3) [Time series - Random walk] Use the same data CONSUMP for this exercise. The permanent income hypothesis (PIH) states that individuals consume a fix portion of the permanent income (this is the expected lifetime earnings) every time period. Therefore, consumption should be equal to a constant plus a noise. That is, consumption is a random walk with drift! That is gct = ln(ct) In(Ct-1) = Be + Ut. = + a i) Test this hypothesis by running the regression 9ct Bo + B19Ct-1 + ut State the null and the alternative hypothesis of the PIH (that is, consumption is a random walk). ii) To the regression of part (i) add the variables gyt-1 and i3t-1. Are these new variables individually significant at 5% level? iii) In the regression from part ii), does the t statistic on gct-1 becomes less than 1.96 in absolute value (that is, becoming insignificant at 95%). Is now the PIH supported by the data