1. Obtain a time-series of 30+ observations, for three variables of your own choosing,for any country. However, the choice of variables should be guided by an economic model from Macroeconomics. These variables should be capable of representing an economic modelwhere one variable (Y) is deemed to be dependent on the others (Xs are independent vari-ables). Provide full sources and definitions for these series.(10 marks)
Q2. Produce (separate) time-series graphs of these variables in levels and in first differ-ences. Examine these graphs for evidence of drift (trend) or unit roots.(5 marks)
Q3. Write down fully an estimating equation which represents your underlying economicmodel (Hint: it may be easier to use natural logs) in the form:lnYt=β0+β1lnX1t+β2lnX2t+ut(1)wheretis the time subscript anduis the disturbance term.(5 marks)
Q4. Perform appropriate tests for the presence of a unit root in each of the three seriesand comment on the findings.(15 marks)
Q5. Estimate your model by OLS and use this to perform a test for cointegration (DoNOTprovide a discussion of the standard significance tests for this model)(20 marks)
Q6. Compute the first difference for each time series. Perform appropriate tests for thepresence of a unit root in each of the three series in the first differences and comment on thefindings. Estimate your model in the first differences:∆ lnYt=β0+β1∆ lnX1t+β2∆ lnX2t+εt(2)(20 marks)
Q7. Given your answers in Q4-Q6, explain whether Model (1) or Model (2) is more ap-propriate. Discuss the findings in your preferred equation with respect to the sign, size andstatistical significance of the individual coefficients and the overall performance of the model2
represented by your chosen equatio
