Population Change and International and Internal Migration in Italy, 2002-2017: Ravenstein Revisited Population Change and International and Internal Migration in Italy, 2002-2017: Ravenstein Revisited* Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci Abstract: In 1885, Ravenstein formulated his “laws” of migration, based on the ex- perience of the British Isles. In a further 1889 paper, he extended his analysis as a tour d’horizon of migration and population changes in other nations, including Italy. Even if social and economic processes including globalisation and rising mobility have changed the world since then, Ravenstein’s “laws” remain a point of reference today. Harnessing theoretical and methodological advances made since the 19th century, this paper describes and seeks to explain the role of international and inter- nal migration in regional population change in Italy from 2002-2017. This paper pro- vides the fi rst geographically detailed migration analysis for the country’s 611 Local Labour Market Areas (LLMAs), using register-based migration and population data. Our contribution focuses on several of Ravenstein’s “laws” relating to gender (differ- ences between men and women), natives and non-natives (differences between the Italian and the foreign population), distance migrated from origin to destination, and the role of the economy in shaping push and pull factors of migration. The results show that international migration is more prominent among men than women. In the case of internal moves, the rates of migration among men and women are simi- lar, and internal migration is more prominent among the foreign than the native Ital- ian population. Overall, international migration gains contribute substantially more to population change than internal migration gains and losses do. In Italy, the effects of persistent economic imbalances and of distance on migration patterns are not in line with Ravenstein’s hypotheses: not all areas with high unemployment show an effect of dispersion, nor does distance always act as a deterrent to migration. The geographically detailed analysis presented here illustrates the temporal and spatial coexistence of diverse international and internal migration processes depending on local characteristics, as well as the importance of the economic or administrative centres as the driving force behind national patterns. Our results show that, even 130 years after their formulation, Ravenstein’s migration “laws” (more accurately Comparative Population Studies Vol. 44 (2019): 497-532 (Date of release: 08.09.2020) Federal Institute for Population Research 2020 URL: www.comparativepopulationstudies.de DOI: https://doi.org/10.12765/CPoS-2020-16 URN: urn:nbn:de:bib-cpos-2020-16en8 * This article belongs to a special issue on “Internal Migration as a Driver of Regional Population Change in Europe: Updating Ravenstein”. • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci498 called “hypotheses” today) are still a valuable starting point in assessing and under- standing migration processes and their role in regional population change. Keywords: Internal migration · International migration · Population redistribution · Socio-economic factors · Italy · Local Labour Market Areas 1 Introduction Ernst Georg Ravenstein’s work (1876, 1885 and 1889) laid the foundations for many of the theories of migration we know today. Whereas Ravenstein formulated the “laws” in his 1885 paper using data for the British Isles, his 1889 paper attempted to verify whether these “laws” also held in the context of the contemporary political entities of continental Europe and North America. Ravenstein’s work continues to be infl uential and can be found in most of the theoretical approaches and interpreta- tive models on internal and international migration (for example, Lee 1966; Tobler 1995; Piché 2013; Stillwell/Thomas 2016). The gravitational approach, rural-urban migration and urbanisation, the relationship between migration and development, the two-way dynamics of migration (net migration and return migration), the gen- dering of migration, and the economic determinants of migration are only some of the most relevant issues that are based on Ravenstein’s seminal works (King 2012). Ravenstein also understood that the forces underlying migratory movements were “complex and include the quality of the public infrastructure, such as roads, climate, taxation and more” (Greenwood 2019: 269). The validity of Ravenstein’s laws has been tested in numerous contributions. Macisco and Pryor (1963), for ex- ample, used the laws to provide a framework to review the many studies of migra- tion in the USA. More recent studies have focused on some points directly related to Ravenstein’s hypotheses, such as the differentials in mobility between women and men, underlining the importance of available data and the historical context in the formulation of the “laws” of migration. For example, Alexander and Steidl (2012) show that Ravenstein’s affi rmation that women are more migratory is based on, for Ravenstein, hidden age structure differences for men and women due to differ- ences in mortality and out-migration. However, Ravenstein himself answered his critics: “Of course I am perfectly aware that our laws of population, and economic laws generally, have not the rigidity of physical laws, as they are continually being interfered with by human agency.” (1989: 241). The direct and indirect presence of Ravenstein’s “laws” and ideas in the actual academic discussion show that empiri- cal migration studies are not imaginable without the foundations laid by Ravenstein in his contributions. Clearly, after more than 130 years and fundamental changes in social and economic processes, as well as increasing globalisation and mobility, not all of Ravenstein’s observations can be confi rmed, and it seems easy to fi nd excep- tions to the “rules”. Population Change and International and Internal Migration in Italy • 499 In this paper we investigate whether the pioneering statements by Ravenstein remain relevant for the analysis of migration in Italy from a geographically detailed perspective, focussing on the period 2002-2017. In Italy, internal and international migration contributed decisively to demographic change and population growth. Nevertheless, partially brought on by the Great Recession after 2008, the role of in- ternational migration diminished and, as a consequence, total population decreased slightly after 2014. The relation between migration and population change has be- come complex and diversifi ed at the geographic level, especially when focussing on Local Labour Market Areas (LLMAs) as units of analysis. In the Italian context, a subset of the Ravenstein “laws” (1885) is particularly relevant: the differences between men and women and the differences, between the Italian and the foreign population regarding international and internal migra- tion processes, as well as his observations regarding the role of distance and the importance of economic context. In Italy, differences in the roles and opportunities of women and men remain important. Similarly, the presence of foreigners in Italy is a current and relevant phenomenon, not least because of the signifi cant socio- demographic differences between foreigners and Italians. Finally, Italy continues to be characterised by a high degree of heterogeneity in socio-economic regional development and accessibility. The paper will examine these premises in four further sections. First, it gives an overview of the post-war international and internal migration trends in Italy (Sec- tion 2). It then presents the data and methods (Section 3) and puts some of Raven- stein’s “laws” through an Italian checklist, underlining the nexus between migration and population change (Section 4). The concluding section summarises the fi ndings (Section 5). 2 Post-war international and internal migration trends in Italy Internal migration has always played a major social and economic role in Italy, but the peak of its importance was certainly achieved in the years of the economic boom of the 1950s and 1960s, with important long-distance migration fl ows origi- nating in the Northeast and the South of Italy and directed to the industrial areas of the Northwest and to Rome (Bonifazi 2013a; Gallo 2012; Golini 1974). The internal movements of the period after the Second World War are linked to the massive eco- nomic and demographic growth of urban areas. This growth was to the detriment of rural areas, including the Apennines (the 1,200 km mountain range that stretches from the Northwest to the South), and the Mezzogiorno (the Southern regions of Italy including the islands of Sardinia and Sicily). It was partly determined by the Fordist economic model that generated a strong demand for a young workforce in the factories of the industrial triangle of the Northwest. In this period the migra- tion fl ows between the Mezzogiorno and the Centre-North and the resulting net migration balance in favour of the Centre-North became one of the most important characteristics of the Italian internal migration pattern, attracting the attention of • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci500 academic and political debate with many analyses focusing on the subordinate po- sition of the Mezzogiorno (Bonifazi 2013b; Impicciatore/Strozza 2016, for example). With the decline of the Fordist model and with the oil price crises of the 1970s, internal mobility decreased and net migration patterns changed. The economic re- covery of the 1980s and the development of the Third Italy led to a shift of the net migration patterns away from the Industrial Triangle of the Northwest to the Third Italy with its smaller and often family-led companies. This created economic well-being in extended areas of Central and Northeastern Italy, such as the cities along the Via Aemilia, a Roman road in Northern Italy from Rimini on the Adriatic coast to Piacenza. However, this did not cause a new rise in internal migration fl ows, since the production processes were not particularly labour-intensive and local la- bour supply increased, especially through the higher labour market participation of women. The slight increase in internal migration in the 1990s was mainly driven by short-distance movements, even though the fl ows between the South and the other Italian territorial divisions continued to be important. However, the main novelty of Italian internal mobility in the last thirty years has been the growing role of the foreign population (Bonifazi et al. 2009, 2016, 2017, 2020, Gallo 2012) with its higher propensity to change residence (Bonifazi et al. 2012; De Filippo/Strozza 2011), even if immigration has declined in recent years. The higher internal mobility of the foreign population is probably due to the inter- twined dynamics of their international and internal migration and to being exposed to greater uncertainty regarding housing and employment. As matter of fact, in the last decades, foreign immigration was unexpectedly the main novelty of Italy’s demographic profi le and the most important characteristic of the Italian migration system. The growth of the foreign population has been sub- stantial: from 0.22 million in 1981 to 5.26 million at the end of 2018. As a result of this considerable migratory infl ow, the total population continued to grow until reaching 60.8 million in early 2015, notwithstanding the low fertility. At the regional level, the foreign population is mainly concentrated in the affl uent Centre-North, which has also continued to benefi t from internal migration from Southern Italy. This paper follows a long tradition of studying a wide range of aspects of Italian internal migration processes and patterns in different academic disciplines, includ- ing demography, geography, sociology and economics. Bonifazi and Heins (2000) gave an overview of internal migration from a demographic point of view including also geographic aspects. Since then, several reviews of internal migration in Italy have been published. For example, Panichella (2012) offers a detailed review of the sociological contributions to the study of internal migration in Italy, confi rming a selection process regarding education and social extraction of the migration fl ows between the South and the Centre-North. Several analyses of specifi c sociological aspects are found in Bubbico et al. (2011). Results of interregional migration analysis of the Italian situation and the the- oretical approaches from an economic point of view are presented, for example, by Biagi and Dotzel (2018) and Faggian et al. (2015). An econometric analysis of the Italian situation is offered by Biagi et al. (2011), analysing the internal migration fl ows between the 103 provinces in 2001-2002 and including economic, social and Population Change and International and Internal Migration in Italy • 501 environmental variables. Distinguishing between short- and long-distance fl ows, Biagi et al. (2011) fi nd different models and state that short-distance migration fl ows do not seem to be determined by economic variables, while in the case of long- distance migration fl ows, “economic/labour market variables play a dominant role” (Biagi et al. 2011: 123). Piras (2017) offers a further review of the empirical economic literature on re- gional patterns of internal migration in Italy and fi nds that most studies confi rm the role played by per capita GDP and unemployment rates as push and pull factors. The analysis covers a longer timeframe (1970-2005), uses the 20 Italian regions, and considers the push and pull factors separately since the variables could have dif- ferent effects if considered at the origin or at the destination of the migration fl ow. Already in the 1990s, the absence of a geographic relationship between unemploy- ment rates and internal migration was found “puzzling” (Faini et al. 1997), with a recent contribution elaborating on this aspect (Faggian et al. 2017). Several analyses focus on the relationship between international immigration in general and on for- eign and Italian internal migration (see for example Mocetti/Porello 2010; Benassi et al. 2019). Basile et al. (2018) probe whether international migrants are complements or substitutes to the Italian workforce. They identify a displacement effect of foreign immigration for 2003-2011 regarding the foreign-born and the low-qualifi ed (with a lower secondary, and especially primary, level of education) Italian population, as well as a complementary effect to the more highly educated Italian population. The last twenty years have been characterised by a slight upturn of internal mo- bility that has been halted by the recent economic and fi nancial crisis (Bonifazi et al. 2017). Notwithstanding the low internal migration intensities and the reduction of the positive international migration balance, the geographic net migration patterns of foreigners and Italians seem to have changed markedly over recent years. 3 Data and methods In his work, Ravenstein used migration data by place of birth from the 1871 and 1881 British censuses, along with similar data from censuses in North America and Europe (Grigg 1977). The defi nition of the person as a migrant is taken from the information on the place at birth and the place of residence at the census dates. If the two places differ, the person is considered a migrant, though there is no data on when the migration actually took place. The data available to Ravenstein allowed a distinction to be drawn between international migration and an internal move. Ravenstein defi ned movements between adjacent counties as “short-distance” and showed that the majority of migrants only migrated in this manner. The present analysis is based on data from the Italian population registers, us- ing offi cial statistics on migratory movements occurring from 2002-2017. The pop- ulation registers are organised at the municipal level. Italy has almost 8,000 mu- nicipalities and every movement between them is counted in these registers. The population registers provide the information on migration fl ows that are recorded when an individual or a family registers a change of residence to the municipality • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci502 of destination. This registration is obligatory and gives access to many administra- tive advantages (local or regional differences in taxes, fees, insurance rates etc.), so it is likely that the vast majority of individuals indeed register their migration. On the other hand, particular subgroups, such as students, are less inclined to com- municate their (temporary) residence to the new municipality. For each change of address, information on the municipalities of origin and destination as well as the main socio-demographic information of the individual that changes residence (citi- zenship, place and date of birth, sex, marital status and educational attainment) is also available. In 2012, a specifi c law (Decree-Law February 9th, 2012 N. 5, converted into law April 4th 2012 N. 35) led to the processing of the information on changes of residence in the municipal population registers almost in real time, with a noticeable reduction of the time elapsed between the request by the citizen and the fi nal trans- mission of the information to the registers. This change in registration methods resulted in an overcount of moves in 2012, which is why the data for 2012 should be interpreted with great caution. To provide spatially detailed analyses of the structure of migration fl ows in Italy, we use fl ow data for 611 Local Labour Market Areas (LLMAs) or Sistemi Locali del Lavoro. These areas were defi ned by Istat (2015) and express the organisation of the Italian national territory based on the relationships between individuals and the social and economic context. The LLMAs are based on the daily commuting pat- terns between the municipalities of residence and the place of work and defi ne the urban systems where most of the daily activities and movements of people are concentrated, providing a grid whose boundaries derive from actual social and economic processes and not historic or geographic ones. The information on the daily commuting fl ows for work were collected in the 15th Population and Housing Census, carried out in 2011 (Istat 2015). Until March 2018, Italy was divided into 611 LLLMAs: 225 in the North, 105 in the Centre, and 281 in the South and the Islands (or Mezzogiorno). To facilitate the comparison of geographic patterns over time, the contribution focuses on two different periods: 2004-2008 and 2013-2017, divided by the effects of the great economic recession after 2008. In Italy, a second economic downturn followed in 2012 and 2013 and the country does not seem to have recovered fully in the years thereafter. Whereas the fi rst period is characterised by a (slowly) growing economy, the second period is characterised by high unemployment in the after- math of the economic crisis. We omit the years 2009 to 2012 in this comparison, since these are years of transition without clear regional patterns and because 2012 presents the abovementioned data situation. The analysis of the role of migration for population change is based on net mi- gration rates at the regional level and aggregate net migration rates (ANMRs) at the national level. The latter is used for the comparison of internal net migration over the entire study period (2002-2017). The aggregate rate is defi ned as ANMR = 100 * 0.5 Σi |Di – Oi| /P (i), where the variables Di and Oi are infl ows to and outfl ows from region i, and P(i) is the population at risk in region i (Bell et al. 2002; Rees et al. 2017). The ANMR thus measures the impact of migration on population redistribution: it identifi es the net shift of population between regions per hundred persons resident Population Change and International and Internal Migration in Italy • 503 in the country. In a second part of the analysis, we focus on geographic differences and compare the two periods (2004-2008 and 2013-2017) based on net migration rates. Annual citizenship, sex and age data by Istat and updated after the 2011 popula- tion census serve as the denominator for calculations of the rates (see http://demo. istat.it/ and http://demo.istat.it/ricostruzione2013/index.php). The net migration rates NMR (NMR = 100 * (Di – Oi) /P(i), with the variables Di and Oi denoting infl ows to and outfl ows from region i, and P(i) as the average population at risk in region i) are calculated for the total population and by citizenship, sex and age groups. It should be noted that since the fi rst years of the study period, the foreign population resident in Italy has grown signifi cantly, growing from 1.5 million in 2003 to more than 5.2 million in 2019; 8.7 percent of the total population. This process goes hand in hand with an increasing number of naturalisations. Throughout this paper, the terms “Italian population” or “Italians” and “foreign population“ or “foreigners“ are used to distinguish the resident population by citizenship. In recent years, the num- ber of naturalisations increased considerably, with a maximum of 185,000 in 2016. When discussing distance, one important point is certainly the way distance is measured. Whereas in the past the geographic distance was the point of reference, distance can today be expressed in many other ways, including expenditures in money and time. We nonetheless prefer the distance between the geographic cen- tres of the LLMAs for simplicity’s sake. Socio-economic indices were used to test the possible factors that infl uence the internal and international net migration rates (for a complete list, see Annex 2). The rates of economic activity and unemployment rates are annually estimated by Istat (see https://www.istat.it/it/archivio/217437). Other and more detailed socio- economic information is drawn from the 2011 population census. In general, the geographic patterns of the indices are stable over time and a high collinearity is observed between indices. For example, the geographic differences of economic activity, unemployment, and the share of foreign population are very similar. This is not surprising given that Italy is a country divided into two macro-areas of eco- nomic well-being. A further dimension includes population density, which is generally used to de- fi ne settlement characteristics of the LLMAs and to describe processes of urbanisa- tion (Champion 2001; Bonifazi/Heins 2003). Population density is the ratio between resident population of a territorial unit (here the LLMAs) and their surface area (ex- pressed in square kilometres). In the Italian context, population density does not de- scribe settlement characteristics and the process of urbanisation very effectively, as rural areas can also have a relatively high population density. Additional socio-eco- nomic information is therefore used to systematically describe the territorial differ- ences in the Italian context. Regarding LLMA age structure, our main focus is on the share of the 20-44-year-olds, the age group with the highest migration intensities. The geographic differences in the age structure are the result of past patterns of demographic change, e.g. lower fertility rates in the Centre-North and higher ones in the Mezzogiorno, and more recent international immigration. The age structure patterns are only partially related to socio-economic disparities. Urbanity (defi ned • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci504 as high proportions of university degree-holders and employment in sectors such as IT or fi nancial services) is a less decisive socio-demographic and socio-economic aspect. The same holds for the late departure from the parental home, a specifi - cally Southern European and Italian cultural aspect infl uencing internal migration patterns of the local population, is represented through the share of the 20-44-year- olds living in the same household as their parents. Population-weighted correlation coeffi cients are used to test for links between international and internal net migra- tion patterns and these socio-economic indices. Inspired by Ravenstein’s work, this paper aims to provide further insights into the recent migration patterns in Italy by analysing the socio-demographic and geo- graphic detail of international and internal net migration patterns. The analysis of internal migration considers the distance migrated and the abovementioned factors linked to the Italian settlement system and the economic situation that are hypoth- esised to infl uence the migration patterns at the level of the LLMAs. 4 Ravenstein’s seven “laws”: an Italian checklist Our analysis of migration fl ows at the nation-wide scale and for selected case study regions aims to verify some of the “laws” observed by Ravenstein over the past 20 years for the Italian context. Ravenstein’s seven “laws” (1885: 198-199) are: 1. “... the great body of our migrants only proceed a short distance” forming “ ‘currents of migration’ setting in the direction of the great centres of com- merce and industry which absorb the migrants.” 2. Migration is leading to population gain through a process of single short dis- tance migration fl ows towards attractive areas. The areas of absorption are areas of positive net migration. 3. Migration is leading to population loss in the areas that are feeding the pro- cess of absorption. The areas of dispersion are areas of negative net migra- tion. In these processes the communication facilities are offsetting the deter- rence effect of distance or “facilities of communication may frequently the disadvantage of distance.” In fact the second and third “laws” are mirror images of one other: areas of popu- lation gain depend on areas of population loss. Ravenstein associates the former with industry and commerce, and the latter with agriculture. Migrations from the areas of dispersion to the areas of absorption were expected to occur in a cascad- ing manner. 4. “Each main current of migration produces a compensating counter-current.” 5. “Migrants proceeding long distances generally go by preference to one of the great centres of commerce and industry”, or towards today’s larger met- ropolitan areas. Population Change and International and Internal Migration in Italy • 505 6. The natives of towns are less migratory than those of the rural parts of the country.” Or, residents of urban areas have a lower propensity to migrate than rural residents. 7. “Females are more migratory that males.” Or, women have a higher propen- sity to migrate. Needless to say, due to the lack of data, the statements about the different in- clinations (being “less” and “more migratory”) were not based on detailed demo- graphic analysis, but on the observation of frequencies. Nonetheless, Ravenstein argued for a demographic approach taking both the population at origin and desti- nation into account. Of the seven “laws” put forward by Ravenstein, this paper focuses specifi cally on the role of two socio-demographic variables: citizenship and sex. We analyse the differences between men and women and between Italians and foreign nationals in relation to international and internal migration and the impact on regional popula- tion change. Further items on our checklist are the distance mentioned in several of Ravenstein’s “laws” and the socio-economic characteristics of the areas of origin and destination. Even though Ravenstein does not explicitly mention the predomi- nant role of the economic drivers of migration, several references are made to the “great centres of commerce and industry” as well as to rural areas when discussing the direction of migration fl ows and the processes of absorption and dispersion. Before discussing the case studies, however, we will give a brief overview of recent trends in international and internal migration and population change in Ita- ly, and the regional relationship between international and internal migration and population change when comparing the two periods 2004-2008 and 2013-2017 by citizenship, sex, and age group. From 2002 through 2017, the total population resident in Italy grew from 56.996 to 60.484 million. This increase is entirely due to the foreign component, since the Italian population slightly decreased from 55.646 million at the beginning of 2002 to 55.340 million at the end of 2017. International migration has shaped national and regional demographic change in Italy over the past two decades. Figure 1 presents annual international net migration in relation to the respective population. The for- eign population grew from 1.357 million at the beginning of 2002 to 5.144 million at the end of 2017 (5.256 at the end of 2018). For the foreign population, the years when measures to legalise irregular migrants were enacted clearly stand out. For example, the years 2002-2003 – after the Bossi-Fini law (Law 189/2002) came into force – saw the legalisation of the legal status of about 650,000 irregular foreigners who presumably immigrated before 2002-2003. The 2007 EU enlargement brought another important migratory infl ux, especially from Romania. The decrease of immigrants after the start of the economic crises and especially after 2010 brought the increase of the number of the foreign nationals resident in Italy to a halt by 2014. Only the arrival of refugees and asylum seekers after 2014 once again led to a slight increase in the international net migration rates. Linked to the immigration fl ows, the high internal aggregate net migration rates of the foreign population are indicated in Figure 2. Until 2008, their total internal • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci506 migration change was signifi cantly higher than the Italian one and only thereaf- ter values declined to around 3 per mille. Between 2013 and 2017, rates fl uctuated between 2 and 3 per mille. As mentioned earlier, the 2012 values are primarily in- fl uenced by the changes in administrative procedure, rather than by the economic downturn. The consistently decreasing internal mobility rate of the foreign popula- tion brings the total internal aggregate net migration rates to levels more similar to those recorded for the Italian population. International net migration change is dominated in the fi rst years by the age groups 20-34, whereas in more recent years they have been superseded by the 15-19-year-olds, with increasing migration gains for the 15-19-year-olds and relative decreasing ones for the 30-34-year-olds (Fig. 3). In fact, since 2014, the characteristics of immigration have changed profoundly: there has been a shift from work-related to family-related immigration. In addition, it should be noted that humanitarian emergencies have played an important role for the migratory infl ows. These changes are obviously refl ected in the international net migration rates by sex and age. However, it should also be noted that emigration fl ows might not always be registered properly due to a lack of incentives for people leaving Italy to formally de-register (UNECE/EUROSTAT 2010). Until 2014, generally (slightly) more foreign women immigrated to Italy than for- eign men did (52.6 percent in 2002, 55.4 percent in 2010), but since then their share in immigration fl ows has declined to 41.6 percent in 2017. In the cases of the immi- Fig. 1: International net migration rates, Italy 2002-2017 (per 1,000) -20 0 20 40 60 80 100 120 140 160 180 200 220 240 2002 2004 2006 2008 2010 2012 2014 2016 Year Italians Foreigners Total International net migration rate (per 1,000) 2018 Source: Authors’ calculations based on Istat: “Iscrizioni e cancellazioni all’anagrafe per trasferimento di residenza”; population estimates by citizenship, sex and age [http://demo.istat.it/] Population Change and International and Internal Migration in Italy • 507 gration and emigration of the Italian population, the shares of women are lower than those of men in all years. The internal net migration dynamics are dominated by the 20-34-year-olds, with the highest aggregate net migration rates for the age group 25-29. Whereas an increasing trend is observed for the 25-34-year-olds, the trend for the younger age group (20-24) is decreasing. This is in line with the observation of the continuous postponing of life course events by young adults in Italy (Istat 2014). The overall internal aggregate net migration declined after 2008, most clearly for the young adults, and recovered in the last years. This paper examines a total of 22 cases in four groups. The fi rst contains the most important Italian metropolitan areas: Milan, Genoa, Bologna, Rome, Naples, Bari and Palermo, with the fi rst four Central-Northern metropolitan areas (except Genoa) being areas of attraction (or “absorption”). The second covers Mezzogiorno LLMAs characterised by migration loss: Isernia, Torre del Greco, Foggia, Melfi , Reg- gio Calabria, Vibo Valentia, Caltanissetta and Gela. Third, the Tuscan LLMAs of Flor- ence, Livorno (“Leghorn”), Pisa and Prato is taken into account because Ravenstein (1889) referred to them explicitly. Fourth and fi nally, three “outliers” are considered: Desenzano del Garda on the shores of Lake Garda and Meran in South Tyrol show special patterns of net migration gain over recent years, since they are important tourist destinations, and L’Aquila is of particular interest because it was hit by an earthquake in 2012. Fig. 2: Internal aggregate net migration rates, Italy 2002-2017 (per 1,000) 0 1 2 3 4 5 6 7 8 9 10 11 12 2002 2004 2006 2008 2010 2012 2014 2016 2018 Year Italians Foreigners Total Internal aggregate net migration rate (per 1,000) Source: Authors’ calculations based on Istat: ”Iscrizioni e cancellazioni all’anagrafe per trasferimento di residenza”; population estimates by citizenship, sex and age [http://demo.istat.it/] • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci508 Table 1 gives an overview of the case studies, drawing on some basic indicators regarding the settlement structure and the demographic and socio-economic situ- ation. For several indicators, the contrast between Central-Northern and Southern Italy is clearly visible. These include the presence of foreigners, the level of eco- nomic activity, and unemployment rates. Another factor that leads to similarities between some of the selected LLMAs is the role of universities. The LLMAs with the highest ratio of enrolled students to the general population are by far Pisa (254 to 1,000) and L’Aquila (171), as well as Bologna (96), Florence (77), and Bari (72). This is particularly important for the Tuscan LLMAs, as is the very high share of foreign population. Florence is an international centre attracting many different nationali- ties, while Prato is one of the centres of Chinese clothing manufacturing in Italy. Whereas Ravenstein’s observations refer to migration fl ows in a historical period characterised by industrialisation and urbanisation, today’s regional patterns of so- cio-economic disparities in Italy are more varied. The economy is characterised by the continual growth of the service sector rather than manufacturing, whereas the urbanisation process is substituted by locally differentiated processes of sub- and re-urbanisation. The geographic pattern of international net migration 2004-2008 refl ects the consistent infl ow of migrants in this period and shows that most areas of Central-Northern Italy gained population (Fig. 4), compared to the less attractive Southern areas. The maps presented in Figures 4 and 5 are cartograms in which land area is proportionally rescaled to match each LLMA’s 2011 census population, consequently “distorting” their shape. Fig. 3: International net and internal aggregate net migration rates by age, Italy 2002-2017 (per 1,000) Source: Authors’ calculations based on Istat: ”Iscrizioni e cancellazioni all’anagrafe per trasferimento di residenza”; population estimates by citizenship, sex and age [http://demo.istat.it/] 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 2002 2004 2006 2008 2010 2012 2014 2016 2018 Year 0-14 15-19 20-24 25-29 30-34 35-44 45-54 55+ International migration International net migration rate (per 1,000) Internal migration Internal aggregate net migration rate (per 1,000) 0 2 4 6 8 2002 2004 2006 2008 2010 2012 2014 2016 2018 Year Population Change and International and Internal Migration in Italy • 509 A re a P o p u la ti o n A n n u a l P o p u la ti o n P o p u la ti o n F o re ig n E co n o m ic U n e m - P o p u la - 2 0 -6 4 - P o p u la ti o n (i n 1 ,0 0 0 ) p o p u la ti o n d e n si ty 2 0 -4 4 p o p u la ti o n a c ti v it y p lo y m e n t ti o n 2 0 -6 4 y e a r- o ld s 2 0 -4 4 g ro w th , (p e r km 2 ) (% ) (% ) ra te , r a te , w it h a u n i- e m p lo y e d liv in g 2 0 0 2 -2 0 17 2 0 -6 4 2 0 -6 4 v e rs it y in t h e s e r- a s so n o r (% ) (% ) d e g re e v ic e s e c to r d a u g h te r (% ) (% ) (% ) M e tr o p o li ta n L L M A s M ila n C -N 3 ,6 8 2 .6 + 0 .6 3 7 2 0 0 4 3 2 .5 10 .0 7 6 .6 6 .6 2 1. 0 3 0 .1 2 7. 4 G e n o a C -N 6 7 9 .5 -0 .2 6 9 74 9 2 8 .1 7. 0 7 2 .9 7. 5 2 0 .1 2 3 .9 2 9 .7 B o lo g n a C -N 8 4 7. 1 + 0 .6 18 3 3 8 3 1. 6 9 .8 7 8 .8 6 .2 2 3 .3 2 2 .8 2 3 .1 R o m e C -N 3 ,4 7 7. 0 + 0 .9 3 5 8 9 4 3 2 .9 8 .8 7 2 .6 9 .8 2 3 .8 2 9 .9 3 2 .3 N a p le s M e zz . 2 ,5 0 9 .6 + 0 .1 3 3 3 10 5 3 5 .3 2 .3 5 7. 1 2 6 .2 13 .3 19 .2 3 8 .9 B a ri M e zz . 7 3 6 .6 + 0 .2 3 3 4 5 8 3 4 .1 2 .3 6 3 .0 16 .0 15 .7 18 .7 3 9 .1 P a le rm o M e zz . 8 7 9 .5 + 0 .1 5 8 7 5 8 3 4 .4 2 .5 6 0 .6 2 4 .1 14 .7 2 0 .5 3 7. 1 L L M A s w it h m ig ra ti o n l o ss Is e rn ia M e zz . 7 9 .6 -0 .2 19 6 2 3 3 .1 2 .5 6 5 .6 13 .1 17 .6 13 .0 4 2 .4 T o rr e d e l G re co M e zz . 2 51 .7 -0 .1 9 3 2 6 3 4 3 4 .6 1. 7 5 6 .5 2 4 .5 12 .3 14 .5 3 9 .9 F o g g ia M e zz . 2 6 3 .1 -0 .2 0 1 12 4 3 3 .3 2 .5 6 0 .7 17 .4 14 .4 14 .6 4 0 .0 M e lfi M e zz . 6 5 .0 -0 .0 8 5 5 8 3 4 .0 2 .9 6 6 .1 17 .0 11 .6 11 .0 3 7. 5 R e g g io C a la b ri a M e zz . 2 17 .4 -0 .0 0 1 4 6 3 3 4 .1 4 .3 6 3 .8 2 1. 8 2 0 .8 17 .9 41 .0 V ib o V a le n ti a M e zz . 10 2 .8 -0 .0 5 4 19 3 3 4 .4 3 .5 61 .4 17 .8 15 .0 12 .8 4 0 .4 C a lt a n is se tt a M e zz . 11 2 .3 -0 .1 5 5 16 4 3 2 .6 2 .5 61 .2 18 .7 14 .4 15 .3 3 8 .2 G e la M e zz . 10 3 .7 + 0 .0 8 6 2 7 6 3 5 .0 1. 3 5 4 .0 2 3 .5 9 .0 9 .4 3 4 .3 Tu sc a n L L M A s F lo re n ce C -N 6 8 6 .6 + 0 .4 41 5 6 9 3 0 .2 10 .3 7 6 .3 6 .4 2 2 .6 2 4 .2 2 9 .8 L iv o rn o C -N 17 8 .6 + 0 .1 17 6 0 9 3 0 .7 5 .5 7 3 .0 10 .3 16 .6 19 .3 3 1. 5 P is a C -N 17 9 .1 + 0 .3 17 4 0 0 3 1. 9 7. 2 74 .7 8 .0 2 6 .2 2 2 .9 3 0 .0 P ra to C -N 2 7 2 .8 + 0 .7 0 3 6 6 6 3 3 .1 12 .8 7 6 .4 9 .1 11 .4 16 .7 3 1. 8 “O u tl ie r” L L M A s D e se n za n o d e l G a rd a C -N 8 5 .6 + 1. 4 51 3 13 3 4 .3 12 .4 7 5 .2 6 .5 14 .0 16 .4 2 5 .2 M e ra n C -N 8 0 .3 + 1. 0 11 12 0 3 3 .3 9 .2 8 0 .3 2 .9 14 .4 13 .5 2 9 .5 L’ A q u ila M e zz . 9 6 .3 + 0 .1 7 0 61 3 2 .8 6 .0 71 .8 7. 4 2 4 .6 19 .7 41 .7 It a ly - 5 9 ,1 3 2 .0 + 0 .3 7 2 19 7 3 2 .8 6 .8 7 0 .0 11 .0 15 .8 18 .1 3 3 .6 T a b . 1: S e le c te d I ta lia n L L M A s – b a si c in fo rm a ti o n , 2 01 1 C -N = C e n tr e -N o rt h , M e zz . = M e zz o g io rn o . Fo r th e d e fi n it io n o f v ar ia b le s se e A n n e x 2 . S o u rc e : A u th o rs ’ ca lc u la ti o n s b as e d o n I st at : p o p u la ti o n e st im at e s b y c it iz e n sh ip , se x a n d a g e [ h tt p :/ /d e m o .is ta t. it /] , 1 5 th P o p u la ti o n a n d H o u si n g C e n su s 2 0 1 1 • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci510 Internal net migration shows the North/South divide of Italy, with most areas of the Mezzogiorno losing population due to internal migration in 2004-2008 and, consequently, most of the Northern LLMAs gaining. Some metropolitan LLMAs of Central-Northern Italy such as Florence, Milan, and Turin also lost population in this period due to internal migration processes, whereas LLMAs adjacent to the metro- politan areas, as well as areas of central Italy, show positive internal net migration rates. From 2013-2017, the geographic pattern of international net migration changes (Fig. 5) was dominated by large net gains in the metropolitan LLMAs such as Milan and adjacent areas, Bologna, Florence, and Rome, and even some smaller LLMAs in the Mezzogiorno. Specifi c cases such as the LLMAs of Crotone, Isernia, and Ragusa, where some of the reception centres for refugees and asylum seekers are located, appear clearly on the map because they often represent their fi rst residence. Some- what surprisingly, some smaller LLMAs closer to the Austrian border as well as in the South have lost population due to a negative migration balance. These areas are especially losing young Italian adults because of a lack of economic opportunities. Before the Great Recession, this trend was hardly observed. Internal net migration continues to divide Italy, with the LLMAs of the Mezzogiorno losing and those of Central-Northern Italy almost homogenously moderately gaining population. Only Fig. 4: International and internal net migration rates, Italian LLMAs 2004-2008 (per 1,000) > Source: Authors’ calculations based on Istat: ”Iscrizioni e cancellazioni all’anagrafe per trasferimento di residenza”; population estimates by citizenship, sex and age [http://demo.istat.it/] � � International Internal Population Change and International and Internal Migration in Italy • 511 the LLMAs of Bologna and Parma, as well as smaller touristic or scenic areas con- tinue to have a positive net migration above 3 per mille. The general decrease in regional disparities of international and internal net mi- gration between the two periods 2004-2008 and 2013-2017 is worth highlighting. We give an indication of the similarity between the geographic patterns of the net migration rates by LLMA, sex, and international/internal migration by comparing population-weighted correlation coeffi cients (2011 census LLMA population data). Regarding geographical patterns of international and internal migration, it can be noted that the correlation coeffi cients between the net migration rates in 2004-2008 are close to +0.450 for the general population as well as for men and women, and in 2013-2017 reach +0.368 in total, +0.186 for men and +0.516 for women. Seemingly the attractiveness of areas is similar comparing the processes of international and internal migration, but by no means identical. Interestingly, the similarity increases between the two periods for women, but diminishes for men. Comparing the two periods, we note a more intense change in the geographical patterns than expected, since the net rates in the case of international migration for 2004-2008 and 2013-2017 correlate by +0.520 (+0.331 for men and +0.655 for women) and those for internal migration by +0.618 (+0.590 for men and +0.637 for women). It would be too easy to attribute these changes solely to the Great Reces- sion, nevertheless it is an indication that changes (but not reversals of geographic patterns) occurred during the study period. Fig. 5: International and internal net migration rates, Italian LLMAs 2013-2017 (per 1,000) Source: Authors’ calculations based on Istat: ”Iscrizioni e cancellazioni all’anagrafe per trasferimento di residenza”; population estimates by citizenship, sex and age [http://demo.istat.it/] � � International Internal • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci512 In general, men and women have very similar geographical patterns of net mi- gration rates: in the case of the internal migration balance, the coeffi cients are +0.958 in the fi rst and +0.943 in the second period. The patterns are also similar in the case of international moves in the fi rst period (+0.945), but less so in the second one (+0.678) due to the already mentioned relative decrease of the share of women in immigration fl ows after 2014. Whereas for 2004-2008 a generally high similarity between the internal net mi- gration rates and the age group-specifi c ones is observed (over +0.600 for all age groups), in all other cases (international net migration rates of the two periods and internal net migration rates for 2013-2017), the differences in the patterns of the age group-specifi c net migration rates and the total one are more accentuated and close to zero at the age of retirement. Since most migrants are young adults, their geographic patterns should determine the patterns of the total. Whereas internal net migration in the pre-crises period did show less geographic variability, the geo- graphic patterns of international net migration by age group and for 2013-2017 show a great variability, indicating a variety of factors infl uencing these patterns. The migration patterns of the case study LLMAs are summarised in Table 2 and show the existing geographic disparities and the persistence as well as the dynam- ics of the migration patterns of the LLMAs and over time. All Central-Northern LL- MAs have a positive internal migration balance in the second period, except for the Fig. 6: Correlation coeffi cients of international and internal net migration rates by age group, 2004-2008 and 2013-201 0.0 0.2 0.4 0.6 0.8 1.0 0-4 15-19 30-34 45-49 60-64 75+ Age groups 2004-08 Internal 2013-17 Internal 2004-08 International 2013-17 International Correlation coefficients Source: Authors’ calculations based on Istat: ”Iscrizioni e cancellazioni all’anagrafe per trasferimento di residenza”; population estimates by citizenship, sex and age [http://demo.istat.it/] Population Change and International and Internal Migration in Italy • 513 T a b . 2 : S e le c te d I ta lia n L L M A s – in te rn a ti o n a l a n d in te rn a l i n -, o u t- a n d n e t m ig ra ti o n r a te s, 2 0 0 4 -2 0 0 8 a n d 2 01 3 -2 01 7 (p e r 1, 0 0 0 ) S o u rc e : A u th o rs ’ ca lc u la ti o n s b as e d o n I st at : ”I sc ri zi o n i e c an ce lla zi o n i al l’a n ag ra fe p e r tr as fe ri m e n to d i re si d e n za ”; p o p u la ti o n e st i- m at e s b y c it iz e n sh ip , se x a n d a g e [ h tt p :/ /d e m o .is ta t. it /] 2 0 0 4 -2 0 0 8 2 0 13 -2 0 17 In te rn a l m ig ra ti o n In te rn a ti o n a l m ig ra ti o n In te rn a l m ig ra ti o n In te rn a ti o n a l m ig ra ti o n In - O u t- N e t In - O u t- N e t In - O u t- N e t In - O u t- N e t M e tr o p o li ta n L L M A s M ila n 2 9 .0 3 1. 2 -2 .2 9 .5 1. 1 + 8 .3 2 7. 7 2 5 .4 + 2 .3 7. 3 2 .8 + 4 .5 G e n o a 13 .5 14 .9 -1 .4 7. 1 1. 0 + 6 .1 13 .6 14 .0 -0 .4 5 .0 2 .1 + 3 .0 B o lo g n a 3 5 .9 3 2 .2 + 3 .7 9 .2 0 .9 + 8 .2 3 2 .4 2 7. 9 + 4 .5 6 .8 2 .7 + 4 .1 R o m e 18 .6 19 .5 -0 .9 10 .1 1. 2 + 8 .9 17 .7 16 .0 + 1. 7 6 .7 2 .2 + 4 .4 N a p le s 2 1. 6 2 8 .0 -6 .4 2 .4 0 .3 + 2 .1 2 0 .6 2 4 .0 -3 .4 3 .2 1. 3 + 1. 9 B a ri 12 .6 15 .9 -3 .3 2 .4 0 .6 + 1. 8 13 .2 15 .6 -2 .3 3 .1 1. 6 + 1. 5 P a le rm o 18 .0 2 1. 6 -3 .6 2 .2 0 .9 + 1. 3 17 .0 19 .5 -2 .5 2 .3 2 .2 + 0 .1 L L M A s w it h m ig ra ti o n l o ss Is e rn ia 16 .5 17 .9 -1 .4 4 .5 0 .8 + 3 .6 18 .0 2 0 .2 -2 .2 5 .9 2 .4 + 3 .4 T o rr e d e l G re co 18 .2 2 6 .8 -8 .6 2 .3 0 .3 + 1. 9 18 .2 2 2 .4 -4 .2 1. 9 1. 0 + 0 .9 F o g g ia 8 .8 14 .8 -6 .1 3 .5 0 .4 + 3 .1 8 .1 12 .8 -4 .7 4 .0 1. 0 + 3 .0 M e lfi 8 .6 13 .1 -4 .5 3 .8 0 .8 + 3 .0 9 .6 13 .2 -3 .6 3 .8 1. 7 + 2 .0 R e g g io C a la b ri a 9 .4 12 .5 -3 .2 5 .0 0 .5 + 4 .5 9 .0 13 .4 -4 .4 4 .3 1. 8 + 2 .5 V ib o V a le n ti a 14 .3 19 .7 -5 .5 5 .2 0 .9 + 4 .3 15 .6 19 .6 -4 .0 4 .6 1. 1 + 3 .5 C a lt a n is se tt a 8 .7 12 .8 -4 .1 3 .0 1. 5 + 1. 6 9 .8 14 .4 -4 .6 8 .0 2 .2 + 5 .9 G e la 8 .5 14 .8 -6 .3 2 .7 2 .8 + 0 .0 6 .6 12 .0 -5 .4 2 .8 3 .0 -0 .2 Tu sc a n L L M A s F lo re n ce 2 3 .8 2 6 .6 -2 .8 9 .3 0 .9 + 8 .4 2 6 .0 2 3 .1 + 2 .8 7. 8 2 .1 + 5 .8 L iv o rn o 14 .5 14 .4 + 0 .1 5 .8 0 .6 + 5 .2 15 .9 14 .3 + 1. 6 4 .7 2 .1 + 2 .6 P is a 2 9 .0 2 6 .4 + 2 .6 7. 4 1. 0 + 6 .4 2 9 .0 2 7. 2 + 1. 8 6 .3 2 .3 + 4 .0 P ra to 2 2 .6 2 2 .8 -0 .2 10 .2 0 .6 + 9 .5 2 1. 6 2 1. 1 + 0 .5 8 .4 2 .0 + 6 .4 “O u tl ie r” L L M A s D e se n za n o d e l G a rd a 4 3 .9 3 5 .0 + 8 .9 11 .9 1. 0 + 10 .9 3 5 .8 3 1. 2 + 4 .7 5 .9 3 .7 + 2 .2 M e ra n 2 7. 8 2 4 .2 + 3 .6 9 .8 2 .6 + 7. 2 2 8 .5 2 3 .6 + 4 .9 7. 2 5 .4 + 1. 8 L’ A q u ila 17 .1 16 .9 + 0 .2 6 .5 0 .7 + 5 .8 19 .2 2 2 .3 -3 .2 6 .3 1. 9 + 4 .5 It a ly - - - 6 .9 0 .9 6 .0 - - - 5 .0 2 .4 + 2 .6 • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci514 old industrial city of Genoa. It seems that the economic crises amplifi ed the eco- nomic weaknesses of the Mezzogiorno, leading to migration gains in the North. The special cases are two attractive tourist destinations and the LLMA of L’Aquila, which was hit by an earthquake in 2009. Regarding international migration, the gains de- creased but continued to play an important role in regional population change, con- tributing to population growth or at least limiting losses due to negative internal net migration. When observing the single case studies, it is obvious that a great geographic var- iability exists regarding the impact of internal and international migration on region- al population change. Among the selected metropolitan areas, the high migration gain of Milan in 2013-2017 for young adults (Fig. 7a), a turnaround from the internal losses of the earlier period due to sub-urbanisation processes, contrasts with the consistent losses of the LLMA of Naples (Fig. 7c). Bari and Palermo, the other two Southern metropolitan LLMAs, show similar patterns of the process of age selectiv- ity of internal migration. In the region of Rome, similarly to Milan, internal migration gains increased for most age groups between 2004-2008 and 2013-2017 (Fig. 7b). However, migration losses for younger and older persons hint at the persistent se- lectivity of migration fl ows. The international migration gains of the Rome LLMA diminished during the study period in all age groups. University towns such as Pisa and Bologna show high migration gains for young adults. Foggia is an example of the other side of the coin: an LLMA that consistently loses young adults and does not even seem to take advantage of return migrants (Fig. 7d). While international immigration leads to a generalised migration gain in all categories, the demographic effect is most impressive in the metropolitan areas of Central-Northern Italy, with extremely high gains for the young foreign adults. In the two periods, all selected LLMAs show a gain from international migration. In some cases the co-location of institutions for persons in search of international protection leads to extremely high, if often only temporary, net migration gains. Regarding the observation that women have a higher propensity to migrate, we already mentioned that Ravenstein seems not to have conducted his analyses using a reference population or a population at risk. When analysing age-specifi c patterns of migration intensities (Bonifazi et al 2020), it becomes clear that women anticipate many life events including migratory moves, but that over a lifetime it seems diffi - cult to postulate higher propensities for women in the case of Italy. Only in the case of foreign women an exception was noted in recent years: higher propensities to migrate of older adult women due to their engagement as care providers for indi- viduals and families. However, focusing on the change of residence from and to the selected seven metropolitan LLMAs, we observe an overall higher propensity to migrate for men (Fig. 8). However, the differences between men and women vary between the LL- MAs, between the in-migration and out-migration fl ows, and between the two peri- ods considered. Italian men consistently seem to migrate more than Italian women. Regarding the foreign population, the patterns completely change between 2004- 2008 and 2013-2017. In the fi rst period, foreign men show a clearly higher propen- sity to migrate (even though surprising exceptions such as the infl ows to the LLMAs Population Change and International and Internal Migration in Italy • 515 of Bari and Genoa are observed). In the second period, the situation changes, as the sex differences are less strong, and women seem to migrate more than men (with the exception of in-migration to Naples and out-migration from Naples and Palermo). As for the international migration fl ows (Fig. 9), Italian men migrate more than Italian women. For the foreign population, the immigration rates completely change between 2004-2008 and 2013-2017. The sex distribution varied greatly between LL- MAs for 2004-2008 in particular. In the second period, the situation changed and the differences between women and men are more apparent and always in favour of the latter. Regarding the international emigration of foreigners, the LLMAs of the Centre-North show a higher propensity of women, while the opposite is true for the LLMAs of the South in both periods. Fig. 7: International and internal net migration patterns by age-group of selected Italian LLMAs, 2004-2008 and 2013-2017 (per 1,000) -10 -5 0 5 10 15 20 25 30 35 40 0-4 10-14 20-24 30-34 40-44 50-54 60-64 70-74 Age groups 2004-2008 Internal 2004-2008 International 2013-2017 Internal 2013-2017 International A - Milan Net migration rates (per 1,000) -10 -5 0 5 10 15 20 25 30 35 40 0-4 10-14 20-24 30-34 40-44 50-54 60-64 70-74 Age groups 2004-2008 Internal 2004-2008 International 2013-2017 Internal 2013-2017 International B - Rome Net migration rates (per 1,000) -30 -25 -20 -15 -10 -5 0 5 10 15 20 0-4 10-14 20-24 30-34 40-44 50-54 60-64 70-74 Age groups 2004-2008 Internal 2004-2008 International 2013-2017 Internal 2013-2017 International C - Naples Net migration rates (per 1,000) -30 -25 -20 -15 -10 -5 0 5 10 15 20 0-4 10-14 20-24 30-34 40-44 50-54 60-64 70-74 Age groups 2004-2008 Internal 2004-2008 International 2013-2017 Internal 2013-2017 International D - Foggia Net migration rates (per 1,000) Source: Authors’ calculations based on Istat: “Iscrizioni e cancellazioni all’anagrafe per trasferimento di residenza”; population estimates by citizenship, sex and age [http://demo.istat.it/] • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci516 The development of the communication and transportation infrastructure and the existing migration networks between the South and the administrative and eco- nomic centres of the Centre-North certainly offset the deterrent effect of distance. However, most migrants within Italy still only move a short distance. Since Italian data on changes of residence refer to the municipality as smallest territorial unit, not all changes of residence over short distances are included in the present analysis. As for Ravenstein, the main focus is on longer-distance migra- tion fl ows. In 2013-2017, 34 percent of migratory fl ows took place below the 10km threshold (distances of intra-LLMA migration fl ows are estimated: 1/2 of the radius of the circle with the LLMA’s surface area), 72 percent below 50km, 79 percent be- low 100km, and only 13 percent exceed 300km. This leads to an average migration distance of 115km in the 2013-2017 period in total (slightly down from 121km in Fig. 8: Internal in-migration and out-migration of the metropolitan Italian LLMAs by citizenship and sex, 2004-2008 and 2013-2017 (per 1,000) In-migration 2004-2008 Internal in-migration rates 2004-2008 (per 1,000) In-migration 2013-2017 Internal in-migration rates 2013-2017 (per 1,000) Out-migration 2004-2008 Internal out-migration rates 2004-2008 (per 1,000) Out-migration 2013-2017 Internal out-migration rates 2013-2017 (per 1,000) 0 20 40 60 80 100 120 Milan Genoa Bologna Rome Naples Bari Palermo Italian men Italian women Foreign men Foreign women 0 20 40 60 80 100 120 Milan Genoa Bologna Rome Naples Bari Palermo 0 20 40 60 Milan Genoa Bologna Rome Naples Bari Palermo 0 20 40 60 Milan Genoa Bologna Rome Naples Bari Palermo Source: Authors’ calculations based on Istat: ”Iscrizioni e cancellazioni all’anagrafe per trasferimento di residenza”; population estimates by citizenship, sex and age [http://demo.istat.it/] Population Change and International and Internal Migration in Italy • 517 2004-2008), 120km for Italians and 92km for foreign citizens. When excluding the migration fl ows between municipalities within a single LLMA, the respective val- ues roughly double, leading to an average migratory distance of 211km, down from 224km. The average migratory distances range between 95km and 135km (179km and 244km when excluding intra-LLMA changes of residence). Distances increase with age for young people, reaching their highest values for the 25-29-years-olds, then decreasing until the 50-54 age group, followed by another rise, except for 75-year-olds and above. Between 2004-2008 and 2013-2017, the average distances migrated diminished for the under-30-year-olds. It is certainly true for Italy that most migrants proceed only a short distance, with the important exception of migration fl ows from the Mezzogiorno to some metro- Fig. 9: International immigration and emigration rates of the metropolitan Italian LLMAs by citizenship and sex, 2004-2008 and 2013-2017 (per 1,000) Immigration 2004-2008 International immigration rates 2004-2008 (per 1,000) Immigration 2013-2017 International immigration rates 2013-2017 (per 1,000) Emigration 2004-2008 International emigration rates 2004-2008 (per 1,000) Emigration 2013-2017 International emigration rates 2013-2017 (per 1,000) 0 2 4 6 8 10 12 Milan Genoa Bologna Rome Naples Bari Palermo 0 2 4 6 8 10 12 Milan Genoa Bologna Rome Naples Bari Palermo 0 20 40 60 80 100 120 140 160 180 Milan Genoa Bologna Rome Naples Bari Palermo 0 20 40 60 80 100 120 140 160 180 Milan Genoa Bologna Rome Naples Bari Palermo Italian men Italian women Foreign men Foreign women Source: Authors’ calculations based on Istat: ”Iscrizioni e cancellazioni all’anagrafe per trasferimento di residenza”; population estimates by citizenship, sex and age [http://demo.istat.it/] • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci518 politan areas of the Centre-North, e.g. Turin, Milan, Bologna, or Rome. This con- fi rms Ravenstein’s observation that long-distance migrations are usually directed towards the larger metropolitan areas (“great centres of commerce and industry”). For Italy, this observation can be confi rmed – at least for the industrial and commer- cial centres of the Centre-North and for the LLMA of Milan in particular. However, the ubiquity of information and the existence of migratory networks make it easier for internal as well as international migrants to choose the destination they desire, so the concept of the great centres of commerce and industry has to be adapted to the present day realities. Specifi c functions of an LLMA (higher education, for exam- ple) might be of greater importance for attracting long-distance migrants, as well as the presence of specifi c sub-divisions of the tertiary sectors. Finally, we analyse socio-economic aspects of Italian international and internal migration patterns following the implicit comments of Ravenstein regarding the drivers of migration when mentioning the “great centres of commerce and indus- try”. Rural areas no longer play a signifi cant role in the migration system due to the decreasing importance of the Italian agricultural sector (Bonifazi/Heins 2000) and the ensuing out-migration from rural areas. Table 3 reports selected results and confi rms the importance of the socio-economic dimension (represented by the incidence of the foreign population (%) and other information of economic well- being) for the geographical patterns of net migration rates. Figure 10 shows the two most telling examples: the international net migration rates in 2004-2008 and the internal net migration rates in 2013-2017. Both show a positive relationship be- tween economic activity rates (for the 20-64-year-olds) and net migration rates. The activity rates show a clear difference between the Mezzogiorno (with lower values) and the, on average, economically better-off Central-Northern Italian LLMAs (with higher ones). International migration gains in the years preceding the economic crises are clearly determined by the economic and labour situation of the LLMAs, although important local variations are observed. These are attributable to the important im- migration fl ows of this period and local labour market specifi cities regarding, for ex- ample, caregiving and the agricultural sector, where many foreigners are employed. In the more recent period (2013-2017), characterised by lower immigration fl ows, the correlation coeffi cients are lower. (Clearly, the correlation analysis only assess- es a linear relationship between the distributions, a limitation when considering the relationship between net migration rates and socio-economic variables.) Regarding internal net migration, the positive relationship with the activity rates (+0.580) in 2004-2008 intensifi ed in 2013-2017 (+0.788), probably due to lower variability and a clearer linear relationship. In general, the combination of different socio-economic information does not increase the share of the geographic differences of the net mi- gration rates “explained” by them. Based on the role played by population density in the Italian settlement system, it is not surprising that the correlation coeffi cients are relatively low. In the pre-crises period, a positive migration balance could be observed for mid-sized LLMAs to the detriment of high- and low-density LLMAs. However, the geographical variability is so high that it would be incorrect to de- fi ne this period as counter-metropolitan or characterised by movement away from Population Change and International and Internal Migration in Italy • 519 T a b . 3: C o rr e la ti o n c o e ffi c ie n ts b e tw e e n n e t m ig ra ti o n p a tt e rn s an d s o ci o -e co n o m ic v a ri a b le s, 6 11 I ta lia n L L M A s, 2 0 0 4 - 2 0 0 8 a n d 2 01 3 -2 01 7 E co n o m ic E co n o m ic E co n o m ic P o p u la ti o n P o p u la ti o n F o re ig n 2 0 -6 4 -y e a r- o ld s P o p u la ti o n a c ti v it y a c ti v it y a c ti v it y d e n si ty , 2 0 -4 4 , p o p u la ti o n e m p lo y e d 2 0 -4 4 ra te s, 2 0 - ra te s 15 + , ra te s 15 + , 2 0 11 2 0 11 ( % ) (% ) in t h e s e rv ic e liv in g a s 6 4 , 2 0 11 2 0 0 6 -2 0 0 8 2 0 13 -2 0 17 (L o g 10 ) se c to r, 2 0 11 so n o r (% ) (% ) (% ) (% ) d a u g h te r (% ) In te rn a l n e t m ig ra ti o n r a te s (p e r 1, 0 0 0 ) T o ta l p o p u la ti o n 2 0 0 4 -2 0 0 8 + 0 .5 8 0 ** + 0 .6 3 1 * * + 0 .5 7 5 * * -0 .2 2 6 * * -0 .1 8 8 * * + 0 .4 8 6 * * -0 .0 6 3 -0 .4 8 1 * * T o ta l p o p u la ti o n 2 0 13 -2 0 17 + 0 .7 8 8 ** + 0 .7 6 7 ** + 0 .7 6 4 ** + 0 .1 7 9 ** -0 .3 4 6 ** + 0 .6 5 2 * * + 0 .4 9 9 * * -0 .7 3 6 * * In te rn a ti o n a l n e t m ig ra ti o n r a te s (p e r 1, 0 0 0 ) T o ta l p o p u la ti o n 2 0 0 4 -2 0 0 8 + 0 .7 3 8 ** + 0 .7 5 3 ** + 0 .7 7 7 * * + 0 .1 15 * * -0 .2 5 2 * * + 0 .9 0 3 * * + 0 .3 7 9 * * -0 .7 18 * * T o ta l p o p u la ti o n 2 0 13 -2 0 17 + 0 .3 4 4 ** + 0 .3 16 ** + 0 .3 4 7 ** + 0 .2 2 0 ** -0 .0 9 9 * + 0 .4 6 9 * * + 0 .4 8 2 * * -0 .4 15 * * ** s ig n ifi c an t at t h e 0 .0 1 le v e l a n d * s ig n ifi ca n t at t h e 0 .0 5 le v e l. S o u rc e : A u th o rs ’ c al cu la ti o n fr o m Is ta t: ” Is cr iz io n i e c an ce lla zi o n i a ll’ an ag ra fe p e r tr as fe ri m e n to d i r e si d e n za ”; p o p u la ti o n e st im at e s b y ci ti ze n sh ip , s e x a n d a g e [ h tt p :/ /d e m o .is ta t. it /] . I st at : 1 5 th P o p u la ti o n a n d H o u si n g C e n su s 2 0 1 1 a n d Is ta t: “ O cc u p at i r e si d e n ti e p e rs o n e i n c e rc a d i o cc u p az io n e n e i S is te m i lo ca li d e l la v o ro ” [h tt p s: // w w w .is ta t. it /i t/ in fo rm az io n i- te rr it o ri al i- e -c ar to g ra fi ch e / si st e m i- lo ca li- d e l- la v o ro ] • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci520 metropolitan areas. Rather, employment in the private service sector (along with a general higher level of educational attainment), which is higher in the metropolitan areas of the Centre-North, is positively correlated with net migration rates in 2013- 2017. This is probably an indication that the LLMAs with a higher share of employ- ment in the private service sector were more resilient in the face of recent economic crises, compared to manufacturing-dependent LLMAs. (In a certain way, the discus- sion regarding population density does apply.) As with the economic variables, the data regarding young adults living with their parents is characterised by a North/ South gradient, with a higher share of 20-44-year-olds living with their parents in the Mezzogiorno and with a lower or negative migration balance (although regional specifi cities do exist). Finally, in lieu of a summary, some results of a simple regression analyses of net migration fl ows between the case study LLMAs and all other LLMAs are discussed. These population-weighted regression analyses for net migration rates for single migration fl ows use the independent variables listed in Annex 2. The analyses of the single net migration fl ows confi rm the general patterns al- ready discussed. The quantitatively most important net migration fl ows and the most unbalanced ones are those to neighbouring LLMAs. For the Milan LLMA, these include Como, Bergamo, Busto Arsizio, Pavia, Crema, Lodi, Lecco, Vigevano; for the Rome LLMA, Pomezia (bordering Rome to the south, it gained 16,200 individuals through the migratory exchange with Rome in 2004-2008, but only 2,000 in 2013- 2017; in total it has the highest volume and net migration in the Italian case), Viterbo, and Rieti; but the same is true also for the Tuscan LLMAs and the other larger LL- Fig. 10: Two examples of the relationship between net migration rates (per 1,000) and economic activity rates (per 100), Italian LLMAs Source: Authors’ calculations based on Istat: ”Iscrizioni e cancellazioni all’anagrafe per trasferimento di residenza”; population estimates by citizenship, sex and age [http://demo.istat.it/]. Istat: 15th Population and Housing Census 2011 -5 0 5 10 15 50 55 60 65 70 75 80 85 Economic activity rates of 20-64 years old, 2011 (%) 10,000 residents or less 10,001 to 50,000 residents 50,001 to 100,000 residents 100,001 to 500,000 residents 500,001 or more residents International net migration rates 2004-2008 (per 1,000) International net migration rates 2004-2008 -10 -5 0 5 10 15 50 55 60 65 70 75 80 85 Economic activity rates of 20-64 years old, 2011 (%) Internal net migration rates 2013-2017 (per 1,000) Internal net migration rates 2013-2017 Desenzano del Garda Rome Bologna Florence Prato Milan TorinoVenezia Busto Arsizo ComoPadova Livorno Genoa Cagliari Melfi Isernia Reggio Calabria Catania Palermo Gela Naples Vibo Valetia Bari Foggia Bologna MeranDesenzano del Garda Livorno Florence Como Bergamo Busto Arsizo Padova Prato Rome Venezia Genoa Cagliari L'Aquila Catania Bari Palermo Vibo Valentia Melfi Isernia FoggiaCalttanissettaNaplesGela Torre del Greco Population Change and International and Internal Migration in Italy • 521 MAs. Between the two periods (2004-2008 and 2013-2017), the net migration loss of the metropolitan LLMAs diminished signifi cantly, leading to a slowdown of the suburbanisation process. However, in the case of the Southern LLMA of Naples, the net migration loss to Rome and Milan is far more important than the suburban net migration losses to the neighbouring LLMAs of Caserta and Nola. Interestingly, some of the long-distance net migration losses towards the metropolitan LLMAs of Central-Northern Italy increased after the economic downturn. This is especially true for the net migration fl ows to the advantage of Milan, but also Turin, Florence, and Rome as the largest LLMAs. The interesting case of the net migration fl ow be- tween Rome and Milan increased in its volume and changed from a quite balanced fl ow with a slight gain for Milan in the entire fi rst period (about 500 persons) to a net gain of more than 3,100 in the second period. In general, the internal net migra- tion gains increased for Milan between the two periods and the Italian migration system seems to be more focused on Milan today than before the great economic recession, when fl ows were more balanced and general net migration losses of the metropolitan LLMAs to their surroundings could be observed. The metropolitan LLMAs of the Mezzogiorno generally only have few net migra- tion fl ows leading to a net migration gain, mostly these are fl ows that are linked to their function as regional capitals (Naples, Bari, Palermo, and Reggio Calabria). LLMAs such as Vibo Valentia (Calabria), Caltanissetta and Gela (Sicily) only have negative net migration fl ows. All these LLMAs have been losing population – espe- cially after the start of the crises – to Milan, Turin, Rome, and the other economic centres of Central-Northern Italy. Traditional links between origins and destinations, such as a preference for the Adriatic corridor for areas in Apulia to Bologna and the Northeast, or Rome and Lombardy for the Sicilian areas, seem to have played a smaller role in 2013-2017. For the Tuscan LLMAs, we observe a net migratory loss only for neighbouring or suburban LLMAs, or even between the LLMAs included here as case studies, such as the net fl ows between Florence and Prato or Pisa and Livorno, where a turnaround in net gains (or net losses) is observed between the two periods. In the case of Pisa in particular, the presence of installations of the military marine could explain the migratory exchange with places such as La Spezia and Taranto. The two special cases of Desenzano del Garda and Meran are gaining population from their local or regional migration fi eld and through positive net migration fl ows from Milan, Rome, and Naples. For L’Aquila, an increasing net migration loss to Pescara, as well as Milan and Rome, is observed. The results of regression models for net migration patterns of single LLMAs are partially inconclusive (Table 4). The adjusted r2 of the regression models seems to be too low for drawing reliable conclusions in the great majority of instances. Only in the cases of Milan and Rome do the results seem to be more robust, because their migration fi eld spans the entire nation and the remotest and smallest LLMAs are connected to these centres by some migratory fl ows. The differences in the results of the two case studies are certainly due to their geographic locations, for which the differences in the results for the distance pa- rameter are an obvious indication. Moreover, while Milan and Rome are important centres, their roles in the Italian economic and migratory system are different. • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci522 r2 D is ta n ce P o p u la ti o n P o p u la ti o n P o p u la ti o n F o re ig n E co n o m ic U n e m p lo y - P o p u la ti o n 2 0 -6 4 - P o p u la ti o n (L o g 10 ) g ro w th , d e n si ty 2 0 -4 4 ( % ) p o p u la ti o n a c ti v it y r a te m e n t ra te 2 0 -6 4 w it h y e a r- o ld s 2 0 -4 4 l iv in g 2 0 0 2 -2 01 7 (L o g 10 ) (% ) 2 0 -6 4 ( % ) 2 0 -6 4 ( % ) a u n iv e rs it y e m p lo y e d a s so n o r d e g re e ( % ) in t h e d a u g h te r se rv ic e (% ) se c to r( % ) M il a n 2 0 0 4 -2 0 0 8 T o ta l 0 .5 2 6 -1 .0 8 8 0 .3 8 9 -0 .1 9 0 -0 .3 8 6 -0 .2 5 5 0 .1 3 7 0 .1 74 M e n 0 .5 3 1 -1 .1 0 1 0 .3 9 5 -0 .1 9 9 -0 .3 7 8 -0 .2 6 8 0 .1 4 3 0 .1 8 9 W o m e n 0 .5 0 5 -0 .9 4 8 0 .3 61 -0 .0 9 6 -0 .4 6 7 0 .1 5 8 It a lia n s 0 .5 17 -1 .0 19 0 .3 8 7 -0 .2 17 -0 .3 9 6 -0 .1 8 4 0 .1 3 7 0 .1 6 5 F o re ig n e rs 0 .5 0 5 -1 .0 0 6 0 .2 7 6 -0 .1 8 4 -0 .2 12 -0 .1 8 9 0 .1 3 7 0 .2 0 1 2 0 13 -2 0 17 T o ta l 0 .6 6 4 -0 .6 4 6 0 .4 7 9 -0 .2 8 6 -0 .2 8 9 0 .2 51 -0 .2 16 0 .2 0 8 M e n 0 .6 4 7 -0 .6 13 0 .4 3 4 -0 .1 2 0 -0 .2 3 3 -0 .2 5 9 0 .2 4 7 -0 .2 8 5 0 .1 5 2 0 .1 7 7 W o m e n 0 .6 5 5 -0 .6 61 0 .4 5 2 -0 .2 6 3 -0 .2 9 2 0 .2 4 6 -0 .2 9 5 0 .1 0 9 0 .2 17 It a lia n s 0 .6 6 8 -0 .6 4 9 0 .4 5 9 -0 .2 6 4 -0 .2 5 9 0 .2 5 7 -0 .2 19 0 .2 18 F o re ig n e rs 0 .2 0 7 -0 .2 0 3 0 .2 9 0 -0 .2 5 7 -0 .1 0 1 R o m e 2 0 0 4 -2 0 0 8 T o ta l 0 .4 7 0 -0 .5 0 8 0 .4 5 5 -0 .3 0 5 -0 .1 8 0 -0 .4 3 0 0 .4 6 8 M e n 0 .4 7 9 -0 .5 3 0 0 .4 3 5 -0 .3 2 7 -0 .1 5 9 -0 .4 16 0 .4 6 8 W o m e n 0 .4 5 7 -0 .4 8 0 0 .4 74 -0 .2 7 9 -0 .2 0 4 -0 .4 4 2 0 .4 6 5 It a lia n s 0 .4 6 7 -0 .4 9 6 0 .5 2 1 -0 .3 0 8 -0 .2 3 0 -0 .1 0 1 -0 .4 51 0 .4 8 2 F o re ig n e rs 0 .4 4 5 -0 .5 4 2 0 .3 8 7 -0 .3 0 5 -0 .1 3 0 -0 .3 3 1 -0 .3 19 0 .4 16 2 0 13 -2 0 17 T o ta l 0 .6 2 2 0 .4 5 4 -0 .4 16 0 .2 8 4 -0 .4 5 0 0 .3 6 9 M e n 0 .5 8 6 -0 .1 2 1 0 .4 3 2 -0 .1 2 0 -0 .3 9 1 0 .2 7 9 -0 .4 8 2 0 .4 6 6 W o m e n 0 .6 10 0 .3 74 0 .0 9 3 -0 .3 4 7 -0 .1 6 5 0 .3 8 1 -0 .4 2 5 0 .2 6 7 -0 .1 7 5 It a lia n s 0 .5 9 4 -0 .0 4 8 0 .4 3 4 -0 .4 10 0 .2 9 7 -0 .4 3 1 0 .3 4 9 F o re ig n e rs 0 .2 3 0 -0 .1 18 0 .2 0 9 -0 .1 6 3 -0 .2 3 5 S o u rc e : A u th o rs ’ c al cu la ti o n fr o m Is ta t: ” Is cr iz io n i e c an ce lla zi o n i a ll’ an ag ra fe p e r tr as fe ri m e n to d i r e si d e n za ”; p o p u la ti o n e st im at e s b y c it iz e n sh ip , s e x an d a g e [ h tt p :/ /d e m o .is ta t. it /] . Is ta t: 1 5 th P o p u la ti o n a n d H o u si n g C e n su s 2 0 1 1 . O n ly s ig n ifi ca n t re g re ss io n c o e ffi c ie n ts a t th e 0 .0 5 le v e l ( fu ll m o d e l) a re r e p o rt e d . T h e s o ci o -e co n o m ic v ar ia b le s re fe r to t h e s it u at io n in t h e o th e r LL M A s. F o r th e d e fi n it io n o f v ar ia b le s se e A n n e x 2 T a b . 4: R e g re ss io n r e su lt s o f in te rn a l n e t m ig ra ti o n fl o w s o f th e L L M A s o f M ila n a n d R o m e w it h t h e o th e r 61 0 I ta lia n L L M A s, 2 0 0 4 -2 0 0 8 a n d 2 01 3 -2 01 7 Population Change and International and Internal Migration in Italy • 523 Rome has a more administrative and political, Milan a more economic function. In fact, distance plays a smaller role in the case of Rome. However, the distance pa- rameter is systematically lower in the second period and also has a negative effect (lower migratory gains or losses to areas of high population density disappear in the second period). The regression results linked to population growth and age struc- ture, as well as the presence of the foreign population and the activity rates, refl ect general characteristics of areas of migration gains or losses. They do not suggest direct determinants of net migration patterns. Interestingly, the presence of foreign population seems to infl uence the net migration patterns most strongly in the case of Milan. At the same time, unemployment does not emerge in any case as a factor of migration gains or losses. In general, no fundamental differences between the results for women and men are observed. However, the results for the foreign popu- lation seem to indicate different factors that determine the net migration patterns, factors that are not captured through the set of socio-economic variables here. As observed in another context (Greenwood 2019), individuals, families, and places are heterogenous, and the increasing segmentation of society and the hous- ing and labour markets makes observing generalised migratory behaviour more diffi cult. The challenge is not only one of analytical precision, but also in the fact that social (and migratory) processes are less “standardised”. Greenwood (2019: 273) also offers an interesting discussion regarding the role of differential economic opportunities, such as those in Italy today. However, in the Italian context, the set of potential motivations for internal mobility is so varied and complicated that it seems to have escaped the attention of modellers. 5 Summary of fi ndings Ravenstein’s empirical approach seems absolutely valid and useful today when ap- plied to present-day Italy, although some of his observations are – naturally – a prod- uct of the demographic and socio-economic situation of his times, the geographical setting, and data availability. Regarding migration patterns, Italy seems to present a special case not easily comparable to other European countries. Therefore it would seem inappropriate to expect Ravenstein’s 130-year-old “laws” to fully hold when applied to present-day Italy. This contribution attempted to focus on the empirical method followed by Ravenstein, for which he had great sense and sensibility. Two specifi c factors predominantly affected recent aspects of Italian migration trends: First, around the turn of the millennium, international migration fl ows completely changed the Italian setting. Increasing immigration and the specifi c migration pat- terns of the foreign population due to their socially and economically more pre- carious living conditions (e.g. higher migration intensity, shorter average distance) continue to drive a general change in migration patterns. In addition with the great economic recession, the emigration of Italians is once again playing a role as an alternative to internal migration fl ows, because internal migration is associated with relatively high indirect and direct costs due to the rigidities of the housing and la- • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci524 bour markets. These circumstances make emigration to study and work in another EU country attractive, while freedom of movement in the EU makes it widely fea- sible. In recent years, the number of Italian emigrants both less and highly qualifi ed increased, following existing networks of Italians abroad that were created by ear- lier waves of emigration as well as new destinations. Second, the growth of the metropolitan LLMAs of Central-Northern Italy seems to have picked up in the wake of the Great Recession, even if this growth is generally more directed to the centres than the peripheries. After an apparent stagnation of internal migration gains, or even losses, in the years up to the crises, these metro- politan areas now again show internal migration gains. Our results show that more men than women migrate internationally; in the case of internal moves the rates of migration by men and women are similar; internal migration is more important for the foreign than the Italian population; and inter- national migration gains contribute substantially more to overall population change than internal migration gains and losses do. Not all areas in Italy characterised by high unemployment show an effect of dis- persion, and distance does not seem to be a deterrent to migrate. The geographical- ly detailed analysis illustrates the temporal and spatial coexistence of many diverse international and internal migration processes according to local socio-economic characteristics. Notwithstanding the geographic heterogeneity we observe regard- ing the nexus between the international and internal migration patterns and region- al population change, we must acknowledge the importance of the economic and administrative centres of Italy (e.g. Milan, Rome, Naples, and Turin) as the driving force behind the national patterns. However three main fundamental differences between England in the 1870/1880s and Italy in the 2000s and 2010s exist: First, the type and detail of migration data are markedly different. Ravenstein could not analyse migration fl ows in great demographic detail (Greenwood 2019: 276) because the necessary data simply did not exist. For Italy, we were able to show how selective migration fl ows are – not only regarding sex, but also regarding citizenship and age. Unfortunately, other socio-demographic characteristics of the migrating individuals and families are not so accurate and continue to be diffi cult to obtain. Obviously the differences between place of birth and register data, dis- cussed in section 2, play a role. Second, today’s demand for labour in Italy is more segmented and specialised, or, in some cases, weaker and more diffuse. The openings of mines or of manufac- turing sites that created a demand for unspecifi c labour in Ravenstein’s analysis has long been supplanted regarding international migration by the demand for labour in caregiving, agriculture, and manufacturing – all of which in the Italian case often imply the “3 D’s”: dirty, dangerous, and demeaning. In the case of internal migra- tion, labour demand seems to play a minor or less specifi c role: Since the Great Recession, the low-wage sector has been the only sector with a certain dynamic, but it hardly offers opportunities that would infl uence internal migration decisions. Third and fi nally, knowledge about potential opportunities is now widely avail- able at one’s fi ngertips though modern communication and information technology. Population Change and International and Internal Migration in Italy • 525 This includes information about areas of destination and alternatives for internal and international migration. Even 130 years after their formulation, Ravenstein’s migration “laws” are still a valuable starting point in assessing and understanding migration processes and their role in regional population change in Italy. The authors agree with Green- wood’s conclusion (2019: 277): “Thus, in general we can conclude that Ravenstein provided a remarkable study of internal migration, the likes of which few have come close to matching.” – Including this study. Acknowledgements The authors are very grateful to the editors of the special issue (Prof. Philip Rees and Dr. Nikola Sander) for their support, help and patience. We would also like to thank the two anonymous reviewers for their excellent suggestions, which helped us to improve the paper considerably. We also appreciate very much the thorough and dedicated language editing provided by CPoS. References Alexander, J. Trent; Steidl, Annemarie 2012: Gender and the “Laws of Migration”: a reconsideration of Nineteenth-Century Patterns. In: Social Science History 36,2: 223- 241 [https://doi.org/10.1017/S0145553200011779]. Basile, Roberto; Causi, Marco 2007: Le determinanti dei fl ussi migratori nelle province ital- iane: 1991-2001. In: Economia & Lavoro 41,2: 139-159 [https://doi.org/10.7384/72341]. Basile, Roberto et al. 2018: The impact of immigration on the internal mobility of na- tives and foreign-born residents: evidence from Italy. In: Spatial Economic Analysis [https://doi.org/10.1080/17421772.2020.1729997]. Bell, Martin et al. 2002: Cross-national comparison of internal migration: issues and measures. In: Journal of the Royal Statistical Society: Series A (Statistics in Society) 165,3: 435-464 [https://doi.org/10.1111/1467-985X.00247]. Benassi, Federico; Heins, Frank; Tucci, Enrico 2019: Residential migrations in Italian metropolitan Local Labour Market Areas: spatial patterns and age-structure effects. In: Canepari, Eleonora; Crisci, Massimiliano (Eds.): Moving Around in Town Practices, Pathways and Contexts of Intra-Urban Mobility from 1600 to the Present Day. Viella Historical Research 15. Roma: Viella: 165-180. Biagi, Bianca; Dotzel, Kathryn R. 2018: Theoretical advances on interregional migration models. In: Biagi, Bianca et al. (Eds.): New Frontiers in Interregional Migration Re- search. Advances in Spatial Science. Cham: Springer International: 21-47 [https://doi. org/10.1007/978-3-319-75886-2_2]. Biagi, Bianca; Faggian, Alessandra; McCann, Philip 2011: Long and short distance mi- gration in Italy: the role of economic, social and environmental characteristics. In: Spa- tial Economic Analysis 6,1: 111-131 [https://doi.org/10.1080/17421772.2010.540035]. Billari, Francesco C.; Liefbroer, Aart C. 2010: Towards a new pattern of transition to adult- hood? In: Advances in Life Course Research 15,2-3: 59-75 [https://doi.org/10.1016/j. alcr.2010.10.003]. • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci526 Bonaguidi, Alberto; Terra Abrami, Valerio 1996: The pattern of internal migration: the Italian case. In: Rees, Philip H. et al. (Eds.): Population migration in the European Union. Chichester: John Wiley and Sons: 231-245. Bonifazi, Corrado 1992: Saldi migratori e studio delle migrazioni. In: Genus 48,1-2: 47-67. Bonifazi, Corrado (Ed.) 1999: Mezzogiorno e migrazioni interne. Irp-Cnr Monografi a 10. Roma: Irp-Cnr. Bonifazi, Corrado 2013a: L’Italia delle migrazioni. Bologna: Il Mulino. Bonifazi, Corrado 2013b: Mobile per forza. Spostamenti di popolazione nell’Italia della crisi. In: Il Mulino 13,5: 798-805 [https://doi.org/10.1402/74616]. Bonifazi, Corrado; Heins, Frank 2000: Long-term trends of internal migration in Italy. In: International Journal of Population Geography 6,2: 111-131 [https://doi.org/10.1002/ (SICI)1099-1220(200003/04)6:2<111::AID-IJPG172>3.0.CO;2-L]. Bonifazi, Corrado; Heins, Frank 2003: Testing the differential urbanisation model for Italy. In: Tijdschrift voor Economische en Sociale Geografi e 94,1: 23-37 [https://doi. org/10.1111/1467-9663.00234]. Bonifazi, Corrado; Heins, Frank 2009: Ancora migranti: la nuova mobilità degli italiani. In: Coorti, Paola; Sanfi lippo, Matteo (Eds.): Storia d’Italia. Annali 24. Migrazioni. Gran- di Opere. Torino: Giulio Einaudi: 505-528. Bonifazi, Corrado; Heins, Frank; Tucci, Enrico 2012: Le migrazioni interne degli stranieri al tempo dell’immigrazione. In: Meridiana. Rivista di storia e di scienze sociali 75,3: 173-190 [https://doi.org/10.1400/202294]. Bonifazi, Corrado et al. 2016: Le migrazioni interne in Italia nel 2013-2014: gli aspetti ter- ritoriali. In: Colucci, Michele; Gallo, Stefano (Eds.): Fare spazio. Rapporto 2016 sulle migrazioni interne in Italia. Roma: Donzelli: 3-23. Bonifazi, Corrado; Heins, Frank; Tucci, Enrico 2017: Italy: internal migration in a low-mo- bility country. In: Champion, Tony; Cooke, Thomas; Shuttleworth, Ian (Eds.): Internal Migration in the Developed World. Are we becoming less mobile? Oxon: Routeledge: 242-262. Bonifazi, Corrado et al 2020: The regional dynamics of internal migration intensities in Italy. In: Population, Space and Place e2331 [https://doi.org/10.1002/psp.2331]. Bubbico, Davide 2012: Le migrazioni interne dal Mezzogiorno tra ricerca di lavoro e mo- bilità occupazionale. In: Meridiana. Rivista di storia e di scienze sociali 75,3: 149-172 [https://doi.org/10.1400/202292]. Bubbico, Davide; Morlicchio, Enrica; Rebeggiani, Enrico (Eds.) 2011: Su e giù per l’Italia. La ripresa delle emigrazioni interne e le trasformazioni del mercato del lavoro. In: So- ciologia del Lavoro 121 [https://doi.org/10.3280/SL2011-121001]. Cellini, Roberto; Torrisi, Gianpiero 2014: Regional resilience in Italy: a very long-run analysis. In: Regional Studies 48,11: 1779-1796 [https://doi.org/10.1080/00343404.20 13.861058]. Champion, Tony 2001: Urbanization, suburbanization, counter-urbanization and re-ur- banization. In: Paddison, Ronan (Ed.): Handbook of Urban Studies. London/Thousand Oaks/New Delhi: Sage: 143-161 [https://doi.org/10.4135/9781848608375.n9]. Champion, Tony; Cooke, Thomas; Shuttleworth, Ian (Eds.) 2017: Internal migration in the Developed World. Are we becoming less mobile? Oxon: Routeledge. Coorti, Paola; Sanfi lippo, Matteo (Eds.) 2009: Migrazioni. Storia d’Italia. Annali 24. Gran- di Opere. Torino: Giulio Einaudi. Population Change and International and Internal Migration in Italy • 527 De Filippo, Elena; Strozza, Salvatore 2011: Le migrazioni interne degli stranieri in Italia. In: Sociologia del lavoro 121 (Su e giù per l’Italia – La ripresa delle migrazioni interne e le trasformazioni del mercato del lavoro): 168-195 [https://doi.org/10.3280/SL2011- 121010]. De Santis, Gustavo 2010: Mobilità a corto e lungo raggio e pendolarismo della popolazi- one italiana. In: Livi Bacci, Massimo (Ed.): Demografi a del capitale umano. Bologna: Il Mulino: 123-138. Faggian, Alessandra; Corcoran, Jonathan; Partridge, Mark 2015: Interregional migration analysis. In: Karlsson, Charlie; Andersson, Martin; Norman, Therese (Eds.): Handbook of research methods and applications in economic geography. Cheltenham: Edward Elgar: 473-495. Faggian, Alessandra et al. 2018: Regional economic resilience: the experience of the Italian local labor systems. In: The Annals of Regional Science 60,2: 393-410 [https:// doi.org/10.1007/s00168-017-0822-9]. Faggian, Alessandra; Rajbhandari, Isha; Dotzel, Kathryn R. 2017: The interregional mi- gration of human capital and its regional consequences: a review. In: Regional Studies 51,1: 128-143 [https://doi.org/10.1080/00343404.2016.1263388]. Faini, Riccardo et al. 1997: An empirical puzzle: falling migration and growing unemploy- ment differentials among Italian regions. In: European Economic Review 41,3-5: 571- 579 [https://doi.org/10.1016/S0014-2921(97)00023-8]. Ferrara, Antonella Rita; Nisticò, Rosanna 2015: Regional well-being indicators and dis- persion from a multidimensional perspective: evidence from Italy. In: The Annals of Regional Science 55,2: 373-420 [https://doi.org/10.1007/s00168-015-0704-y]. Fratesi, Ugo; Percoco, Marco 2014: Selective migration, regional growth and conver- gence: evidence from Italy. In: Regional Studies 48,10: 1650-1668 [https://doi.org/10.1 080/00343404.2013.843162]. Gallo, Stefano 2012: Senza attraversare le frontiere. Le migrazioni interne dall’Unità a oggi. Roma, Bari: Laterza. Golini, Antonio 1974: Distribuzione della popolazione, migrazioni interne e urbanizzazi- one in Italia. Roma: Istituto di Demografi a, Università di Roma. Greenwood, Michael J. 2019: The migration legacy of E.G. Ravenstein. In: Migration Studies 7,2: 269-278 [https://doi.org/10.1093/migration/mny043]. Grigg, David B. 1977: E. G. Ravenstein and the “laws of migration”. In: Journal of Histori- cal Geography 3,1: 41-54 [https://doi.org/10.1016/0305-7488(77)90143-8]. Impicciatore, Roberto; Strozza, Salvatore 2015: Migrazioni internazionali e interne di italiani e stranieri. In: De Rose, Alessandra; Strozza, Salvatore (Eds.): Rapporto sulla popolazione: l’Italia nella crisi economica. Bologna: Il Mulino: 109-140. Impicciatore, Roberto; Strozza, Salvatore 2016: Internal and international migration in Italy. An integrating approach based on administrative data. In: POLIS 30,2: 211-237 [https://doi.org/10.1424/83908]. Istat 2014: Generazioni a confronto. Come cambiano i percorsi verso la vita adulta. Roma: Istat [https://www.istat.it/it/fi les/2014/09/Generazioni-a-confronto.pdf, 05.08.2020]. Istat 2015: La nuova geografi a dei Sistemi Locali del Lavoro. Letture statistiche – Terri- torio. Roma: Istat [https://www.istat.it/it/fi les//2015/10/La-nuova-geografi a-dei-siste- mi-locali.pdf, 05.08.2020]. • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci528 King, Russell 2012: Theories and Typologies of migration: an overview and a primer. Willy Brandt Series of Working Paper in International Migrations and Ethnic Rela- tions 3/12. Malmö: Malmö Institute for Studies of Migration, Diversity and Welfare (MIM), Malmö University [https://www.mah.se/upload/Forskningscentrum/MIM/WB/ WB%203.12.pdf, 05.08.2020]. Lamonica, Giuseppe Ricciardo; Zagaglia, Barbara 2013: The determinants of internal mobility in Italy, 1995-2006: a comparison of Italians and resident foreigners. In: De- mographic Research 29,16: 407-440 [https://doi.org/10.4054/DemRes.2013.29.16]. Lee, Everett S. 1966: A theory of migration. In: Demography 3,1: 47-57 [https://doi. org/10.2307/2060063]. Macisco, John J.; Pryor, Edward T. 1963: A reappraisal of Ravenstein’s “Laws” of migra- tion: a review of selected studies of internal migration in the United States. In: The Amer- ican Catholic Sociological Review 24,3: 211-221 [https://doi.org/10.2307/3708238]. Mocetti, Sauro; Porello, Carmine 2010: How does immigration affect native internal mobility? New evidence from Italy. Working Paper 748. Roma: Banca d’Italia [https:// www.bancaditalia.it/pubblicazioni/temi-discussione/2010/2010-0748/en_tema_748. pdf, 05.08.2020] Nani, Michele 2013: Alla ricerca di “Leggi delle migrazioni”. Ernst Georg Ravenstein (1834-1913) e lo studio della mobilità fra statistica e cartografi a. In: Biorci, Grazia; Sini- gaglia, Roberto (Eds.): Dialoghi sulle migrazioni. Letteratura, storia e lingua. Genova: Genova University Press: 93-100. Panichella, Nazareno 2012: Le migrazioni interne nel secolo scorso: vecchie e nuove forme a confronto. In: Stato e mercato 95,2: 255-282 [https://doi.org/10.1425/37882]. Piché, Victor 2013: Contemporary migration theories as refl ected in their founding texts. In: Population 68,1: 141-164 [https://doi.org/10.3917/popu.1301.0153]. Piras, Romano 2017: A long-run analysis of push and pull factors of internal migration in Italy. Estimation of a gravity model with human capital using homogeneous and heterogeneous approaches. In: Papers in regional Science 96,3: 571-602 [https://doi. org/10.1111/pirs.12211]. Pugliese, Enrico 2006: L’Italia tra migrazioni internazionali e migrazioni interne. Bologna: Il Mulino. Pugliese, Enrico 2011: Le migrazioni interne nella scena migratoria italiana: novità, per- sistenze, luoghi comuni. In: Sociologia del lavoro 121 (Su e giù per l’Italia – La ripresa delle migrazioni interne e le trasformazioni del mercato del lavoro): 19-29 [https://doi. org/10.3280/SL2011-121002]. Ravenstein, Ernst Georg 1876: Census of the British Isles, 1871. Birthplace and Migra- tion. In: Geographical Magazine 3, July: 173-177, August: 201-206, September: 229- 233 [https://catalog.hathitrust.org/Record/000057985, https://babel.hathitrust.org/ cgi/pt?id=mdp.39015035573404, 03.09.2019]. Ravenstein, Ernst Georg 1885: The Laws of Migration. In: Journal of the Statistical Soci- ety 48,2: 167-227 [https://doi.org/10.2307/2979181]. Ravenstein, Ernst Georg 1889: The Laws of Migration – II. In: Journal of the Statistical Society 52,2: 214-301 [https://doi.org/10.2307/2979333]. Rees, Philip et al. 2017: The impact of internal migration on population redistribu- tion: an international comparison. In: Population, Space and Place 23,6 [https://doi. org/10.1002/psp.2036]. Rees, Philip et al. 1998: Internal migration and regional population dynamics in Italy. (Essays 3). Roma: Istat. Population Change and International and Internal Migration in Italy • 529 Rowe, Francisco et al. 2019: Impact of Internal Migration on Population Redistribution in Europe: Urbanisation, Counterurbanisation or Spatial Equilibrium? In: Comparative Population Studies 44: 201-234 [https://doi.org/10.12765/CPoS-2019-18en]. Stillwell, John; Thomas, Michael 2016: How far do internal migrants really move? Dem- onstrating a new method for the estimation of intra-zonal distance. In: Regional Stud- ies, Regional Sciences 3,1: 28-47 [https://doi.org/10.1080/21681376.2015.1109473]. Stouffer, Samuel A. 1960: Intervening opportunities and competing migrants. In: Jour- nal of Regional Science 2,1: 1-26 [https://doi.org/10.1111/j.1467-9787.1960.tb00832.x]. Tobler, Waldo 1995: Migration: Ravenstein, Thornthwaite, and Beyond. In: Urban Geog- raphy 16,4: 327-343 [https://doi.org/10.2747/0272-3638.16.4.327]. United Nations Economic Commission for Europe (UNECE), Statistical Offi ce of the Euro- pean Union (EUROSTAT) 2010: Guidelines for exchanging data to improve emigration statistics. Prepared by the task force on measuring emigration using data collected by the receiving country. Geneva: United Nations [http://www.unece.org/fi leadmin/ DAM/stats/publications/Guidelines_improve_emigration_statistics.pdf, 05.08.2020] Date of submission: 28.01.2020 Date of acceptance: 22.06.2020 Dr. Federico Benassi, Francesca Licari, Dr. Enrico Tucci. Italian National Institute of Statistics. Rome, Italy. E-mail: benassi@istat.it, licari@istat.it, tucci@istat.it URL: https://www.researchgate.net/profi le/Federico_Benassi https://www.researchgate.net/profi le/Francesca_Licari https://www.researchgate.net/profi le/Enrico_Tucci Dr. Corrado Bonifazi, Frank Heins (). Institute for Research on Population and Social Policies (IRPPS). Rome, Italy. E-mail: corrado.bonifazi@irpps.cnr.it, frank.heins@irpps.cnr.it URL: https://www.irpps.cnr.it/en/staff/corrado-bonifazi-3/ https://www.irpps.cnr.it/en/staff/frank-heins-en/ • Federico Benassi, Corrado Bonifazi, Frank Heins, Francesca Licari, Enrico Tucci530 Appendix Fig. A1: Map of the case study areas � Source: Based on Istat Population Change and International and Internal Migration in Italy • 531 Tab. A1: Variables used in the analysis and in Tables 1, 3, and 4 Indicator Defi nition Source Net migration rates See section 3 Data and Methods Population registers Economic activity rate 20-64, 2011 (%) Share of the 20-64-year- olds employed and unemployed (%) 2011 Population and Housing Census Economic activity rate 15+, 2006-2008 (%) Share of the over-14- year-olds employed and unemployed (%) Istat estimates of the employed and unemployed in the LLMAs Economic activity rate 15+ Share of over-14-year-olds employed and unemployed in (average values of yearly) Istat estimates of the employed and unemployed in the LLMAs [https://www. istat.it/it/informazioni- territoriali-e-cartografi che/ sistemi-locali-del-lavoro] or [https://www.istat.it/it/ archivio/217437] Population density (per km2) Population per surface area per km2 2011 Population and Housing Census Population 20-44 (%) Share of the 20-44-year- olds in the total population (%) 2011 Population and Housing Census Foreign population (%) Share of the foreign population in the total population (%) 2011 Population and Housing Census 20-64-year-olds employed in the service sector (%) Share of the 20-64-year- olds employed in service sector (services for information and communication, as well as fi nancial and insurance activities) (%) 2011 Population and Housing Census Population 20-44 living as son or daughter (%) Share of the 20-44-year- olds living in a household as child (%) 2011 Population and Housing Census Annual population growth 2002-2017 Average population growth from 01.01.2002-31.12.2017 (%) Population registers Unemployment rate 20-64 (%) Share of the 20-64-year- olds unemployed as a share of those that are employed and unemployed (%) 2011 Population and Housing Census Population 20-64 with an university degree (%) Share of the 20-64-year- olds with an university degree (%) 2011 Population and Housing Census Published by Prof. Dr. Norbert F. Schneider Federal Institute for Population Research D-65180 Wiesbaden / Germany 2019 Managing Editor Prof. Philip Rees Dr. Katrin Schiefer Copy Editor Julia Luther Editorial Assistant Beatriz Feiler-Fuchs Wiebke Hamann Layout Beatriz Feiler-Fuchs E-mail: cpos@bib.bund.de Scientifi c Advisory Board Karsten Hank (Cologne) Michaela Kreyenfeld (Berlin) Marc Luy (Vienna) Natalie Nitsche (Vienna) Peter Preisendörfer (Mainz) Zsolt Spéder (Budapest) Rainer Wehrhahn (Kiel) Comparative Population Studies www.comparativepopulationstudies.de ISSN: 1869-8980 (Print) – 1869-8999 (Internet) Board of Reviewers Martin Abraham (Erlangen) Laura Bernardi (Lausanne) Hansjörg Bucher (Bonn) Claudia Diehl (Konstanz) Andreas Diekmann (Zurich) Gabriele Doblhammer-Reiter (Rostock) Jürgen Dorbritz (Wiesbaden) Anette Eva Fasang (Berlin) E.-Jürgen Flöthmann (Bielefeld) Alexia Fürnkranz-Prskawetz (Vienna) Beat Fux (Salzburg) Joshua Goldstein (Berkeley) Sonja Haug (Regensburg) Hill Kulu (Liverpool) Aart C. Liefbroer (The Hague) Kurt Lüscher (Konstanz) Emma Lundholm (Umeå) Nadja Milewski (Rostock) Dimiter Philipov (Vienna) Roland Rau (Rostock) Tomáš Sobotka (Vienna) Jeroen Spijker (Barcelona) Olivier Thévenon (Paris) Helga de Valk (Brussels) Heike Trappe (Rostock) Michael Wagner (Cologne)