Question 4

1. Consider the dividends and profits time series are given to you. Since dividends depend on profits, consider the following simple model:

Dividends_t = ₁ + ₂Profits + u_t.

1. Would this regression suffers from the spurious regression phenomenon? Justify. (2 marks)
2. You have normalised the regression when testing the data. The results are as follows:

Variable	Coefficient	Std. Error	t-Statistic	Sig.
C	-13.3114	7.3626	-1.8079	0.5213
Profits	0.6281	0.0526	11.9253	0.1022
R-squared = 0.6231		d = 0.0712

Are Dividends and Profits time series cointegrated? Justify the reasons for this finding. (3 marks)

3. If, after testing, you find that they are cointegrated, would your answer in (i) change? Explain. (2 marks)
4. If Profits and Dividends are cointegrated. Define cointegration between these two variables. (3 marks)
5. Instead of regressing dividends on profits, you regress profits on dividends. Justify if such a regression valid. (2 marks)
6. If you examine the Dividends and Profits series individually, interpret either Dickey-Fuller test or Augmented Dickey-Fuller test you will use to examine stochastic or deterministic trends. (3 marks)
7. If vector autoregressive (VAR) is used for the estimation, identify the requirement for the application of VAR and interpret the approach to accomplish stationarity if Profits and Dividends are nonstationary. (2 marks)
8. To perform forecasting, lag selection would be critical. Identify the lag selection criterion and justify for the selection. (3 marks)