -20 against an alternative hypothesis that it is larger (less negative) than -20, at 5%
significance level. [20 marks]
2. Are students whose home language is not English, more or less eager readers of fiction than those
who speak English at home? Please justify your answer. [25 marks]
Model 2 could still be further developed by instead of using variable fiction, creating separate
dummy variables for each of the responses (1-5) and using them in the estimation. Lets call those
variables fic1, fic2, fic3, fic4, and fic5.
3. If you would run ‘Model 3? where fiction is replaced with fic2, fic3, fic4 and fic5, what would be
the interpretation of the estimated coefficient for fic3? [20 marks]
4. What could happen to R2 in the hypothetical ‘Model 3?? Be as precise as you can. [25 marks]
5. If you would have only one variable available to you (out of noroom, noeng and fiction) for your
regression model, which one should you use to maximise the explanatory power of your model?
Justify your answer