Prasad Krishnamurthy
Econ 140
Sept. 19, 2009
Suggested Solutions PS 1
Problem 2.6 (see textbook)
Problem 2.7
(a) We know
b
2
=
1
N
2
P
b
e
2
i
. Therefore,
P
b
e
2
i
= 2
:
04672
(51
±
2) = 100
:
2893
(b) We know the standard error of
b
2
is given by:
se
(
b
2
) =
p
:
00098 = 031305
We know that the estimated variance of
b
2
is given by:
var
(
b
2
)
=
b
2
P
(
x
i
±
x
)
2
)
X
(
x
i
±
x
)
2
=
b
2
var
(
b
2
)
=
2
:
04672
:
:
00098
= 2088
:
49
(c) The regresssion results suggest that a 1% increase in the percentage of
males above 18 who are high school graduates
is associated with
a
:
18
²
1000 =
180$ increase in the mean (annual) income of males above 18. Notice that I said
\is associated with," which faithfully represents the fact that we are reporting
a correlation. The regression estimates have very little to say about the
causal
forces that underlie the statistical association between education and income.
(d) Recall that the regression line (i.e. the equation that gives us the pre-

This
** preview**
has intentionally

**sections.**

*blurred***to view the full version.**

*Sign up*
This is the end of the preview.
Sign up
to
access the rest of the document.