10 Studenmund Using Econometrics, Sixth Edition

4-8. )2 = (Y X )2 . To find the minimum, differentiate e 2

We know that e 2i = (Yi Yi i 0 1 i i

with respect to 0 and 1 and set each derivative equal to zero (these are the normal equations):

(e 2i ) / 0 = 2[(Yi 0 1X i )] = 0

or Yi = N( 0 ) + 1 (X i )

(e 2i ) / 1 = 2[(Yi 0 1X i )X i ] = 0

or Yi X i = 0 (X i ) + 1 (X 2i )

Solve the two equations simultaneously and rearrange:

1 = [N(Yi X i ) Yi X i ]/N(X i2 ) (X i )2 ]

where x i = (X i X) and y i = (Yi Y).

0 = [X 2i Yi X i X i Yi ] /[N(X 2i ) (X i )2 ] = Y 1X

To prove linearity:

1 = x i y i /x 2i = x i (Yi Y)/X 2i
= x i Yi /x 2i x i (Y)/X 2i
= x i (Yi ) / x 2i Yx i / X 2i
= x i (Yi ) / x 2i since x i = 0
= k i Yi where k i = x i / x 2i

1 is a linear function of Y, since this is how a linear function is defined. It is also a

linear function of the s and !, which is the basic interpretation of linearity.
1 = 0 k1 + 1 k i x i + k i!1 . 0 = Y 1 (X) where Y = 0 + 1 (X), which is also a
linear equation.

To prove unbiasedness:

1 = k i Yi = k i ( 0 + 1X1 + !i )
= k i 0 + k i 1X i + k i!i

Since k i = x i / X 2i = (X i X) / X i X)2 , then k1 = 0, X 2i = 1/ X 2i , k i x i = k i X i = 1

So, i = 1 + k i!1 , and given the assumptions of !, E( 1 ) = 1 + k1E(!1 ) = 1 , proving 1 is

unbiased.

2011 Pearson Education, Inc. Publishing as Addison Wesley

 Answers to Text Exercises 11

To prove minimum variance (of all linear unbiased estimators): 1 = k i Yi . Since

k i = x i /x 2i = (X i X)/(X i X)/(X i X)2 , 1 is a weighted average of the Ys, and the ki are
the weights. To write an expression for any linear estimator, substitute wi for ki, which are also
weights but not necessarily equal to ki:

1* = wi Yi , so E( 1* ) = x i E(Yi ) = w i ( 0 + 1X1 )
= 0 w i + i w i X i

In order for 1* to be unbiased, w1 = 0 and w i X i = 1. The variance of 1* :

VAR( 1* ) = VAR w i Yi = w i VAR Yi = 2 w i2

[VAR(Yi ) = VAR(!i ) = 2 ]
= 2 (w i x i /X 2i x i /X 2i )2
= 2 (w i x i /X 2i )2 + 2 x(x 2i )2
+ 2 2 (w i x i /X 2i )(x i /X 2i )
= 2 (w i x i /X 2i )2 + 2 /(X 2i )

The last term in this equation is a constant, so the variance of 1* can be minimized only by
manipulating the first term. The first term is minimized only by letting wi = xi /Xi2 , then:

VAR( 1* ) = 2 / X 2i = VAR( 1* )

When the least-squares weights, ki, equal wi, the variance of the linear estimator 1 is
equal to the variance of the least-squares estimator, . When they are not equal,
1

VAR( 1* ) > VAR( 1 ) Q.E.D.

4-9. (a) This possibly could violate Assumption III, but its likely that the firm is so small that no
simultaneity is involved. Well cover simultaneous equations in Chapter 14.
(b) Holding constant the other independent variables, the store will sell 134.4 more frozen yogurts
per fortnight if it places an ad. If we ignore long-run effects, this means that the owner should
place the ad as long as the cost of the ad is less than the increase in profits brought about by
selling 134.4 more frozen yogurts.
(c) The result doesnt disprove the owners expectation. School is not in session during the prime
yogurt-eating summer months, so the variable might be picking up the summer time increased
demand for frozen yogurt from nonstudents.
(d) Answers will vary wildly, so perhaps its best just to make sure that all suggested variables are
time-series for 2-week periods. For students who have read Chapters 14 only, the best
answer would be any variable that measures the existence of, prices of, or advertising of local
competition. Students who have read Chapter 6 might reasonably be expected to try to find a
variable whose expected omitted-variable bias on the coefficient of C is negative. Examples
include the number of rainy days in the period or the number of college students returning
home for vacation in the period.

2011 Pearson Education, Inc. Publishing as Addison Wesley