Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The Structural Equation Model in the University Research System

The Structural Equation Model in the University Research System In this paper we want to show the propensity to return to Italy of a sample of Italian academics who have emigrated abroad and we provide new empirical evidence through structural equation models (SEM). With SEM we have obtained more information about the relations between the variables and the propensity to return. Moreover, the estimate of the endogenous variables used to identify significant factors as contributing to generate indirect influences on the propensity to return, via the direct relationship with the exogenous factors. Keywords: structural equation model, propensity to return, brain drain, academics. JEL Classification: F22, J24, O15. 1. Introduction The term drain researcher is not used in literature before with it we want to specify that the sample used in the paper regards researchers (assistant professors) and professors in various universities worldwide. This means that we introduce a sub category of a wider term like brain drain. The present paper aims to elaborate an empirical model to identify the main factors leading to the Italian brain drain (or drain researcher), assesses the propensity to return of highly qualified Italian emigrants. This goal is achieved by means of a sample survey. Interviewees were selected at random among Italian graduates, doctoral students, assistant professors and academics from those who emigrated from Italy. The paper shows the results of empirical analysis: the model considers the propensity to return to Italy of Italians abroad. Our empirical analysis envisaged the use of multivariate and non-parametric statistical model. The model used is structural equation model (SEM). This model produces the relationship between the different variables that affect the propensity to return to Italy. SEM methodology makes to go through the limits of general regression models (that measures just the direct relationship) because it allows measuring the indirect, indirect and latent relationships. A review of the literature shows that the return migration to Europe and other regions, such as Asia, is far from marginal. Many emigrants return to their home country and an even greater proportion is highly qualified. Recent studies considered the migration of qualified agents as transitory (Batista et al. 2007, Gundel and Peters 2008, Mayr and Peri, 2008, Dustmann and Weiss, 2007). Micro-data analysis allows us to overcome the informational limits imposed by the use of macro-data (Brandi, 2001; Piras 2005; Lacuesta 2006, Cattaneo 2009). The Likert scales used in the survey provide an indepth view of the attitude of interviewees to return to Italy than would have been obtained using dichotomous variables (Dustmann 1996 e 2007). There are relatively few empirical studies of the brain drain in Italy (Avveduto and Brandi, 2004; Becker, Ichino and Peri, 2004; Brandi and Cerbara, 2004; Gagliarducci et al., 2005; Brandi and Segnana, 2008). However, little importance has been attached to distinguish between the permanent or transitory nature of highly qualified emigration. The paper is organized as follows. Section 2 presents the data set and the methodology. The main results are described in section 3, while section 4 shows the model. Section 5 presents some policy implications and conclusion remarks. 2. Data set and methods Recently, a number of authors have undertaken the study of the stock of skilled workers in different countries of origin with a view to obtaining information on the brain drain (Carrington and Detragiache, 1998; Docquier and Marfouk, 2006; Doquier Lowell and Marfouk, 2009; Monteleone, 2011). Doquier and Rapoport (2009) assess the overall impact the brain drain has upon the countries of origin, evaluating the costs and benefits of such migration for developing countries both in macro- and micro-economic terms. Micro-economic analysis appears to offer the more interesting field of study. Assessing the brain drain and testing hypotheses Journal of Advanced Research in Management through micro-data have been scantily covered in the literature, at least as far as Italy is concerned (BrandiCerbana, 2004; Becker, Ichino and Peri, 2004). However, while there is undeniably a rich theoretical literature, the empirical literature is scarce (Torrisi, Skonieczny, 2009 and 2011; Biondo et al., 2012). This paper continues the development of results obtained in earlier surveys in 2009 and then expanded again in 20101. It is based on data sets of micro-data: 1400 individual on-line interviews of Italian researchers abroad. The data set is based on a sample of 1400 contacts among Italian PhD researchers (assistant professors) and Italian professors in various universities worldwide. A data platform is developed in relation to the participation and involvement of a sample of Italian immigrant researchers in countries with strong research appeal: Canada, Germany, France, Switzerland and Australia. Out of 1400 individual contacts only 68% (955/1400) responded in full. The sample of respondents comprises individuals who are highly educated in different fields of scientific research or highly skilled workers. The lack of official statistics or surveys on the size of population, did not allow any estimate of the number to be sampled and primarily of a criterion of selection of units. The 1400 contacts (Source: From the Catania University StatEcon2 database - Year 2010) are classified in 60 variables into the following macro areas of interest (see Monteleone, Skonieczny and Torrisi, 2010, 2012 and Biondo A. et al., 2012 for a detailed description). Each variable was analyzed according to different scales of measurement on a case-by-case basis. For the most part, the study uses Likert scales, while for some variables it was necessary to associate ordinal, nominal and interval-based scales3. The SEM instruments are able to provide adequate representations of causal relations, both in terms of the total impact, that part of the independent variables on employees. A regression equation can be interpreted as a causal model where the different coefficients give some support to the causal relationship between the variables, but the causal nature of this model is pragmatic and does not have theoretical support about the internal mechanisms of influence between variables (the regressive models do not estimate the underlying or causal relationships between variables). To assess the report in detail, splitting the total direct and indirect relationship, we used a SEM attempting, through a set of equations, to reproduce the structure of causal processes between variables. The methods used for the construction of structural models are based on models of "recursive type". 3. Descriptive results Descriptive analysis generated the following results: the subjects who leave Italy to go to another country that can offer them better living and job conditions. Respondents' preferred destinations are Britain, Switzerland and Canada (see Table 1). Table 1. Percentage of respondents by host country4 Nh Wh USA 749.7 35.7% UK 504 24.0% Canada 281.4 13.4% France 170.1 8.1% Spain 94.5 4.5% Germany 90.3 4.3% Switzerland 79.8 3.8% Austria 42 2.0% Holland 31.5 1.5% Other 42 2.0% Africa 14.7 0.7% Total 2100 100,0% Source: Stat Econ database ­ Updated 2010.12.31. nh 500 336 188 113 63 60 53 28 21 28 10 1400 nh complete 376 221 144 66 43 33 25 15 13 13 6 955 Skonieczny G., Torrisi B. (2009 - 2011), Monteleone S., Torrisi B. (2011 - 2012), Torrisi B. (2012), Biondo A. et al. (2012). StatEcon is the economic statistics science of the Department of Enterprise Culture and Society, University of Catania. 3The Likert scale measures attitudes. 4 Where Nh is the size of h-th layer and H the number of layers that should be considered with h=1,...,H e h Nh = N ; nh ( h nh = n ) is the sample size in the generic layer h; nh completi is the number of completed questionnaires that were returned for each layer and reclassified. Issue 2(6) Volume III Winter 2012 The results following the prevailing literature have identified these countries as the most capable of attracting workers, especially highly qualified workers. The position of researcher or professor is correlated with the level of basic knowledge obtained in Italy or with working experience in Italy. This result is what affects the experience developed in Italy in relation to basic training ever developed in Italy. This means that the preparation affects the current academic status in the host country, unlike the academic experience had in Italy by the interviewees. The analysis shows that those individuals who leave Italy are well-informed about research in Italy; individuals who have had working experience in Italy before leaving the country demonstrate significant understanding (p-value=0,0034) of how research is financed. The perception is that those who have had work experience in Italy consider that the system is not meritocratic for access to research funding in Italy. A fundamental aspect of the survey is to decide how the host countries perceive the career of individuals engaged in research and what mechanisms govern career progression. A clear majority of researchers (84.45% as the sum of absolutely fairly meritocratic and meritocratic) confirmed that career progress is judged as significantly meritocratic (p-value = 0.0000). These results represent two important elements: 1. expressed opinions about how host country considers the state of meritocratic career progression, 2. that this opinion is shared among all countries surveyed with the exception of some European countries. A significant emigration aspect concerns the relationship between age and career progress. At the age of 30 years, subjects go abroad to become researchers; older agents become teachers. The targets for young migrants are significantly age-correlated. Contrary to the prevalent thrust of the literature which sees recent migration as a transitory phenomenon, the results of our analysis show that in Italy it is permanent. This result is obtained by evaluating the emigrants' propensity to return. This degree of propensity is assessed on the basis of the percentage of responses given in relation to a scale of evaluations designed to highlight the subjects' attitude to the idea of returning to their home country. The 73.3% of respondents have a low, or no, propensity to return to Italy (see Figure1). higt 6,5% 41,2% nothing 20,2% average 32,1% low Figure 1. Percentage distribution relative to propensity to return Source: Stat Econ database ­ Updated 2010.12.31. 4. The model results We considered a database of 955 lines (statistical units) and 60 rows (variables). In relation to the complexity of variables, the propensity to return was studied through multivariate analysis models. For the type of variables, we applied different regressive models (OLS, GLM and LISREL) and finally we chose the best data fit. The SEM is a combination of models of path analysis and factor analysis models. The first model study the causal relationships among multiple variables while the latter allows the analysis of complex phenomena characterized by a set of variables unobservable (latent) through the use of manifest variables. The use of structural equation models allows the simultaneous analysis of latent structures underlying the relationship between endogenous and exogenous variables. The application of structural methods is designed to check the causal pathways between the determinants of the flight and the propensity to return. Descriptive analysis and factor analysis identified the determinants that explain the tendency to migrate and to return, that is, those latent factors capable of synthesizing correlations between aspects of Italian researchers abroad. The exploratory factor analysis examines the relationships between a set of variables, and summarizes the information related to them by identifying a smaller number of behaviors with a small loss of information. Journal of Advanced Research in Management We performed a multivariate analysis of the information provided by the PCA. We analyzed the number of PCs that explained 73% of the total variance of the data set. The PCA analysis produced two components that show higher variability: in the first component there are 22 indicators out of 52, and in the second, 23 out of 52. The latent structure emerged from the analysis is composed of two systems of variables correlated (see Figure 2): the first, is characterized by factors that represent the state of satisfaction perceived by the host country 1, such as salary (X20), the general organization of work (X26a), public policies in support of research (X26c), career opportunities (X26e), management of working hours (X26f), relations with superiors (X26g), the similarities in work teams (X26l); the second, represented by the expectations associated with the Italian research system 2, as the career opportunities (X28a), greater availability of research funds (X28b), the availability of advanced technologies (X28d), appropriate compensation levels (X28e). X20 X26a X26c X26e X26f X26g X26l X28a X28b X28d X28e Figure 2. Diagram of the functional systems identified Source: Elab. Stat Econ database ­ Updated 2010.12.31. The coefficients of the function are highly significant and confirm that X27 depends on the predictor's combination at 95% probability and a good adaptation of 0.985 (p=0.000) (see Table 2). The SEM estimated fits good (see table 2). If the GFI index exceeds 0.95 indicates a good fit of the model to the observed data. This result is also underlined by the good fit RMSEA index. With SEM we have obtained more information about the relations between the variables and the propensity to return. The Propensity to return depends mainly by factors that are associated with the expectations on the Italian research (28a, 28b and 28e), than those related to the representation of satisfaction abroad (20, 26a, 26c and 26e). Table 2. Indicators of fit Lisrel model Fit measure 2 p-value RMSEA GFI Results 208.75 0.5113 0.023 0.985 Good fit 022df 0.05p1 0 RMSEA 0.05 0.95 GFI 1.00 Acceptable fit 2df 23df 0.01p0.05 0.05 RMSEA 0.08 0.90 GFI 0.95 Source: Elab. Stat Econ database ­ Updated 2010.12.31. This early indication shows that the best conditions of the research system in Italy greatly influence the propensity to return. Moreover, the estimate of the endogenous variables used to identify significant factors as contributing to generate indirect influences on the propensity to return, via the direct relationship with the exogenous factors. In summary, the propensity to return is dependent on relationships: direct, (+0.49) with the lower employment opportunities in Italy (28a) and indirect (+0.18), higher employment opportunities abroad (24g). This result represents the state of dissatisfaction of the researchers emigrated; direct, with the increased availability (+0.23) of research funds in Italy (28b) and indirect (+0.04), from positive comments on the relationship between universities and businesses in Italy (17); direct (+0.37) related with the Italian salary (28e) and indirect (+0.08), with positive comments on the relationship between universities and businesses in Italy (17); direct (-0.38) with the degree of satisfaction related to career opportunities abroad (26e) and indirect (+0.33) to the low skill enhancement curriculum in Italy (24e); direct (-0.17) with the policies in support of research (26c) and indirect (+0.13) with on the relationship between universities and enterprises in the host country (18); direct (-0.31), with the degree of satisfaction perceived abroad on work organization as a whole (26a) and indirect (+0.15) with the availability of more equipment (26i), (+0.09) with better satisfaction in relationships with superiors (26g) and (+0.12) with the workplace satisfaction (26b); Issue 2(6) Volume III Winter 2012 and finally, direct (-0.19) with the assessment of the salary currently received (20). The estimated parameters and the relationships generated between the variables examined were significant (*p-value <0.05 see Figure 3). 26b .49* 26i 26g 26g +.15* +.09* +.13* +.33* 26b +.12* 26i 18 24d 18 24e 24g 24g +.18* .45* +.04* .75* +.08* .15* 26a .52* . 26c 36* 26e -. 38* . 49* 28a +.23* + 28b +.37 .60* 28e -.19* -.31* -.17* Y=X27 Result's ordinal variables Nothing ­ Low ­ Average - High Figure 3. Representation of the structural equation model Source: Elab. Stat Econ database ­ Updated 2010.12.31. 5. Policy implications and conclusion The results of this paper provide highly stimulating policy implications and confirm that: more employment opportunities in Italy, compared to abroad, produce positive effects on the propensity to return to Italy; greater availability of research funds in Italy, combined with the best reviews on the relationship between universities and business, would produce positive effects on the return; wages more satisfactory in Italy, combined with better policies for development in the relations between universities and business, to produce positive effect on the propensity to return; compared with the best policies to support research abroad, combined with a positive feedback in the relationship between research and business abroad, to produce a negative effect on the tendency to return to Italy; higher levels of satisfaction, the organization of work, positive evaluations of the workplace, the relationships with superiors and the availability of research facilities, generate negative effects on the propensity to return; best salary levels recognized abroad produce a negative effect on the tendency to return. These results are the inputs on which the Italian research system has to invest. To increase the attractiveness of research in Italy and the return of its researchers need to improve the performance of endogenous and exogenous variables of the relationship estimated in the model. The greater the supply of these factors, the greater the propensity to return to Italy, despite the well-being registered abroad adversely affect the return. Such information could be represented and estimated only through structural equation models. The evidence obtained in this study should lead policymakers in both developing and developed countries not to focus their attention on restricting migration flows of educated individuals. Journal of Advanced Research in Management References [1] Avveduto, S. and Brandi, M.C. 2004. Le migrazioni qualificate in Italia, Studi Emigrazione XLI: 797-829. [2] Batista, C., Lacuesta, A. and Vicente, P. 2007. Brain Drain or Brain Gain: Evidence from African Success Story, IZA discussion paper 3035, Bonn, September. [3] Beccari, A., and Torrisi, B. 2003. New Statistical Methodology for Variable Selection: a Chemiometric Application, Journal of Applied Statistics 15(3). [4] Becker, S.O., Ichino, A., and Peri, G. 2004. How large is the "brain drain" from Italy?, Giornale degli Economisti e Annali di Economia 63: 1-32. [5] Biondo, A.E., Monteleone, S., Skonieczny, G., and Torrisi, B. 2012. Propensity to return: theory and evidence of Italian brain drain, Economics Letters N. 115: 359­362. [6] Brandi, M.C. 2001. Evoluzione degli studi sulle skilled migration: brain drain e mobilità, Studi Emigrazione, XXXVIII no. 141: 75-93. [7] Brandi, M.C., Cerbara, L., I Ricercatori stranieri in Italia: fattori di push e pull, Studi Emigrazione, anno XXXI , No 156 (2004). [8] Brandi, M.C., Segnana, M.L. 2008. Lavorare all'estero: fuga o investimento?, Consorzio Interuniversitario Alma Laurea (ed.) X Indagine Alma Laurea sulla condizione occupazionale dei laureati, Il Mulino. [9] Carrington, W.J., and Detragiache, E. 1998. How big is the brain drain?, IMF Working paper no, 201. [10] Cattaneo, C. 2009. The Decision to Migrate and Social Capital: Evidence from Albania, Fondazione Eni Enrico Mattei and University of Sussex. [11] Docquier, F., Lowell, B.L. and Marfouk, A. 2009. A gendered assessment of the brain drain, Population and Development Review 35(2): 297-321. [12] Docquier, F., and Marfouk, A. 2006. International migration by educational attainment (1990-2000), in Ozden, C., and Schiff (eds). International migration, remittances and the brain drain, Chapter 5, Palgrave-Macmillan. [13] Dustmann, C. 1996. Return migration, The European experience, Economic Policy 22. [14] Dustmann, C. and Weiss, Y. 2007. Return migration: Theory and Empirical evidence, CReAM, CDP No 02 London. [15] Gagliarducci, S., Ichino, A., Peri, G., and Perotti, R. 2005. Lo Splendido Isolamento dell'Università Italiana, Working Paper, Fondazione Rodolfo De Benedetti, Milano, www,igier,uni-bocconi,it/perotti. [16] Gundel, S., and Peters, S. 2008. What determines the duration of stay of immigrants in Germany? Evidence from a longitudinal duration analysis, SOEP papers 79, DIW Berlin. [17] Kendall, M. A New Measure of Rank Correlation, Biometrika, 30, (1938): 81-89. [18] Lacuesta, A. 2006. Emigration and human capital: who leaves, who comes back and what differences does it make?, Working paper Bank of Spain 0602. [19] Mayr, K. and Peri, G. 2008. Return Migration as a Channel of Brain Gain, Working Paper no, 14039. [20] Monteleone, S. 2011. Brain Drain e Crescita Economica: Una Rassegna Critica sugli Effetti Prodotti", QA n.1Rivista dell'Associazione Rossi-Doria (2011): 29-51. [21] Monteleone S., Torrisi B. 2012. Italian Researchers Abroad: A Multivariate Analysis of Migration Trends, Rivista Italiana degli Economisti XVII n.1 (2012): 101-128. [22] Monteleone, S., and Torrisi, B. 2012. Geographical analysis of the academic brain drain in Italy, Scientometrics Vol. 93, Issue 2 (2012): 413-440. [23] Piras, R. 2005. Un'analisi dei flussi migratori interregionali dei laureati: 1980-1999, Rivista Economica del Mezzogiorno, Vol, XIX, (2005): 129-162. [24] Torrisi, B., and Skonieczny, G. 2009. A statistical approach to study the determinants geographic mobility of brain drain, Proceedings Second Arab Statistical Conference 2-4 (2009): 543-552. [25] Torrisi, B., and Skonieczny, G. 2011. Fuga del capitale umano italiano di alta qualificazione: esclusione sociale o povertà indotta?, Rivista della SIEDS n. 3/4 anno 2011. [26] Torrisi, B. 2012. La produttività accademica correlata al benessere lavorativo dei ricercatori italiani in Italia e all'estero", Rivista Italiana di Economia Demografia e Statistica, in press. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Advanced Research in Management de Gruyter

The Structural Equation Model in the University Research System

Loading next page...
 
/lp/de-gruyter/the-structural-equation-model-in-the-university-research-system-kfyIrUfpS0
Publisher
de Gruyter
Copyright
Copyright © 2012 by the
ISSN
2068-7532
DOI
10.2478/v10258-012-0010-x
Publisher site
See Article on Publisher Site

Abstract

In this paper we want to show the propensity to return to Italy of a sample of Italian academics who have emigrated abroad and we provide new empirical evidence through structural equation models (SEM). With SEM we have obtained more information about the relations between the variables and the propensity to return. Moreover, the estimate of the endogenous variables used to identify significant factors as contributing to generate indirect influences on the propensity to return, via the direct relationship with the exogenous factors. Keywords: structural equation model, propensity to return, brain drain, academics. JEL Classification: F22, J24, O15. 1. Introduction The term drain researcher is not used in literature before with it we want to specify that the sample used in the paper regards researchers (assistant professors) and professors in various universities worldwide. This means that we introduce a sub category of a wider term like brain drain. The present paper aims to elaborate an empirical model to identify the main factors leading to the Italian brain drain (or drain researcher), assesses the propensity to return of highly qualified Italian emigrants. This goal is achieved by means of a sample survey. Interviewees were selected at random among Italian graduates, doctoral students, assistant professors and academics from those who emigrated from Italy. The paper shows the results of empirical analysis: the model considers the propensity to return to Italy of Italians abroad. Our empirical analysis envisaged the use of multivariate and non-parametric statistical model. The model used is structural equation model (SEM). This model produces the relationship between the different variables that affect the propensity to return to Italy. SEM methodology makes to go through the limits of general regression models (that measures just the direct relationship) because it allows measuring the indirect, indirect and latent relationships. A review of the literature shows that the return migration to Europe and other regions, such as Asia, is far from marginal. Many emigrants return to their home country and an even greater proportion is highly qualified. Recent studies considered the migration of qualified agents as transitory (Batista et al. 2007, Gundel and Peters 2008, Mayr and Peri, 2008, Dustmann and Weiss, 2007). Micro-data analysis allows us to overcome the informational limits imposed by the use of macro-data (Brandi, 2001; Piras 2005; Lacuesta 2006, Cattaneo 2009). The Likert scales used in the survey provide an indepth view of the attitude of interviewees to return to Italy than would have been obtained using dichotomous variables (Dustmann 1996 e 2007). There are relatively few empirical studies of the brain drain in Italy (Avveduto and Brandi, 2004; Becker, Ichino and Peri, 2004; Brandi and Cerbara, 2004; Gagliarducci et al., 2005; Brandi and Segnana, 2008). However, little importance has been attached to distinguish between the permanent or transitory nature of highly qualified emigration. The paper is organized as follows. Section 2 presents the data set and the methodology. The main results are described in section 3, while section 4 shows the model. Section 5 presents some policy implications and conclusion remarks. 2. Data set and methods Recently, a number of authors have undertaken the study of the stock of skilled workers in different countries of origin with a view to obtaining information on the brain drain (Carrington and Detragiache, 1998; Docquier and Marfouk, 2006; Doquier Lowell and Marfouk, 2009; Monteleone, 2011). Doquier and Rapoport (2009) assess the overall impact the brain drain has upon the countries of origin, evaluating the costs and benefits of such migration for developing countries both in macro- and micro-economic terms. Micro-economic analysis appears to offer the more interesting field of study. Assessing the brain drain and testing hypotheses Journal of Advanced Research in Management through micro-data have been scantily covered in the literature, at least as far as Italy is concerned (BrandiCerbana, 2004; Becker, Ichino and Peri, 2004). However, while there is undeniably a rich theoretical literature, the empirical literature is scarce (Torrisi, Skonieczny, 2009 and 2011; Biondo et al., 2012). This paper continues the development of results obtained in earlier surveys in 2009 and then expanded again in 20101. It is based on data sets of micro-data: 1400 individual on-line interviews of Italian researchers abroad. The data set is based on a sample of 1400 contacts among Italian PhD researchers (assistant professors) and Italian professors in various universities worldwide. A data platform is developed in relation to the participation and involvement of a sample of Italian immigrant researchers in countries with strong research appeal: Canada, Germany, France, Switzerland and Australia. Out of 1400 individual contacts only 68% (955/1400) responded in full. The sample of respondents comprises individuals who are highly educated in different fields of scientific research or highly skilled workers. The lack of official statistics or surveys on the size of population, did not allow any estimate of the number to be sampled and primarily of a criterion of selection of units. The 1400 contacts (Source: From the Catania University StatEcon2 database - Year 2010) are classified in 60 variables into the following macro areas of interest (see Monteleone, Skonieczny and Torrisi, 2010, 2012 and Biondo A. et al., 2012 for a detailed description). Each variable was analyzed according to different scales of measurement on a case-by-case basis. For the most part, the study uses Likert scales, while for some variables it was necessary to associate ordinal, nominal and interval-based scales3. The SEM instruments are able to provide adequate representations of causal relations, both in terms of the total impact, that part of the independent variables on employees. A regression equation can be interpreted as a causal model where the different coefficients give some support to the causal relationship between the variables, but the causal nature of this model is pragmatic and does not have theoretical support about the internal mechanisms of influence between variables (the regressive models do not estimate the underlying or causal relationships between variables). To assess the report in detail, splitting the total direct and indirect relationship, we used a SEM attempting, through a set of equations, to reproduce the structure of causal processes between variables. The methods used for the construction of structural models are based on models of "recursive type". 3. Descriptive results Descriptive analysis generated the following results: the subjects who leave Italy to go to another country that can offer them better living and job conditions. Respondents' preferred destinations are Britain, Switzerland and Canada (see Table 1). Table 1. Percentage of respondents by host country4 Nh Wh USA 749.7 35.7% UK 504 24.0% Canada 281.4 13.4% France 170.1 8.1% Spain 94.5 4.5% Germany 90.3 4.3% Switzerland 79.8 3.8% Austria 42 2.0% Holland 31.5 1.5% Other 42 2.0% Africa 14.7 0.7% Total 2100 100,0% Source: Stat Econ database ­ Updated 2010.12.31. nh 500 336 188 113 63 60 53 28 21 28 10 1400 nh complete 376 221 144 66 43 33 25 15 13 13 6 955 Skonieczny G., Torrisi B. (2009 - 2011), Monteleone S., Torrisi B. (2011 - 2012), Torrisi B. (2012), Biondo A. et al. (2012). StatEcon is the economic statistics science of the Department of Enterprise Culture and Society, University of Catania. 3The Likert scale measures attitudes. 4 Where Nh is the size of h-th layer and H the number of layers that should be considered with h=1,...,H e h Nh = N ; nh ( h nh = n ) is the sample size in the generic layer h; nh completi is the number of completed questionnaires that were returned for each layer and reclassified. Issue 2(6) Volume III Winter 2012 The results following the prevailing literature have identified these countries as the most capable of attracting workers, especially highly qualified workers. The position of researcher or professor is correlated with the level of basic knowledge obtained in Italy or with working experience in Italy. This result is what affects the experience developed in Italy in relation to basic training ever developed in Italy. This means that the preparation affects the current academic status in the host country, unlike the academic experience had in Italy by the interviewees. The analysis shows that those individuals who leave Italy are well-informed about research in Italy; individuals who have had working experience in Italy before leaving the country demonstrate significant understanding (p-value=0,0034) of how research is financed. The perception is that those who have had work experience in Italy consider that the system is not meritocratic for access to research funding in Italy. A fundamental aspect of the survey is to decide how the host countries perceive the career of individuals engaged in research and what mechanisms govern career progression. A clear majority of researchers (84.45% as the sum of absolutely fairly meritocratic and meritocratic) confirmed that career progress is judged as significantly meritocratic (p-value = 0.0000). These results represent two important elements: 1. expressed opinions about how host country considers the state of meritocratic career progression, 2. that this opinion is shared among all countries surveyed with the exception of some European countries. A significant emigration aspect concerns the relationship between age and career progress. At the age of 30 years, subjects go abroad to become researchers; older agents become teachers. The targets for young migrants are significantly age-correlated. Contrary to the prevalent thrust of the literature which sees recent migration as a transitory phenomenon, the results of our analysis show that in Italy it is permanent. This result is obtained by evaluating the emigrants' propensity to return. This degree of propensity is assessed on the basis of the percentage of responses given in relation to a scale of evaluations designed to highlight the subjects' attitude to the idea of returning to their home country. The 73.3% of respondents have a low, or no, propensity to return to Italy (see Figure1). higt 6,5% 41,2% nothing 20,2% average 32,1% low Figure 1. Percentage distribution relative to propensity to return Source: Stat Econ database ­ Updated 2010.12.31. 4. The model results We considered a database of 955 lines (statistical units) and 60 rows (variables). In relation to the complexity of variables, the propensity to return was studied through multivariate analysis models. For the type of variables, we applied different regressive models (OLS, GLM and LISREL) and finally we chose the best data fit. The SEM is a combination of models of path analysis and factor analysis models. The first model study the causal relationships among multiple variables while the latter allows the analysis of complex phenomena characterized by a set of variables unobservable (latent) through the use of manifest variables. The use of structural equation models allows the simultaneous analysis of latent structures underlying the relationship between endogenous and exogenous variables. The application of structural methods is designed to check the causal pathways between the determinants of the flight and the propensity to return. Descriptive analysis and factor analysis identified the determinants that explain the tendency to migrate and to return, that is, those latent factors capable of synthesizing correlations between aspects of Italian researchers abroad. The exploratory factor analysis examines the relationships between a set of variables, and summarizes the information related to them by identifying a smaller number of behaviors with a small loss of information. Journal of Advanced Research in Management We performed a multivariate analysis of the information provided by the PCA. We analyzed the number of PCs that explained 73% of the total variance of the data set. The PCA analysis produced two components that show higher variability: in the first component there are 22 indicators out of 52, and in the second, 23 out of 52. The latent structure emerged from the analysis is composed of two systems of variables correlated (see Figure 2): the first, is characterized by factors that represent the state of satisfaction perceived by the host country 1, such as salary (X20), the general organization of work (X26a), public policies in support of research (X26c), career opportunities (X26e), management of working hours (X26f), relations with superiors (X26g), the similarities in work teams (X26l); the second, represented by the expectations associated with the Italian research system 2, as the career opportunities (X28a), greater availability of research funds (X28b), the availability of advanced technologies (X28d), appropriate compensation levels (X28e). X20 X26a X26c X26e X26f X26g X26l X28a X28b X28d X28e Figure 2. Diagram of the functional systems identified Source: Elab. Stat Econ database ­ Updated 2010.12.31. The coefficients of the function are highly significant and confirm that X27 depends on the predictor's combination at 95% probability and a good adaptation of 0.985 (p=0.000) (see Table 2). The SEM estimated fits good (see table 2). If the GFI index exceeds 0.95 indicates a good fit of the model to the observed data. This result is also underlined by the good fit RMSEA index. With SEM we have obtained more information about the relations between the variables and the propensity to return. The Propensity to return depends mainly by factors that are associated with the expectations on the Italian research (28a, 28b and 28e), than those related to the representation of satisfaction abroad (20, 26a, 26c and 26e). Table 2. Indicators of fit Lisrel model Fit measure 2 p-value RMSEA GFI Results 208.75 0.5113 0.023 0.985 Good fit 022df 0.05p1 0 RMSEA 0.05 0.95 GFI 1.00 Acceptable fit 2df 23df 0.01p0.05 0.05 RMSEA 0.08 0.90 GFI 0.95 Source: Elab. Stat Econ database ­ Updated 2010.12.31. This early indication shows that the best conditions of the research system in Italy greatly influence the propensity to return. Moreover, the estimate of the endogenous variables used to identify significant factors as contributing to generate indirect influences on the propensity to return, via the direct relationship with the exogenous factors. In summary, the propensity to return is dependent on relationships: direct, (+0.49) with the lower employment opportunities in Italy (28a) and indirect (+0.18), higher employment opportunities abroad (24g). This result represents the state of dissatisfaction of the researchers emigrated; direct, with the increased availability (+0.23) of research funds in Italy (28b) and indirect (+0.04), from positive comments on the relationship between universities and businesses in Italy (17); direct (+0.37) related with the Italian salary (28e) and indirect (+0.08), with positive comments on the relationship between universities and businesses in Italy (17); direct (-0.38) with the degree of satisfaction related to career opportunities abroad (26e) and indirect (+0.33) to the low skill enhancement curriculum in Italy (24e); direct (-0.17) with the policies in support of research (26c) and indirect (+0.13) with on the relationship between universities and enterprises in the host country (18); direct (-0.31), with the degree of satisfaction perceived abroad on work organization as a whole (26a) and indirect (+0.15) with the availability of more equipment (26i), (+0.09) with better satisfaction in relationships with superiors (26g) and (+0.12) with the workplace satisfaction (26b); Issue 2(6) Volume III Winter 2012 and finally, direct (-0.19) with the assessment of the salary currently received (20). The estimated parameters and the relationships generated between the variables examined were significant (*p-value <0.05 see Figure 3). 26b .49* 26i 26g 26g +.15* +.09* +.13* +.33* 26b +.12* 26i 18 24d 18 24e 24g 24g +.18* .45* +.04* .75* +.08* .15* 26a .52* . 26c 36* 26e -. 38* . 49* 28a +.23* + 28b +.37 .60* 28e -.19* -.31* -.17* Y=X27 Result's ordinal variables Nothing ­ Low ­ Average - High Figure 3. Representation of the structural equation model Source: Elab. Stat Econ database ­ Updated 2010.12.31. 5. Policy implications and conclusion The results of this paper provide highly stimulating policy implications and confirm that: more employment opportunities in Italy, compared to abroad, produce positive effects on the propensity to return to Italy; greater availability of research funds in Italy, combined with the best reviews on the relationship between universities and business, would produce positive effects on the return; wages more satisfactory in Italy, combined with better policies for development in the relations between universities and business, to produce positive effect on the propensity to return; compared with the best policies to support research abroad, combined with a positive feedback in the relationship between research and business abroad, to produce a negative effect on the tendency to return to Italy; higher levels of satisfaction, the organization of work, positive evaluations of the workplace, the relationships with superiors and the availability of research facilities, generate negative effects on the propensity to return; best salary levels recognized abroad produce a negative effect on the tendency to return. These results are the inputs on which the Italian research system has to invest. To increase the attractiveness of research in Italy and the return of its researchers need to improve the performance of endogenous and exogenous variables of the relationship estimated in the model. The greater the supply of these factors, the greater the propensity to return to Italy, despite the well-being registered abroad adversely affect the return. Such information could be represented and estimated only through structural equation models. The evidence obtained in this study should lead policymakers in both developing and developed countries not to focus their attention on restricting migration flows of educated individuals. Journal of Advanced Research in Management References [1] Avveduto, S. and Brandi, M.C. 2004. Le migrazioni qualificate in Italia, Studi Emigrazione XLI: 797-829. [2] Batista, C., Lacuesta, A. and Vicente, P. 2007. Brain Drain or Brain Gain: Evidence from African Success Story, IZA discussion paper 3035, Bonn, September. [3] Beccari, A., and Torrisi, B. 2003. New Statistical Methodology for Variable Selection: a Chemiometric Application, Journal of Applied Statistics 15(3). [4] Becker, S.O., Ichino, A., and Peri, G. 2004. How large is the "brain drain" from Italy?, Giornale degli Economisti e Annali di Economia 63: 1-32. [5] Biondo, A.E., Monteleone, S., Skonieczny, G., and Torrisi, B. 2012. Propensity to return: theory and evidence of Italian brain drain, Economics Letters N. 115: 359­362. [6] Brandi, M.C. 2001. Evoluzione degli studi sulle skilled migration: brain drain e mobilità, Studi Emigrazione, XXXVIII no. 141: 75-93. [7] Brandi, M.C., Cerbara, L., I Ricercatori stranieri in Italia: fattori di push e pull, Studi Emigrazione, anno XXXI , No 156 (2004). [8] Brandi, M.C., Segnana, M.L. 2008. Lavorare all'estero: fuga o investimento?, Consorzio Interuniversitario Alma Laurea (ed.) X Indagine Alma Laurea sulla condizione occupazionale dei laureati, Il Mulino. [9] Carrington, W.J., and Detragiache, E. 1998. How big is the brain drain?, IMF Working paper no, 201. [10] Cattaneo, C. 2009. The Decision to Migrate and Social Capital: Evidence from Albania, Fondazione Eni Enrico Mattei and University of Sussex. [11] Docquier, F., Lowell, B.L. and Marfouk, A. 2009. A gendered assessment of the brain drain, Population and Development Review 35(2): 297-321. [12] Docquier, F., and Marfouk, A. 2006. International migration by educational attainment (1990-2000), in Ozden, C., and Schiff (eds). International migration, remittances and the brain drain, Chapter 5, Palgrave-Macmillan. [13] Dustmann, C. 1996. Return migration, The European experience, Economic Policy 22. [14] Dustmann, C. and Weiss, Y. 2007. Return migration: Theory and Empirical evidence, CReAM, CDP No 02 London. [15] Gagliarducci, S., Ichino, A., Peri, G., and Perotti, R. 2005. Lo Splendido Isolamento dell'Università Italiana, Working Paper, Fondazione Rodolfo De Benedetti, Milano, www,igier,uni-bocconi,it/perotti. [16] Gundel, S., and Peters, S. 2008. What determines the duration of stay of immigrants in Germany? Evidence from a longitudinal duration analysis, SOEP papers 79, DIW Berlin. [17] Kendall, M. A New Measure of Rank Correlation, Biometrika, 30, (1938): 81-89. [18] Lacuesta, A. 2006. Emigration and human capital: who leaves, who comes back and what differences does it make?, Working paper Bank of Spain 0602. [19] Mayr, K. and Peri, G. 2008. Return Migration as a Channel of Brain Gain, Working Paper no, 14039. [20] Monteleone, S. 2011. Brain Drain e Crescita Economica: Una Rassegna Critica sugli Effetti Prodotti", QA n.1Rivista dell'Associazione Rossi-Doria (2011): 29-51. [21] Monteleone S., Torrisi B. 2012. Italian Researchers Abroad: A Multivariate Analysis of Migration Trends, Rivista Italiana degli Economisti XVII n.1 (2012): 101-128. [22] Monteleone, S., and Torrisi, B. 2012. Geographical analysis of the academic brain drain in Italy, Scientometrics Vol. 93, Issue 2 (2012): 413-440. [23] Piras, R. 2005. Un'analisi dei flussi migratori interregionali dei laureati: 1980-1999, Rivista Economica del Mezzogiorno, Vol, XIX, (2005): 129-162. [24] Torrisi, B., and Skonieczny, G. 2009. A statistical approach to study the determinants geographic mobility of brain drain, Proceedings Second Arab Statistical Conference 2-4 (2009): 543-552. [25] Torrisi, B., and Skonieczny, G. 2011. Fuga del capitale umano italiano di alta qualificazione: esclusione sociale o povertà indotta?, Rivista della SIEDS n. 3/4 anno 2011. [26] Torrisi, B. 2012. La produttività accademica correlata al benessere lavorativo dei ricercatori italiani in Italia e all'estero", Rivista Italiana di Economia Demografia e Statistica, in press.

Journal

Journal of Advanced Research in Managementde Gruyter

Published: Dec 1, 2012

There are no references for this article.