Homework 1
Each student will need to:
• estimate 3 variants of a model of consumption of a given commodity,
• prepare a printout with regression results (do not include descriptions, interpretations etc in the printout!), using the G7 “r” command result format (just copy the results from the G7 screen),
• be prepared to shortly comment on the results,
• be able to interpret regression coefficients and the basic statistics,
• model variants must include a linear model and a log-linear model.
The estimation will be done in the G7 software (PDG), using pce data (pce.txt). Both the program and the data are available at www.inforum.uni.lodz.pl.
Hints:
• Use real income and consumption in your models, not the nominal ones!
• If you include prices into your equations, these have to be relative prices, e.g. ppce1/cpi (this is the price of rice relative to the “average price level”), or ppce1/ppce5 (this is the price of rice relative to the price of pasta) etc.
• To estimate a log-linear model you should type for example:
“r @log(pcer1)=@log(yr),@log(ppce1/cpi)”
Please learn the interpretation of regression coefficients (elasticities!) and other statistics for log-linear models.
• If you find an explanatory variable insignificant you might want to exclude this variable in another variant of the model.
• You should prepare critical values for the t-Student’s and the Durbin-Watson test.
• Use G7 help to learn the syntax of G7 commands. G7 reference can also be found at http://www.inforum.umd.edu/papers/inforum/software/GREF.pdf
Warning – not preparing the homework or absence in the next class will lead to an additional homework with much extra effort!