Il comportamento insediativo della popolazione straniera nell'Italia meridionale

Since the closing two decades of the last century dimension and features of immigration in Italy have become so considerable to affect noticeably the structure and the dynamics of total population. Through the data drawn from a recent (2005) survey on both legal and illegal foreigners attending the Italian territory ¬ carried out by Fondazione ISMU on behalf of the Welfare Ministry ¬ we will try to explain the settling behaviour of the foremost ethnic groups of foreign immigrants with regards to the geographical area of the Southern Italy provinces. For this purpose, two objectives are pursued. The first objective consists in verifying if the settling behaviour of foreign immigrants in the provinces is are deeply analyzed by means of proper statistical methods. The second objective consists in explaining the differences in settling behaviour in the light of the characteristics (e.g. the duration of the presence in the province and/or in Italy, the legal/illegal condition, the socio-economics peculiarities) of each ethnic group. In order to face the first objective of this study the concentration level of each of the 14 nationalities has been considered by means of the Gini Ratio. Then, to identify the provinces in which the relative density of each of the 14 selected nationalities is significantly different from the regional average, the location quotients have been computed. Therefore, the 14 nationalities have been classified into three groups on the basis of Gini Ratio. In the first group (GR > 0.75), the nationalities (Bangladesh, Mauritius and Sri Lanka) are highly concentrated in very few provinces of south Italy. The concentration observed for the nationalities of the second group (Ukraine, Philippines, India, Algeria, Senegal and Tunisia) is considered medium (0.60<GR<=0.75). Finally, the nationalities of the third group (Albania, China, Poland, Morocco and Romania) have a very low concentration ratio (GR<= 0.6) and show a wide spreading settling behaviour. The analysis of the settling patterns has been deepened taking the Moran Index in order to capture for each ethnic group any tendency of its relative frequency to spread out in similar/dissimilar value clusters of contiguous provinces (i.e. positive or negative spatial autocorrelation). Finally and solely for the nationalities with a statistically significant Moran Index the Local Indicator of Spatial Autocorrelation has been computed in order to provide a measure, for each nationality, of the province’s tendency to have a value that is highly correlated with values in nearby provinces. Therefore it has been possible to detect the cluster centre, and in this case to recognize the kind of cluster (hot or cold spot), the outliers and, finally, the provinces with no significant local autocorrelation. AQs regard the second objective, to describe the components that affect the settling behaviour, four groups have been identified, on the bases of settling pattern and of some social, demographic and migratory characteristics observed at the province level. Then the resulting clusters have been analyzed taking into account some extra characteristics: the nationality, the share of women, the share of illegal workers and that of people living with their relatives. Some conclusions can be drawn. In the South of Italy the settling pattern of foreign immigrants seems not to be a consequence of the developing immigration process and this may depend on the geographical level chosen for analysis of the geographical settling behaviour. With reference to this point it’s worth to underline that despite the awareness of a theoretical framework of a modifiable areal unit problem, in real data applications it is hardly possible to choose the geographical level of the analysis, as far as data are already collected. Therefore the results summarized hereafter are to be considered with awareness that the measurement of concentration and of spatial autocorrelation may vary with the geographical level of analysis. It can be argued that with a more detailed geographical level (i.e. the municipalities) the spatial variation of the relative presence of each nationality increases, so that the related Moran index I would turn out to be almost always lower than it is at the province level. On the other hand, at a lesser detailed geographical level the decreasing of spatial variation of each nationality could point out spatial autocorrelations not visible at the province level. Therefore, while it will be interesting, if possible, to examine spatial autocorrelation at a more detailed level, the results coming from a wider zoning (i.e. the region) are of no interest for the purpose of this paper. Anyway, at the actual level of analysis it has been possible to recognize well defined clusters principally on the basis of the amount of presence, both legal and illegal, and of the job specialization. Therefore, the settling behaviour of foreign immigrants can also be considered as an indirect marker of the characteristics of labour demand in each province. With reference to the illegal part of immigration, it worth noticing that such a group is generally composed of people whose duration of presence is lower than the average, as the network of relationships among countrymen already settled supports new arrivals in living arrangements and job opportunities, so as to create some sort of ethnic “enclaves”. The network strength is more visible in smaller communities than in bigger ones, where spreading people across a large territory weakens the links among countrymen.