International Migration: What is all the fuss about?
Migration from places of economic disparity and the lack of opportunity is fundamental to our species existence. Dating back to the exodus of Homo sapiens out of eastern Africa, migration has been one of the great constants of human history. Long before political borders emerged, humans were walking the Earth, traveling in the hope of finding new and better lives. Some of these journeys were cyclical, such as the seasonal treks of nomadic tribes, others were more continuous journeys searching for a better home. Humanity’s migration journey was not always voluntary; disruptions could take the form of wars, invasion, or famines. Whatever the reason may be, population movements across geographies created a long-lasting impact on the host and sending countries.
Until the early 20th century, migrants from much of the world traveled with few restrictions, pro-vided they could aﬀord and survive the administrative and health checks. At this time, immigrants rarely needed visas to make a new start in another country. Between 1920 and 1930, the trauma of unprecedented global warfare, economic recession, and growing xenophobia resulted in a sharp decline in the liberty of migration. By the time it re-emerged in the mid-1940s, the beginning of the “thirty glorious years,” its character had changed.
Fast-forwarding today, around 243 million – or 3.3% – of the Earth’s 7.3 billion people live outside their country of birth (2). Although this share in the global population remained roughly constant since 1960, the diversity in terms of composition (i.e., origin) and destination has increased. Today, international migration involves migrants originating from many more countries, going to many more destinations generating more origin-destination pairs than ever observed before (Özden et al., 2011). International migration remains a story of movements of people from less developed to more developed countries. Still, there is a substantial amount of movement between developed countries (i.e., “the north”) and between developing countries (“the south”). Overall, about one-third of the world’s migrants travel from north to north, another third travel from south to north, and the final third travel from south to south.
Note: Figure plots foreign-born stocks (left-axis) and as a share of total poulation (right-axis).
Source: United Nations Migration Database
Just another factor of production in a globalized world?
In the globalized world where capital, goods, and services move freely across borders, international migration remains limited. As more people look to live and work overseas, immigration has become a broad social and economic phenomenon. Although freedom of movement is not necessarily increasing, there is growing recognition of the role of migration as a component of globalization and a source of economic growth.
Despite evidence suggesting substantial eﬃciency and equity gains from freer international migra-tion(Clemens, 2011), lowering barriers to immigration is not a popular idea supported in many dominat-ing political circiles. Going back in history, one will quickly realize that the cyclicality of anti-immigrant attitudes is not limited to our current times. Today, the global news is occupied with US President Donald Trump’s constant thrum of anti-immigrant rhetoric and policy announcements and growing anti-immigrant attitudes in the US, Europe, and in war-bordering countries. As Howard Zinn shows in his popular book People’s History of America, resisting immigration has been a permanent part of American history. As Zinn explains, members of the earlier waves of immigration always made the newcomers’ lives diﬃcult. Older immigrants from England and Germany were hostile toward the Irish, who had come during the Potato Famine between 1845 and 1855, and the Irish were subsequently hostile towards the Chinese immigrants 30 years later. Throughout the Age of Mass Migration (1850-1913) where about million immigrants arrived in the United States, Italians, Jews, Polish, and Mexicans took their turn in being the subject of anti-immigrant rhetoric. Much like some immigrants today, these “new immigrants” were perceived by sectors of the US as being unlikely to assimilate.
Beyond the political debate, migration is a demographically important issue. For instance, between 2000 and 2018 (3), the increase in the foreign-born population accounted for more than three-quarters of the total population increase in European OECD countries, and for almost 40% of the rise in the United States (OECD, 2019). The increase in the demographic importance of international migration over the past decade has given more audience to the economic and political consequences of immigration.
The recent increase in the interest in the migration question boosted the amount of research directed to the topic. As the issue gained public attention, researchers have tried to establish the impact of migration on the host and sending countries. For a long time, the literature focused on the impact of immigration on labor-related outcomes of the native population. On the side of the sending country, most attention has been given to the importance of remittances for the economy and the cost of brain-drain on human capital accumulation. In the last decade, empirical literature expanded, providing a growing amount of evidence on migratory moves’ impact on the economy and society in both their receiving and sending countries. Today, a growing number of papers show that migration boosts bilateral trade and investment, improve public finances, accelerate the diﬀusion of innovation, and generate shifts in values and political preferences. Furthermore, these eﬀects can have heterogenous eﬀects in the short and long-term.
Productivity and Migration
In addition to the benefits mentioned above, migrants also contribute to economic performance through productivity growth. Immigration can aﬀect productivity through several channels: encour-aging eﬃcient specialization, diﬀusion of innovation, enhancing agglomeration externalities by increasing the density in urban areas, and skill-complementarities arising from diversity (see Peri (2016) for a summary).4
Migrants can also boost productivity by diﬀusing knowledge across firms, regions, and countries. For instance, as workers move between firms located within and across countries, they spread knowledge across space(Dasgupta, 2012; Poole, 2013b). As shown in Bahar and Rapoport (2018), migrants can play an important role as vehicles of knowledge transmission between the source and receiving countries, which can lead to increases in sectoral productivity and exports.
Migrants play a unique role in knowledge transmission due to their higher mobility rates compared to natives. Overall, immigrant workers more often move across regions, industries and occupations compared to their native counterparts (Borjas, 2001; Orrenius and Zavodny, 2007). Migrant mobility, thus, can be a mechanism for reviving the diﬀusion machine across regions and country. By working in frontier firms located in the high productive regions, migrants are exposed to knowledge by the leading firms in their sector. At the end of migration episode, migrant returns back to his or her country of origin equipped with productive knowledge. Through this diﬀusion, countries can reduce interregional gaps and generate inclusive and balanced national growth. At the international level, countries lagging in terms of productivity and income can get a chance to improve productivity and catch-up to the leading economies.
Overall, this dissertation explores the consequences of mobility, both within and between countries, on productivity and labor markets. Across three empirical investigations in three diﬀerent countries and period, I attempt to provide evidence on some aspects of migration that is little explored in the literature. Specifically, I explore the links between migrant mobility and labor markets’ adjustment during a demand shock; migration as a channel for diﬀusing knowledge and boosting productivity across borders; and agglomerations and productivity gains associated with larger city sizes in developing economies.
Cushioning Eﬀect of Immigrant Mobility: Evidence from the Great Recession
In the last two decades, the literature focused on the impact of immigration on the labor markets. In the first chapter, I reverse the mirror and explore an aspect little explored in the literature. I attempt to understand what happens to natives’ labor market outcomes when migrants go home. Focusing on the context of Spain during the Great Recession, I study how the departure of immigrant workers impacts the wages and employment of native workers. Specifically, I attempt to show how the departure of foreign-born workers can dampen the eﬀects of negative economic shocks on natives. To guide the empirical exercise and help the interpretation of my empirical findings, I commence by setting out a theoretical framework. This simple model helps derive predictions about the native labor demand while accounting for both negative labor demand shock and a decrease in labor supply due to the departure of immigrant workers. The model’s predictions show that a decrease in overall labor demand due to the economic crisis will decrease the labor demand for natives but also drive some immigrants away from the location. Faced with less competition from the immigrant workforce, labor market outcomes of natives improve. In the context of economic contraction and labor shedding, the decreased competition from immigrants will dampen the harmful eﬀects of the crisis on native employment and wages. In reverse, in case of a positive demand shock, an increase in the labor demand for natives will be slowed by the rise in the competition due to the increased presence of immigrants from the same skill group. I test these predictions empirically using the context of Spain during the Great Recession. I focus on the direct impact of immigrants’ outflow from local labor markets. I start by analyzing the changes in employment and wage outcomes of natives, located in each province, due to the reduction in labor supply due to the departure of immigrant workers. Specifically, I regress the annual growth in employment and wages of natives in a province on normalized net out-migration rates. To make causal claims about the estimates, I use a modified version of the standard shift-share instrument (Card, 2001). I analyze short-term eﬀects on employment and wages for specific groups of workers (e.g., young or unskilled natives) but also types of employment adjustments in response to out-migration.
In my analysis, I use the Spanish Social Security panel, which allows me to track natives over time and across firms, sectors, and regions. I combine this data with individual-level municipal registers that cover 100% of the population to precisely count the number of immigrants residing in each province annually. I measure the intensity of departures across Spanish provinces by the annual decline in the immigrant working-age male population during this period, relative to the entire working-age male population in the previous year.
I find that the outflow of immigrants accelerated the wage and employment growth for natives of all skill and demographic groups. I find that the departure of immigrants increased geographic inflows of natives from other areas, increased the entry to employment from nonemployment, and improved the wages of those who were already employed.
Given the context of economic contraction, these findings show that through their higher mobility, immigrants diﬀuse the incidence of local shocks beyond the local labor market and cushion the natives during an adverse shock. In the context of wage stagnation and employment losses, the departure of immigrants, improved the wage growth while dampening the drop in employment. As the locations which were hit harder by the crisis also lost more immigrants, this smoothing eﬀect worked like an automatic stabilizer stronger in hard-hit areas and worked as a channel for equilibrating diﬀerences across labor markets. This suggests that migrants do not only grease the wheels during the good times (Borjas, 2001) but also during bad times by leaving locations that are hit by a negative shock. This finding underlines an important benefit of immigration that has been explored little in the literature. This is particularly important given the concerns about the relative lack of mobility of natives and especially among less-skilled workers as it leads to significant divergence in local unemployment rates and workers’ earnings across local labor markets (Bound and Holzer, 2002; Cadena and Kovak, 2016; Dao et al., 2017).
Migration and Post-conflict Reconstruction: The Eﬀect of Returning
Refugees on Export Performance in the Former Yugoslavia
The second chapter focuses on another novel angle exploring gains from migration: the role that migrants play in spreading ideas, technology, and knowledge across countries, and how such transfers are reflected in long-term real economic outcomes such as exports. With my co-authors Dany Bahar, Andreas Hauptmann, and Hillel Rapoport, we exploit a natural experiment and document how return migrants –having spent time in a foreign country– explain the subsequent performance of the same export sectors in which they had worked while abroad. Our study focuses on the early 1990s when about 700,000 citizens of the former Yugoslavia fled to Germany to escape the war. Most of the Yugoslavian migrants in the first half of the 1990s were given a temporary legal status, known as Duldung (German for “toleration”), which allowed them to stay and work in Germany. Duldung status was valid for six months and was automatically renewed as long as the war was ongoing. In 1995 the Dayton peace agreements were signed, putting an end to the war in the former Yugoslavia. This led to Germany halting the renewing of the Duldung status of the Yugoslavian refugees, starting an enforced repatriation process. By 2000, the majority of Yugoslavian refugees had been repatriated back to their home country or to other territories of dissolved Yugoslavia.
Within this context, our study uses confidential administrative social security data from Germany that allow us to identify Yugoslavian refugees who had arrived in Germany during the Balkan refugee crisis and returned home after the war following 1995. In particular, we compute the number of these refugees who had worked in each one of the almost 800 four-digit tradable industries in Germany during that period (or ‘treatment’). We link this information to industry-level Yugoslavian export data. We estimate changes in export values from Yugoslavian countries to the rest of the world as a result of returning refugees who were employed in those same sectors in Germany.
We find that industries with a 10% higher number of returning workers have a larger level of exports to the rest of the world from 0.8% to 2.4% between the pre and post-war periods. Furthermore, these positive eﬀects remain stable and keep increasing as time goes by. Our study rules out many explanations that could be driving the results, such as convergence, the inflow of investment linked to migration, or a decrease in information costs linked to international trade due to migrant networks.
Given that productivity is an underlying determinant of exports, we interpret these results as sup-porting the idea of migrants being drivers of know-how and technology transfers between countries. This interpretation is backed by the fact that our results are particularly driven by industries that are knowledge-intensive. They are also stronger when returnees are skilled and in occupations intensive in analytical and cognitive tasks.
A lesson to draw from our results is the importance of integrating refugees in their host economies’ labor markets, as this can really play a significant role in the post-conflict reconstruction of their home countries upon their return.
Agglomeration Eﬀects in a Developing Economy: Evidence from Turkey
Urbanization has shaped developed countries during the 20th century, but it has had transformative eﬀects on developing countries. Several major factors distinguish urbanization in developing countries from urbanization in developed countries. First, urbanization in developing countries is occurring at a quicker pace. In the last 50 years, on the back of a strong wave of rural-urban migration, emerging countries have seen the number of large cities increase. It took more than a century for most developed countries from the time urbanization started to increase markedly until they reached 50%. Today’s developing countries often reach that threshold in less than half the time. Second, much larger numbers are involved making it a relevant issue for large numbers of people.
Given the importance of the topic, the third chapter of this dissertation focuses on understanding the determinants of regional productivity in a developing country context. For such a study, Turkey, an upper-middle-income developing country that has experienced fast urbanization and a high rate of growth of the urban population, seemed like a natural fit. Since the 1950s, the urban population in Turkey has increased dramatically due to massive rural-to-urban migration and a high fertility rate. In 1960, Turkey still featured a largely agrarian economy with 31 percent of its population residing in urban areas. By 2017, 75% of the Turkish population lived in cities, making it a very highly urbanized country. Today there are substantial inequalities between regions on almost every metric (income, production, life quality, etc.), including productivity.
In this study, I focus on understanding the sources of spatial productivity diﬀerences. For this purpose, I use social security records, a new administrative dataset that has only recently become available to researchers and thus has never been used in research before. Combining these records with various other datasets, I apply the standard two-step estimation approach to analyze the relationship between agglomeration economies and productivity in Turkey. Specifically, I attempt to measure the elasticity between employment density and productivity gains associated with denser urban areas. I address the endogeneity bias due to reverse causality by using historical instruments based on census data from the Ottoman Empire and the early years of the Turkish Republic.
My results show that agglomeration economies are powerful and at levels that are expected given Turkey’s urbanization level. The elasticity fits the literature perfectly and contributes to filling the knowledge gap about urbanization dynamics in the developing world. Beyond finding strong gains associated with denser agglomerations and market access, I also find that workers do not sort across locations based on their observable skills. This result is in sharp contrast with what is usually observed for developed countries, where a large fraction of the explanatory power of cities productivity advantages arises from sorting of the workers with higher abilities to denser and larger cities. Overall my findings corroborate earlier findings for developing countries, which show that while the main mechanisms of urban economies are present in the developing world, the current models need to be extended to capture the diﬀerences between Western cities and those in the developing world.
Table of contents :
1 General Introduction
1.1 Cushioning Effect of Immigrant Mobility: Evidence from the Great Recession
1.2 Migration and Post-conflict Reconstruction: The Effect of Returning Refugees on Export Performance in the Former Yugoslavia
1.3 Agglomeration Effects in a Developing Economy: Evidence from Turkey
1.4 Concluding Remarks
2 Introduction Generale
2.1 Effet Amortisseur de la Mobilité des Immigrants
2.2 Migration et Reconstruction Post-conflit : L’Effet du Retour des Réfugiés sur les Exportations en ex-Yougoslavie
2.3 Economies d’Agglomération dans un Economie en Développement : Le Cas de la Turquie
3 The Cushioning Effect of Immigrant Mobility: Evidence from the Great Recession in Spain
3.2 An Equilibrium Model with Heterogeneous Labor Supply, Demand Shock and Wage Rigidities
3.2.2 Equilibrium Effect of Migration and Demand
3.3 Spanish Context
3.3.1 The Great Recession
3.3.2 Immigrant Population in Spain
3.3.3 Immigrant vs. Native Mobility
3.4 Data and Summary Statistics
3.4.2 Summary Statistics
3.5 Empirical Strategy
3.5.1 Econometric Equation
3.6 Main Results
3.6.1 First Stage and Validity of Instruments
3.6.2 Second Stage
3.7.2 Bartik and Other Controls
3.7.3 Alternative Instruments
3.7.4 Alternative Measures of the Supply Shock
3.7.5 Alternative Weights
3.8 Dynamics of Adjustment
3.9 Heterogeneity Analysis
3.9.1 By Skill Group
3.9.2 By Demographics
3.9.3 EU versus non-EU outflows
3.10.1 Geographic Mobility of Natives
3.10.2 Margins of the Employment Effects
3.10.3 Margins of the Wage Effects
3.10.4 Margins by Contract Types
3.11 Understanding the Magnitude of the Estimates
3.12 Concluding Remarks
3.13.1 Derivation of the Firm’s Demand Curve
3.13.2 Derivation of Labor Supply
3.13.3 Derivation of Equilibrium Wage and Employment Responses under Flexible Wages
3.13.4 Firms Adjustment During the Crisis
3.13.5 Differences in Unemployment Rates
3.13.6 Reconciling with Cadena and Kovak (2016)
3.13.7 Immigrant Population
3.13.8 Government Measures
3.13.9 Net Immigrant Departures as a Share of Total Working-age Immigrant Population
3.13.10More on Mobility Data
3.13.11Construction of the instrument
3.13.12 Instrument Validity
3.13.13Construction of Bartik control
3.13.14Immigrant and Native Outflows
3.13.16Differentiating Between Boom and Bust
4 Migration and Post-conflict Reconstruction: The Effect of Eeturning Refugees on Export Performance in the Former Yugoslavia
4.2 Historical Context
4.3 Labor Market Conditions and Mobility of Refugees
4.4 End of the War and Deportation
4.5 Data and Sample
4.6 Natural Experiment: Yugoslavian Refugees
4.6.1 Empirical Strategy
4.6.2 Summary Statistics
4.7.1 Pre-trend and Event-study Estimation
4.7.2 Robustness and Heterogenous Effects
4.8 Mechanisms: Heterogeneous Effects by Workers’ Characteristics
4.9 Concluding Remarks
4.10.1 Anecdotal Evidence: Four Individual Stories
4.10.2 Details on Employment Data, Sample Construction and Variable Description
4.10.3 Zeroes in the Data
4.10.4 Including Exports to Germany
4.10.5 Additional Controls
4.10.6 Further Tests on the Exogeneity of Exit From the Labor Force
4.10.7 Robustness: German WZ 93 3-digit Industries
4.10.8 Other Robustness Checks
4.10.9 Excluding Slovenia and Macedonia from the Sample
4.10.10Occupations by Characteristics
4.10.11Estimations Using Treatments by Educational Attainment and Occupations Characteristics
4.10.12Expanding to All Countries: External Validation
5 Agglomeration Effects in a Developing Economy: Evidence from Turkey
5.2 Literature Review
5.2.1 Agglomeration Economies
5.2.2 Magnitudes for the Effect of Density on Workers’ Productivity
5.2.3 Evidence from Developing Countries
5.3 The Turkish Context
5.4 Literature on Regional Differences
5.5 Data and Sample
5.5.1 Social Security Data
5.5.2 Household Labor Force Survey
5.5.3 Historical Data
5.6 Empirical Strategy
5.6.1 Econometric Equation
5.6.2 Estimation Issues
5.6.3 Estimation Issue 1: Omitted Variables
5.6.4 Estimation Issue 2: Circular Causality
5.6.5 Estimation Issue 3: Accounting for Local Shocks
5.6.6 Estimation Issue 4: Sorting by Ability
5.7 Main Results
5.7.2 Omitted Variable Bias
5.7.3 Reverse Causality
5.7.4 Sorting By Ability
5.8 Multivariate Approach: A Unified Framework
5.9 Infrastructure and Other Amenities
5.10 Concluding Remarks
5.11.1 Summary Statistics
5.11.2 Yearly OLS results
5.11.3 One-step results
5.11.4 Individual Analysis
5.11.5 Informal Employment
5.11.6 First Stage of Multivariate Regression
5.11.7 Additional Results: Infrastructure and Amenities