第一次用frontier4.1 做我国分地区工业部门的sfa。
依次分别为工业增加值和固定资本净值和从业人员，当取 对他们取LN时，sfa结果为：
the final mle estimates are :
coefficient standard-error t-ratio
beta 0 0.18817519E+01 0.64976259E+00 0.28960607E+01
beta 1 0.45808708E+00 0.10302657E+00 0.44463005E+01
beta 2 0.58460205E+00 0.74754257E-01 0.78203179E+01
sigma-squared 0.72700856E+00 0.39853842E+00 0.18241869E+01
gamma 0.98598148E+00 0.86136858E-02 0.11446685E+03
mu is restricted to be zero
eta 0.10599407E+00 0.16705027E-01 0.63450403E+01
log likelihood function = 0.73701977E+02
LR test of the one-sided error = 0.17963373E+03
with number of restrictions = 2
efficiency estimates for year 1 :
firm eff.-est.
1 0.27604668E+00
2 0.30629398E+00
3 0.18969304E+00
4 0.11030116E+00
5 0.17854947E+00
6 0.25160960E+00
7 0.34870753E+00
8 0.26504581E+00
9 0.20947027E+00
10 0.18207734E+00
11 0.23131961E+00
12 0.15093425E+00
13 0.26971874E+00
14 0.18355683E+00
15 0.18624683E+00
16 0.26145344E+00
17 0.15334598E+00
18 0.17967861E+00
19 0.31132137E+00
20 0.16228664E+00
21 0.12858337E+00
22 0.23663766E+00
mean eff. in year 1 = 0.21694901E+00
等等
当取LOG形式后，计算结果为：
the final mle estimates are :
coefficient standard-error t-ratio
beta 0 -0.10477180E+01 0.96442306E+00 -0.10863676E+01
beta 1 0.13310612E+01 0.39961661E+00 0.33308457E+01
beta 2 -0.42518469E-01 0.33313654E+00 -0.12763076E+00
sigma-squared 0.11323663E-01 0.77482017E-01 0.14614569E+00
gamma 0.51459413E+00 0.89275059E+00 0.57641421E+00
mu is restricted to be zero
eta 0.12976893E+00 0.11828013E+00 0.10971321E+01
log likelihood function = 0.16630387E+03
LR test of the one-sided error = 0.10795553E+03
with number of restrictions = 2
efficiency estimates for year 1 :
firm eff.-est.
1 0.82205981E+00
2 0.89495870E+00
3 0.75627103E+00
4 0.63738031E+00
5 0.63054443E+00
6 0.81674798E+00
7 0.81225899E+00
8 0.84694301E+00
9 0.84382039E+00
10 0.84285892E+00
11 0.93148692E+00
12 0.85884530E+00
13 0.87118815E+00
14 0.81898052E+00
15 0.86562027E+00
16 0.85370373E+00
17 0.78331815E+00
18 0.91346922E+00
19 0.94078076E+00
20 0.72564801E+00
21 0.71866759E+00
22 0.70236956E+00
mean eff. in year 1 = 0.81308735E+00
请牛人帮忙，为什么取LN和LOG形式后，差距会如此之大。十分不解
扫码加好友，拉您进群
全部版块
我的主页
收藏
