Economics 140A *Please show all work and EViews output Spring 2016 Homework 4 1. The data set PWT63.wf1 contains data on per capita consumption (CONS) and GDP (RGDPL), for many countries for many years. You're interested in determining how much consumption rises due to an increase in output. a) Estimate a consumption function log( ) = + log( ) + where represents country and represents time. Report what you get for (in a way a nonstatistician can understand). Now allow for country fixed effects (add in @expand(country, @droplast) to your regression) instead of a single constant term so that you are now estimating log( ) = + + log( ) + . What do you get now? b) Describe (in a way that a non-statistician could understand) the theoretical differences (not the differences in your numerical output) between running your regression with country fixed effects and without. 2. Let's try a system of simultaneous equations. Download the \"HW5\" data from Gauchospace. Suppose the system is = 6 4 + + = 6 + 2 + demand supply where and are prices and quantities of a certain book over time, is the book's rating on the New York Times Bestseller list and and are random shocks. Also assume that the book's rating () only affects people's demand for the book and not the supply. a) First, try estimating the demand equation using OLS. What's your estimate for the coefficient on quantity and rating? Does this look correct? b) Second, try estimating the supply equation using OLS, what are your estimates for the coefficients on price? Do these look correct? c) Now try using instrumental variables (i.e., two-stage least squares) to estimate the supply and demand equations. Note that it may not be possible to do both. Do your answers look better than in parts (a) and (b)? page 2 3. Use the data set UWLAW2000.wf1 for this question. To be sure you have valid data, set the sample with smpl if gpa<=4 and lsat>99 and lsat<200 unchanged?)