Prof. Bryan Caplan

http://www.gmu.edu/departments/economics/bcaplan

Econ 345

Fall, 1998

HW#2

Part I: Math

Wallace and Silver, *Econometrics* – do exercises: 2.4, 2.5, 2.6

Freedman and Lane, *Mathematical Methods in Statistics*: p.58-59, #1, #3, #5

Also answer the following questions, and explain your answer:

- True or False?: When you regress a dependent variable (Y) on a constant and an independent variable (X), another way to write the formula for
is .*b* - Prove: If
=0, then . Show that in this case, R*b*^{2}=0. - Your regression errors come out to the following: {4, 3, 6, 1, -1, -3, -6, 2, -6}. You used one independent variable in the regression.

- What is
?*N* - What is (
+1)?*k* - Calculate s
.^{2}

- What is wrong with the following set of errors: {4, 3, 6, 0, -1, -3, -6, 2, -6}?

Part II: Computing

- Download the Eviews workfile hw2.bin; it contains data for for M2 (a measure of money supply), nominal GDP, and real GDP, as well as data on inflation and unemployment.
- Create a Group containing all five series, and get their descriptive statistics using both Common Sample and Individual Samples (print both). What causes the discrepancy between the two sampling methods?
- Using the Equation function, try the following regressions. Make sure you include a constant.

- real GDP on nominal GDP
- nominal GDP on M2
- unemployment on inflation

Comment on each of your results and print them. Is the coefficient positive or negative? Significant at the 5% level or not? What are your R^{2}'s?