Non-causality in Bivariate Binary Panel Data

Table 4: Causality testing

• the hypothesis H₁→_i2, concerning the поп-causality of Y¹ towards Y² is
accepted at any conventional significance level: therefore, we can accept
the hypothesis that Y¹ does not cause Y². Hence, fertility timing does
not seem to have any impact on the marriage and divorce decisions of
American women.

• the hypothesis H₁<_s>₂, concerning the simultaneous independence between
Y² and Y¹ is rejected at any conventional significance level: the same
result had been obtained when testing for restrictions in the estimated
models. Indeed, testing hypothesis H₁<_s>₂ is equivalent to testing for the
restriction of a model with simultaneous dependence to a model without
simultaneous dependence: clearly, the numerical results are hardly com-
parable, but we expect the two testing procedures to be asymptotically
equivalent.

7 An Illustrative Example for Survival Data

To show a potential application of the model developed in Section 5, we investi-
gate the causal relationship between the adoption of two technologies introduced
in the 70’s in the Italian metalworking industry. The dataset involves survival
data, namely the spell of non adoption for both technologies in a sample of Ital-
ian plants, and therefore the analysis described in Section 5 will be performed.
The two technologies considered are Computer Aided Design or Manufacturing
(CADCAM), which will be labelled by 1, and Flexible Manufacturing Systems
(FMS) which will be labelled by 2. Both technologies are originated from the
Flexible Automation (FA) paradigm and therefore they are expected to display
significative interactions.

Data on the diffusion of FA within the Italian metalworking industry are pro-
vided by the FLAUTO database, developed at Politecnico di Milano. FLAUTO
monitors adoption of FA technologies by a sample composed of 782 Italian met-
alworking plants with 10 or more employees. The sample is stratified by size
class, industry, and geographical area, so as to faithfully represent the uni-
verse of Italian metalworking plants with 10 or more employees. The dataset

21

<<   18   19   20   21   22   23   24   >>

More intriguing information