Data analysis in finance:
suppose you had data on stock price indices in two different countries. you are interested in causality issues and want to find out whether movements in stock prices in one country effect stock prices in another. describe all the steps you would go through (i. e. what models you would estimate and what testing procedures you would use) to investigate your question of interest.

a. test for stationarity each series x and y
b. if x and y are nonstationary test for cointeration. regress y on x and c and test whether the residual is stationary. if the residual e=y-c-bx is stationary there is cointegrating relationship between x and y.
c. estimate var or vecm depending on whether you found nonstationarity and cointegration. if y and x are stationary use x and y in var. if y and x nonstationary and not cointegrated use var for ? y and ? x. if you found cointegration use vecm. use cointegrating term e(t-1), ? y and ? x in vecm. stationary you use var. d. test for granger casuality. the coefficients of lagged variables x in equation for y should be tested for significance in order to find out whether x granger causes y and vice versa.