Table 6: Unconditional Convergence with Spatial Regimes, 1963—1997 (continued).
|
Sectors |
______________Transportation/Utihties______________ |
Services |
Financ |
e Insurance |
and Real Estate |
|
Models a |
OLS ______ML-HET_____ |
OLS GM-HET |
OLS |
GM-HET_______ | |
|
Low High |
Low High |
Low |
High | ||
|
Variables | |||||
|
Constant |
8.25*** 12.33*** 5.82 |
3 77*** 5 93** ɛ 92*** |
7.20*** |
13.78*** |
3.86 |
|
Log GDP 1963 |
-0.70*** -1.07*** -0.48 |
-0.33*** -0.55** -0.82*** |
-0.58*** |
-1 15*** |
-0.31 |
|
Lagged GDP growth |
0.37*** | ||||
|
Spatial ARb |
0.05*** |
0.63*** | |||
|
Convergence ratec |
3.44 INDs 1.87 |
1.14 2.28 4.90 |
2.48 |
IND |
1.06 |
|
R2 adjusted |
0.23 |
0.07 |
0.21 | ||
|
AIC |
-55.87 |
-50.11 |
-6.89 | ||
|
LIK |
29.93 |
27.05 |
5.44 | ||
|
JBd |
3.60 |
1.50 |
3.09 | ||
|
BP |
5.06 |
3.68 |
0.00 | ||
|
Chow-Waldf |
1.93 |
3.55 |
5.85*** | ||
|
I |
Q 23*** |
0.14** |
Q 4g*** | ||
|
LM-error |
4.62** |
1.75 |
21.82*** | ||
|
Robust LM-error |
0.98 |
0.28 |
10.31*** | ||
|
LM-Iag |
ɛ IQ*** |
1.50 |
13.70*** | ||
|
Robust LM-Iag |
4.45** |
0.02 |
2.20 | ||
|
LM-SARMA__________ |
9 Q7*** |
1.78_______________________________ |
24.01*** |
,' Significance at the 1, 5 and 10% level is signaled by ***, ** and *, respectively.
b Spatial Autoregressive parameter of the spatial error model.
c In percentage points per year. The convergence rate equals 100 × (ln(⅛+l))/-T, where b is the estimated coefficient for the GDP per capita level in 1963, and T
the length of the 1963—1997 time period.
dJarque Bera test for normality of the errors.
e Breusch-Pagan test with random coefficients as the alternative hypothesis.
f Chow-Wald test for stability of coefficients across regimes.
ɛ IND means the convergence rate could not be calculated because the estimated coefficient of the initial GDP per capita (in absolute value) is greater than one.