Geography Reference
In-Depth Information
Table 12
.
2 Cluster i rms' performance indicators
Indicators of i rm performance
Colline Pisane
(%)
Bolgheri/Val di
Cornia (%)
Valle de
Colchagua (%)
Rating
12.5
22
60
Export
34
34
53
Good performer (Rating + Export)
34
49
72
Table 12
.
3
Results of the MRQAP: the impact of knowledge and business networks on
the good performance of i rms
Independent
(Good performer)
Colline Pisane
Bolgheri/Val di
Cornia
Valle de
Colchagua
Intercept
0.0946
0.1206
0.4522
Knowledge network
0.277***
0.156**
0.215***
Business network
0.021
−0.002
0.121
R
- square
0.029
0.010
0.041
N. permutations
2000
2000
2000
Notes:
*** signii cant at 1%; ** signii cant at 5%.
cluster, Colline Pisane, has two properties. The i rst one is the relatively high presence
of low performers (Table 12.2); the second is the poor density and high fragmentation
of its knowledge network.
12
Table 12.4 reports some descriptive statistics of this cluster's
knowledge network, showing that it has the lowest network density, if compared with the
other two clusters (0.04). More interestingly, it has the highest levels in the Hirschman/
Heri ndahl (HH) score applied to a measure both of actor centrality (
coreness
) (0.311)
and of network fragmentation (0.919 and 0.756). This indicates that the distribution of
knowledge linkages in the cluster is more uneven than in the other two clusters and the
vast majority of i rms are disconnected from the network.
The cluster of Valle de Colchagua shows a quite opposite situation. As shown by
Table 12.2 it has a higher incidence of good performers (70 per cent). It is striking that, if
compared to the other two clusters, it has a relatively higher inclusion i rms in the intra-
cluster knowledge network. This is measured by the several indicators listed in Table
12.4. In more detail, the HH indicator is the lowest among the three (0.046), indicating a
relatively less uneven distribution of knowledge linkages between cluster i rms. Similarly,
the measures of network fragmentation are the lowest among the three clusters (0.690
and 0.442). This seems consistent with the fact that the highest incidence of good per-
formers is matched with the higher participation of i rms in the intra-cluster knowledge
network.
Interestingly enough, the case of Bolgheri/Val di Cornia is in line with the results
found for Colline Pisane and Valle de Colchagua as it is situated between the two, both
in terms of the incidence of good performers (Table 12.2) and of the inclusion of i rms in
the intra-cluster knowledge network (Table 12.4).
These results seem therefore to suggest that the structural properties of these knowl-
