Experience, Innovation and Productivity - Empirical Evidence from Italy's Slowdown

process innovations. Both large and medium-sized firms appear to be more often innovative that
small firms.

As far as our main variables of interest are concerned, the share of temporary workers does not
appear to significantly affect the decision to innovate if we split innovations between product and
process, while the average age of board members does it already at this first stage.

Table 5. Probit Maximum Likelihood for Innovation dummies (first stage).

Table 6 shows the estimates of the second stage switching regression: the Maddala method with
OLS on eq. (9’) and a 2SLS-IV method to take into account the endogeneity of the growth rate of
capital stock per worker.⁹ The instruments used for the growth rate of the capital stock prediction
are the initial level of the capital/labor ratio, age of the firm at the beginning of the sample period,
investment intensity at the beginning of the sample period, size, area, sector and group dummies.

Table 6 indicates that the method of estimation does not matter very much: the estimated
coefficients for innovative firms are in fact very similar when using OLS and 2SLS; the same
applies to the sample of non-innovative firms. It should also be noticed that the switching regression
correction is only positive and significant for the innovative firms. The estimated regressions for the
innovative firms tend to exhibit a larger R² than for the non-innovative firms (respectively 0.22 in

⁹ We run similar regressions of equations derived in Note 8, which allows to test for parameters equality between the
two regimes. The results confirm the presence of significant differences in the estimates of the coefficients, in particular
as far as age and the rate of temporary workers are concerned, between the two regimes. The table with these other
regressions are available upon requests.

20

<<   <   19   20   21   22   23   24   25   >   >>