# Applied Modelling and Data Analysis ECON 1096

Coursework can be submitted any time up to 3.00pm on the 12th August 2013. Please be advised that the re-sit assessment appears on the Moodle site and can be seen by all students. However, you are only required to undertake the assessment if advised by the Progression and Award Board, usually at the end of June. Please check your Banner Student Profile on a regular basis for notification of any resits you will be required to undertake. RESIT – In-Class Test Course Leader – Larry Su Students should answer the test supplied here. A guide of two hours is suggested. Please submit your answers as you would coursework via the turnitin links.

University of Greenwich Business School

ECON1096 Applied Modelling and Data Analysis

In-class Test (Resit) Level: M Duration: 2 hours Date: August 2013 Course Leader: Larry D. Su

Instructions to Candidates: Answer all questions. Answers should be written on single-sided paper. Answers should be justi?ed by showing details where relevant.

ECON1096 In-class Test (Resit)

Larry D. Su

1. A researcher obtained the following ordinary least squares (OLS) estimates for a UK ?rm’s stock price using 120 observations from 1980 month 1 to 1989 month 12:

ln st = 0.87 -0.54 ln pt +0.65 ln yt +0.34 ln rt -0.32 ln mt (1.06) (0.24) (0.30) (0.12) (0.24) R2 = 0.34, DW = 1.65, RSS = 1.24 where (all variables in logarithms): st is the log of the stock price; pt is the log of pro?ts; yt is the log of its output in the UK; rt is the log of expenditure on research and development (R&D); mt is the log of expenditure on marketing; u is an error term. Figures in parentheses are standard errors and RSS is the Residual Sum of Squares. (1) Brie?y evaluate the reasons behind including the above explanatory variables in the regression. [10 marks] (2) Test the null hypothesis that the explanatory variables are jointly signi?cant. [10 marks] (3) If the White’s Test statistic is 18.85, does the model su?er from heterskedasticity? [10 marks] 2. Explain what autocorrelation is and how it may arise and outline the implications of autocorrelation for OLS estimation.

c

University of Greenwich Business School

1

ECON1096 In-class Test (Resit)

Larry D. Su

[20 marks] 3. Consider a model of county-wide health care expenditure: Healtht = ß1 + ß2 Incomet + ß3 Income2 t + ß4 P opt + ß5 Seniort + ut where P op is population and Senior is the percentage of population at least 65 years old. The sample size is n = 51. Regression results follow: Variable Estimate Intercept -5.06 Income 0.1045 2 Income -0.002 P op 0.847 Senior 0.385 2 R = 0.97 T SS = 56.88 Standard Error 2.094 0.025 0.005 0.489 0.164

(1) Explain the meaning of R2 in the above context and derive the adjustedR2 . [15 marks] (2) Propose a test that ?Health/?Income is constant, on average, and does not depend on income. State the null and alternative hypotheses, the test statistic, its distribution under the null, compute the test statistic, state 1%, 5% and 10% critical values and comment. [20 marks] (3) Is it preferable to have an econometric model that has as many explanatory variables as possible, or a smaller more parsimonious model? Explain some of these advantages and disadvantages. [15 marks]