Professional Documents
Culture Documents
0.1884
R
2
= 0.9435
Adj. R
2
=
0.9388
eNe =
0.85163
Variable Coefficient Std. Error T-Ratio
Intercept 1.1706 0.3268 3.5823
LN(Lab) 0.6030 0.1260 4.7875
LN(K) 0.3757 0.0853 4.4022
Estimated Covariance Matrix
Intercept 0.10680
LN(Lab) -0.01984 0.01586
LN(K) 0.00189 -0.00961 0.00728
! For testing the hypothesis that the j
th
coefficient is equal to a
particular value, , the matrix R has a single row, r with a 1
0
j
in the j
th
position. Thus R(XNX)
-1
RNis the j
th
diagonal element
of the inverse matrix. The resulting F-Statistic is:
O 2
1,T-K jS j jS
F ( ) Est.Var( )
which is the square of the t-statistic with (T-K) degrees of
freedom.
! Lets test the null hypothesis that the output elasticity of labor
equals 1, H
0
: .
1
1
! The following F-Statistic is obtained:
F
1,2
= (0.6030 - 1)
2
/(0.0156) = 9.937.
The 5% critical value is 4.26 . We therefore reject the null
hypothesis.
! For testing a single linear restriction of the for H
0:
the
r* q
F-statistic is:
( )
( )
( )
2
*
2 j jS
1,T-K
* *
j k jS kS
1 1
r q
r* - q
F
r* r*
r r Est.Cov.
S
j
K K
j k
| `
. ,
where r* is (1 x K) and r
j
*
is the j
th
coefficient and the
S
where e
R
Ne
R
is the sum of squared errors from the restricted
regression which in our example is the Cobb-Douglas
production function.
! The following table provides the results of estimating the
translog production function regression model:
Estimated Translog Production Function
No. of
Obs.=27
=
u
0.1799
R
2
= 0.9549
Adj. R
2
=
0.9441
eNe =
0.67993
Variable Coefficient Std. Error T-Ratio
Intercept
0.9442 2.9108 0.3244
LN(L)
3.6136 1.5481 2.3343
LN(K)
-1.8931 1.0163 -1.8628
LN
2
(L)
-0.9641 0.7074 -1.3628
LN
2
(K)
0.0853 0.2926 0.2915
LN(L)*LN(K)
0.3124 0.4389 0.7117
! The resulting F-statistic for the null hypothesis of a Cobb-
Douglas model is:
3,21
0.85163 - 0.67993
1.768
3 (0.03238)
F
The critical value from the F table is 3.07, so we could not reject
the hypothesis of a Cobb-Douglas model.
! Note the coefficient on lnK is now negative. This does not mean
that the estimated output elasticity with respect to capital has
the wrong sign.
# 2 4 5
lnY
lnK + l nL
lnK
]
+
]
where is the full
( )
w = 0 0 1 0 ln K ln L and
S