In the mid and late 1990s there was a sharp increase in American stock prices, followed by an even sharper drop, which some economists have labelled a bubble. In this problem you will explore whether...

In the mid and late 1990s there was a sharp increase in American stock prices, followed by an even sharper drop, which some economists have labelled a bubble. In this problem you will explore whether the rising stock market might have boosted the performance of the economy (perhaps due to expectations of future wealth by investors) and, conversely, if the collapse of the market might have contributed to the recession. The cs2e3.xls file contains quarterly observations from the first quarter of 1952 to the fourth quarter of 2002 on the following two variables: rgdp: real GDP (billions of 2001 dollars); rstockval: real total value of American stocks (billions of 2001 dollars). Take the logarithms of these variables and call them lnrgdp and lnrstockval, respectively. In the subsequent tasks always use the log transformed variables. 3 a) Plot the levels and first differences of lnrgdp and lnrstockval and briefly comment on the graphs. b) Perform various unit root tests on the log transformed variables. What do you conclude? If you concluded in Part (b) above that the two variables are I(0), then let y = lnrgdp and z = lnrstockval; if you think that they are I(1), then let y = d(lnrgdp) and z = d(lnrstockval). c) Estimate a VAR(1) of y and z. Determine the optimal lag length, p, assuming that p = 8. Check the appropriateness of the selected lag structure by testing for autocorrelation in the residuals. d) Estimate a VAR(p) of y and z and briefly interpret the results. What do the significant slope coefficients suggest to you about the relationship between the two variables? e) Check the estimated AR characteristic roots and the sample autocorrelation and sample cross correlation functions. What do they imply? f) Obtain the impulse response functions for 24 periods ahead using the Cholesky decomposition with the original ordering of the variables. What do they suggest to you about the relationship between GDP and stock value? Do you expect the impulse response functions to be sensitive to the ordering of the variables? g) Obtain and comment on the variance decompositions for 24 periods ahead based on the original ordering of the variables. h) Consider the Granger Causality/Block Exogeneity Tests view of your VAR object and briefly explain the findings. i) All things considered, does it seem like the fall of the stock market could have contributed to the recession or not? Does it seem like the recession could have contributed to the fall of the stock market, or not?