Do institutions play a role in skilled migration? The case of Italy
Download 0.62 Mb.
|
- Navigate this page:
- Definition Importance value
- Empirical investigation Data
3.2 Methodology The method adopted to obtain IQI from elementary indexes was developed in three main phases: normalization, attribution of weights, aggregation of indexes. Normalization The first step consists in normalizing the elementary indexes, that is, in reformulating each elementary index so that it is measurable in the range 
with  where  and  are respectively the minimum and maximum value assumed by the i-th index in the various provinces and  .
Weight assignment The second phase envisages assignment of a weight for each normalized index. To do this, we use the Analytic Hierarchy Process (henceforth AHP) elaborated by Saaty (1980, 1992) and widely employed in multiple criteria decision-making in different fields such as geographical sciences, land planning and resource allocations9. The AHP method hinges on a predetermined multi-layer framework creating a hierarchy among elementary and aggregate indexes (see Figure 1). To derive a weight for each index of a given rank, representing the importance of that index in determining the index of the upper level, AHP starts from pairwise comparisons between indexes of the same layer based on verbal judgments on the relative importance of each index10. These judgments are then turned into numerical values of importance. According to the following Saaty relative numerical scale:
All the comparisons made between the importance values of the elementary indexes of each dimension are reported in matrix : 
By construction, matrix is a square matrix with all the elements of the diagonal equal to 1: Using and  to denote, respectively, the eigenvalue and the maximum eigenvalue of matrix , if comparisons of importance are fully consistent, then  , otherwise  . Saaty (1980, 1992) proposed a consistency index  which assumes the value of zero in the event of maximum consistency, and positive values otherwise. According to Saaty the threshold value which defines adequate consistency of determinations is  . Application of this procedure to our elementary indexes allows us to obtain weights of the elementary indexes such that  and the weights of dimensions  such that  (Table 2).
Insert Table 2 Aggregation Having defined the weights of each index, the aggregation function used to determine the index of institutions is given by: 
where IQI is the index of provincial institutions which, by construction, assumes values in the range Insert Table 3 Insert Figure 2 Sensitivity analysis Finally, the results of a sensitivity analysis carried out to test the response of the institutions’ index to alternative scenarios are reported in the Appendix. The main issue in the methods of constructing synthetic indexes is known to lie in the subjectivity of the criterion for assigning weights which is affected by the degree of importance assigned to the individual dimension. For this reason we carried out a sensitivity analysis of the IQI with which we tested its performance in eight different scenarios obtained by assigning different weights to different dimensions. The analysis shows that the IQI is, among the various scenarios reproduced, the one that comes closest to intermediate values of variability and one of the best scenario able to minimize the distance from the mean for each of the 107 provinces. Empirical investigation Data The chief source of statistical information which we used was the ISTAT survey on the professional recruitment of graduates in 2004 (ISTAT, 2009)12. The analysis focuses on individual movements implying change of residence between Italian provinces13.To supplement this dataset, we used the provincial data for per capita GDP and the unemployment rate for 25-34 year olds (sources: ISTAT: Conti Economici Regionali 1995-2009 and Istituto Tagliacarne: Atlante di competitività delle provincie italiane 2001) as well as information on the elementary indexes used to construct the IQI indicator (the sources are reported in Table 1). Prior to examining the econometric estimates, let us first focus on the inter-regional net migration balance for graduates, calculated on the basis of the sample in question. Table 4 represents individual movements implying change of residence between Italian regions. Hence, Italian graduates are classified on the basis of comparison between their region of residence in 2004 (i.e. immediately after graduation) and their region of residence in 2007 (i.e. after three years from graduation). As reported in Table 4 both for the regions and the macro areas of Central and Northern Italy, the Mezzogiorno and abroad, Emilia Romagna, Lazio and Lombardy show the highest positive net migration balance (31%, 22% and 16% respectively), thereby proving to be the most attractive regions for young graduates. By contrast, Basilicata (-37%), Puglia (-36%) and Calabria (-31%) are the regions with the greatest loss of young graduates for the period 2004-2007. Significantly, Figure 3, which measures migration balances for 2004-2007 against the values of the indicator IQI14, highlights a marked positive correlation between the two dimensions. Insert Table 4 As regards the macro areas, ISTAT (2009) reports that about 25% of the graduates transferred their place of residence from the South to a region in Italy’s Centre-North or abroad, while there was a positive balance (+11%) in the Centre-North and very positive (+130%) abroad. Also the data on pre-degree migration (i.e. of high-school leavers) confirm the negative trend in the Mezzogiorno: 3,546 students with residency in 2004 in a region in southern Italy graduated at a university in the Centre-North. This means that about 22% of southern Italian high-school leavers who embark on a degree course migrate to a region in the Centre-North to continue their university studies. Insert Figure 3 Directory: uploads uploads -> 587 Return function, r i(X) r i(0) r i(1) r i(2) r i(3) 1 0 2 4 6 Thermal Station, I 2 0 1 5 6 3 0 3 5 6 10 uploads -> Department of science and humanities department of mathematics part-a questions and answers uploads -> The Case Against the nypd’s Quota-Driven ‘Broken Windows’ Policing What Is It? uploads -> For want of a nail uploads -> Ownership Models There are two different types of ownership models uploads -> Police Militarization uploads -> City of rising star regular meting thursday, july 14, 2016 uploads -> Township of winfield uploads -> Stop Militarizing Law Enforcement Act of 2014 uploads -> The Black Panther Party’s Ten-Point Program What We Want Now! Download 0.62 Mb. Share with your friends:
|
. The method we use in this paper is that of the distance from the ideal point. Using to denote the i-th elementary index for the j-th province, the corresponding 
(for ) and 
(for 
. Saaty showed that a weight can be estimated for each elementary index. By calculating the eigenvalues of matrix , considering the eigenvector associated to the maximum eigenvalue, and setting the constraint under which the