|
Empindsh |
-.008 |
-.011 |
-.007 |
-.009 |
|
(0.55) |
(0.73) |
(0.50) |
(0.59) | |
|
PARPYME |
-8*10-4 |
-.002 |
-.003 |
.001 |
|
(0.02) |
(0.06) |
(0.09) |
(0.03) | |
|
IDPIB |
-.195 |
-.297 |
-.174 |
-.209 |
|
(0.52) |
(0.78) |
(0.46) |
(0.55) | |
|
Rho |
.036 |
.005 |
.043 |
.022 |
|
Sigma |
.341 |
.341 |
.399 |
.337 |
|
Lambda |
.012 |
.002 |
.017 |
.007 |
|
Log likelihood |
-687.235 |
-695.461 |
-870.594 |
676.074 |
|
LR test(1) |
.06 |
.00 |
.05 |
.02 |
|
Test for β1 = 1(2 |
49.3 |
9.73 |
39.54 |
14.02 |
|
N° observations |
1255 |
1252 |
1255 |
1255 |
|
Censored |
94 |
96 |
94 |
94 |
|
Uncensored |
1161 |
1156 |
1161 |
1161 |
*** Significant at 99%; ** Significant at 95%; * Significant at 90%..
t- values in parentheses.
(1) LR test of independent equations (rho=0). Chi-square(1).
(2) Chi-square (1)
Gibrat’s law is rejected independently of the employment definition. The estimator associated
to the logarithm of employment in 1998 is always smaller than 1, meaning that small firms
have grown bigger than large ones. At the same time, process innovation is also significant,
supporting the assumption of a positive relationship between innovation and growth. Firm’s
age is also important in the growth equation, been negative the sign of its estimator.
From the point of view of the survival equation, innovation, both process and product, play a
positive role. And standardisation of the product has associated a negative relationship.
Therefore, we can affirm that innovating firms with a differentiate product have a bigger
probability of survival.
Finally, it does not look like regional variables play an important role in the survival and
growth of Spanish firms during the period analysed. Only the share of industrial employment
is significant in three of the four equations.
20