The macroeconomic impact of labour liberalization and policies in MENA countries

Since the end of World War II, migration from the Middle East and North Africa (MENA) has been a widely discussed phenomenon. On the one hand, migration to Europe, driven by the hope of escaping poverty or unemployment, has been steadily increasing since then. Drawing back on the data of 2008, there is a growing tendency to migrate among the youth in several MENA countries. As of 2008, almost 8.2 million people headed towards the European Union, while some were also leaving to work in either other Arab countries (2.7 million) or other destinations abroad (1.7 million). Especially in the MENA countries, where unemployment rates have reached as high as 28% in the past (Nassar, 2005), mostly young, but also a rising number of educated labour market participants find themselves unable to obtain employment in their home country, thus increasing the need to seek jobs abroad.Europe, on the other hand, faces the reverse trend: Its population is growing older and the dependency ratio is steadily increasing. Hence, it seems reasonable that both regions, the EU as labour-importing and MENA as labour-exporting, might benefit from relaxed restrictions on migration (Haas, 2010). Within that scope, the present paper investigates whether this benefit exists and how it can be quantified.

Literature review

Numerous studies have been conducted to analyse the quantitative and qualitative impact migrants have on their host, as well as their home country. Most economists agree that migration has overall positive welfare effects on the labour-importing countries, while evidence for the labour-exporting countries is ambiguous. In general, migration is believed to increase growth in the host countries due to increased output effects. According to some authors, lifting restrictions on unskilled workers in developed countries generate widespread positive effects on production and hence on real GDP, whereas the benefits from skilled labour movements are primarily felt in specific service sectors (Coppel et al. (2001); Boeri and Brücker (2005)).While most recent research has focused on lifting of international trade barriers, overall gains resulting from the elimination of trade barriers are not that impressive, when compared to the elimination of restrictions to labour mobility. In Clemens (2011) various studies were collected so to gauge the efficiency gain in GDP with removal of different barriers which can reach over 140% with labour liberalization compared to only 4.1% with trade liberalization. Taking into account studies with more reasonable assumptions on the scale of migration Clemens (2011) further reviews efficiency gains in GDP with lower rates of emigration out of origin population, the so called the net emigration rate of origin population. These gains range from over 50% to around 1% according to different rates of migration (positively correlated).In order to investigate the effects several approaches have been used by scholars, which can be grouped as simulation based and econometric based. The first group consists of the factor proportions approach (partial equilibrium) and computable general equilibrium (CGE) approach. The second group comprise area analysis, production theory and time-series approaches.There exist a number of different models that are aimed at studying the effects of labour migration on the overall economy. A complex open-economy CGE framework is used in Keuschnigg and Kohler (1999) to assess the EU enlargement on Austrian economy. The authors assume that the number of unskilled and skilled workers is scheduled to increase by 10.5%, and 2.1% respectively. At the same time, it is expected that wages of unskilled employees decline by 5%, whereas those of skilled workers increase by 2.7%. Muller (1997) describes the effects of migration in Switzerland. His exploratory study suggests mild positive results regarding labour market segmentation, capital mobility, and terms of trade in the native economy. A study conducted by Walmsley and Winters (2005) comes to the conclusion that an increase of mere 3% in the quota of temporary movement of skilled and unskilled workers results in a US$156 billion (in 1997 constant prices) gain in global welfare. Their result confirms that restrictions on labour movement in developing countries are costly for all affected countries, especially due to a foregone decrease of costs for unskilled, domestic labour in the host country. Similarly, a dynamic recursive general equilibrium model, used in a World Bank study (2006), concludes that the global welfare gain is an impressive US$ 674 billion in real income in comparison to the baseline scenario (in 2001 prices). Also, it suggests that liberalization in labour migration benefits developed economies through welfare changes, but generally leads to a fall in wages. The model of Boeri and Brucker (2005) shows that positive gains of labour migration are possible when two economies open their borders. When labour markets clear (baseline scenario), a 1% increase in the share of immigrants relative to native population leads to a 0.3% increase in GDP for the EU member states (total region), while the gains/losses are almost 0.7% in the host country and -0.7% in the source country. However, those gains happen at the expense of low-skilled native workers in the host country, who are most hit by comparable foreign workers with similar skills. Moreover, as Weizsäcker (2006) points out, low-skilled migration increases income inequality among the natives and increases wage prospects in the sending country. For high-skilled workers, the reverse holds. Another multiregional CGE model by Iregui (2003) argues similarly and evaluates the gains from an elimination of restrictions on workers’ movements. In this model specification with 1990 data, average wages increase in the source countries, as restrictions for skilled and unskilled workers are lifted. This is driven by labour scarcity relative to capital, making the return to capital lower. The remaining skilled workers in the source country receive wage increases because of raising demand. In the destination country, wages are lower for both types of labour and the return to capital increases. In case of only skilled labour migration, unskilled workers and capital owners are worse off in the home country. In the host economy, a larger number of skilled workers receive average lower wages. An important extension of the model refers to capital mobility along with skilled labour movement: workers move to countries and sectors with higher wages and higher returns to capital, thus in the home region unskilled workers and capital owners are worse off, whereas in the host country the trend is reversed. Additionally, welfare gains are substantially reduced when transaction costs are introduced in the model.The effect of migrants on wages is analysed in more detail by Longhi et al. (2005a). They use the methodology of meta-analysis to review 18 papers with 348 estimates of immigrant inflows on wages of workers in the host country. Their findings suggest that an increase of the immigration population by 1% results in a decrease in wages of the native-born population of 0.1%. Other studies (Altonji and Card (1987); Bean et al. (1988); Borjas (1986), (1987); Grossman (1982); LaLonde and Topel (1987)) confirm that immigrants impact earnings and employment opportunities of natives only to a very small extent, while they do have a significant impact on their own and other immigrants’ wages. A 10% increase in the number of immigrants leads to a reduction in immigrants’ wages of 2% to 3%.Judging from the evidence so far, it can be summarized that there are little or no negative effects of large scale migration on labour market outcomes in terms of wages and employment opportunities. This could be explained by the compensating growth in output and rise in productivity. Priore (1979) also argues that this is either due to the fact that immigrants perform jobs that most natives do not want to engage in, or to a general mismatch on the native labour market.Basically, as shown in Poot and Cochrane (2005), there are three ways in which migration may contribute to greater economic growth. First, immigration may lead to the advance movement on the convergence to the long-run steady-state growth path; second, it may contribute to innovation and changes in total factor productivity; third, it may trigger incremental changes in efficiency, which boosts total factor productivity in the long run.The standard open-economy model describes the steady state convergence mechanism, which assumes acceleration with greater population influx and comes to the steady state growth rate. A simple macro model by Kemnitz (2001) describes the importance of immigration coinciding with capital growth, which suggests that immigration will benefit a native if and only if the average immigrant arrives with more capital than the average native.Employing the framework of CGE, Walmsley and Winters (2003) study the impact of freeing up of border controls on growth. Their estimation shows that a 3% increase in quotas of unskilled and skilled workers will translate into a global economic boon worth 150 billion USD.The second way of increasing growth is through Schumpeterian innovation: as new immigrants tend to bring and exchange new ideas, open new businesses and new industries. They may even attract FDI from their home countries. Theoretical aspect of entrepreneurial activity of immigrants is largely overlooked, but empirical work exists. Ching and Chen (2000) find evidence that immigration may lead to greater international trade between countries, based upon the study of immigrant entrepreneurs of Taiwanese origin in Canada, while their another finding suggests that investor class (passive capitalists) immigration has a detrimental effect of trade. However, this economic aspect of immigration still remains under-researched.The third venue of economic growth is through economic efficiency. Migrants may respond better to economic signals as they are younger on average, and may be better equipped to adjust to changing conditions. Borjas (2001) uses the model of ?greasing the wheels? in order to explain the link between immigration and growth. His paper estimates how the migrants contribute to US growth by ?greasing the wheels?.Several studies of natural experiments describing workers movement from Cuba to Florida, Algeria to France, Angola and Mozambique to Portugal, the former USSR to Israel do not refute the hypothesis that local markets have the capacity to absorb great labour shocks in a relatively short time span.The body of literature that examines the impact of migration on the host country remains much greater than that of dealing with the effects of outward migration on the sending countries. However, the scale of outward migration in the MENA region requires such research to be undertaken in order to estimate short and long-term economic outcomes. Nassar (2005) suggests that outward migration kept unemployment lower in the sending countries among the youth and women.One of the features of the regional migration is that it tends to be temporary or circular. The case in point is Egyptian workers in the Gulf States, whose number depends on oil prices and political situation in the region. Other notable examples are Syrian workers in Lebanon, Egyptian farmers in Jordan and Libya (Fargues, 2009, p 28-31). The framework of circular migration is preferable for the EU countries, some of which have already signed bilateral agreements on temporary workers: Spain and Morocco, France and Tunisia, Italy and Egypt. The impact of this migration has yet been investigated.A study by the scholars (Venturini, Fakhoury and Jouant ,2009) suggests that job creation in Tunisia, Morocco, Algeria and Egypt consistently lags behind so that the labour market creates annually excess supply which cannot be met by demand in the MENA market. Their outcome of the model, based solely on the four countries, shows that close to half a million people become candidates for circular migration each year.For completion purposes, it is also necessary to mention some more qualitative studies analysing some of the effects migration can have, but which are more difficult to quantify. One important aspect is the role of remittances. In the neoclassical context, remittances help to improve the income distribution, decrease inequality and increase overall welfare (Borjas, 1989). In the cumulative causation theory, remittances trigger inflation, thus increasing inequality and allowing for the ?Dutch disease’ effect by leading to an appreciation of the currency and thus weakening the external competitiveness of that country. This is confirmed by some empirical studies (Myrdal, 1957). Also, the importance of ?brain drain’, the emigration of mostly well-educated citizens, or ?brawn drain’, the emigration of young, agricultural labourers, should not be underestimated, since it bears substantial costs to the source country which might not be fully compensated by the inflow of remittances. Nevertheless, evidence for ?brain drain’ can only be found in rather small and very poor economies (Haas, 2010). Other effects might include socio-cultural effects in the home country (consumerism, non-productive, and remittance dependent attitude) or the host country (xenophobia and racism).So far, the almost unanimous conclusion is that the gains in the labour market are positive, albeit negligible at best. There is a growing evidence that some of these gains are at the expense of previous migrants or low-skilled workers in the host community (Walmsley et al., 2007). Moreover, the probability that immigrants increase unemployment in the short run is low and in the long run is zero (Okkerse, 2008). However, overall welfare seems to increase when relaxing migration restrictions.

Descriptive statistics

Due to recent developments in the MENA region, but also with regard to demographic changes in the EU, migration from the MENA countries to the EU has gained some renewed attention. Especially the turmoil in many Arab nations during the Arab spring is seen as a new turning point in migration relations with the EU. So far, the effects of the tremendous, institutional changes that have taken place cannot be clearly interpreted. Another important change within the EU will also have significant effects on the stream of non-EU migrants. As observed by Deutsche Bank Research group (2011), migration from the peripheral EU countries to the core countries increased due to the impact of the recent financial crisis. Nevertheless, current demographic trends might encourage migration and create a more positive perception of migrants. The aging population in Europe will need to find measures in the future to keep their workforce constant, if they are intending to retain their standard of living. Due to the very low fertility rates in the EU, a softening in migration policies could be the answer.We chose to focus more extensively on Algeria, Morocco, Tunisia, Egypt, Turkey, since they account for the largest migration flows from the MENA region to the EU. Especially, Moroccan and Turkish citizens make up large proportions of the foreign population in some EU countries. In Germany, one quarter of all foreigners is Turkish, whereas in Spain 13.1% are Moroccans. The Netherlands have a roughly equal share of Turkish and Moroccans, which sums up to 25% of the total foreign population (EUROSTAT, 2011). Egyptians actually play a minor role in the EU, but they account for a larger proportion of migrants in Italy and Greece (Zohry, 2005).Egypt and Turkey are by far the largest MENA countries in terms of population with more than twice as many inhabitants than Algeria and Morocco. The proportion of the working age population is roughly the same for all the reported countries or regions. However, when looking at the annual growth rates, it becomes apparent that the ratio is at least three times higher in most MENA countries than in the EU. While in 2009 the EU-population grew by just 0.36%, the same figure accounted for roughly 1.87% in the MENA region. In terms of wealth as measured by GDP per capita, the average EU citizen earns 4 to 8 times more than the average MENA citizen. Due to their relative low income in their origin countries, the wage differential based incentive to migrate is highest for Moroccans and Egyptians. Turkey is by far the richest of these countries, but still way below the European average. A last observation can be made concerning the sectorial distribution of labour: Whereas only 5% of the EU labour force is employed in agriculture, the same figure is up to 41% in Morocco and 26% in Turkey.While the labour participation rate is especially high in the EU and Algeria, it is surprisingly low in Turkey, Tunisia, and Egypt. Moreover, the rate of female labour market participation is as low as one quarter in most MENA countries (except Algeria) compared to almost 50% in the EU. In turn, male participation in all MENA countries is extremely high, whereas it only accounts for 65% in the EU. Another interesting observation can be made with regard to the educational status of the labour force. It appears shocking that more than 80% of Moroccans have either no or only primary education. This development seems, albeit less pronounced, to be shared by other MENA countries for which the corresponding data is available.While overall unemployment in MENA is only a few percentage points above the EU average level, unemployment of workers with a university degree is up to four times higher in North Africa than it is in the European Union. Youth unemployment seems to be a similar problem in MENA and in the EU. This problem is especially acute in the EU, Morocco, Algeria, Turkey, and Tunisia, where unemployment among the youth is more than twice as large as the total unemployment rate. Within that scope, it is worth noticing that the reported high GDP growth rates that were observed in some MENA countries in most recent years did not have significant effects on the labour market.Net migration is positive in the EU while negative in the MENA region, which allows classifying the EU as a labour-importing and MENA as a labour-exporting region. Roughly every fifth foreigner in the EU (21.3%) is from one of the five listed MENA countries.As was expected, North Africans tend to migrate mostly towards francophone countries, for example France and Belgium. In consequence of the common language, barriers of entry and transaction costs are significantly lower for citizens from Morocco, Tunisia, and Algeria. Due to colonial ties, already existing social networks ease the transition for migration and reduce transaction costs further. France alone accounts for 63% of the North African migrants with a stay of more than seven years in the EU. Spain is another large recipient of temporary North African migration in relative terms, which is usually explained by its high demand for seasonal, agricultural workers. In contrast, the case of migrants from the Near and Middle East is very different: The majority of the workers migrates to Sweden, Denmark or the UK and is rather evenly distributed among all other countries. It should be noted that some of the countries in the Near and Middle East classification do not officially belong to the MENA region (e.g. countries in Central Asia).It seems clear that, on average, migrants are less likely to find employment in the EU than natives. Especially difficult is the situation for immigrants in Belgium, the Netherlands, Spain, and Sweden, where the unemployment rate among foreigners is up to 3.45 percentage points higher than its national counterpart. In contrast, the labour markets of Hungary, Greece, Ireland, Portugal, United Kingdom, and Italy demonstrate more favourable conditions to foreigners than to their own citizens.Firstly, those MENA migrants who had difficulties in finding employment in countries with high unemployment rates among foreigners are mainly low-skilled. This is especially true for Spain, France, the Netherlands, Finland, and Italy. Secondly, in countries where foreigners exhibit lower unemployment rates than natives (e.g., Hungary, Poland or Luxembourg), hardly any unemployment among the unskilled is observable. The most likely explanation is that these countries do not attract as much unskilled labour (due to low wages or a very specialized labour market), thus giving more weight to more educated migrants. A last observation that can be made is the role of Egyptians. Only in this case, high skilled workers, on average, present higher unemployment rates than unskilled. This is no surprise with respect to the large unemployment rates of young, educated work seekers in Egypt. It does seem surprising in connection with the high demand for labour of the core EU countries. The high unemployment rate in the EU for skilled Egyptian labour might demonstrate another sign of the mismatch in the Egyptian labour market and the inferior quality of Egyptian tertiary education.

Model and Database

In this study, the standard global applied general equilibrium GTAP model, a CGE model with bilateral labour migration, is used (Hertel, 1997). In the standard GTAP framework, conventional neoclassical behaviour (utility maximization, cost minimization) is assumed, with regional utility aggregated over private demands (non-homothetic), public demands, and savings (investment demand). Production is characterized by a perfectly competitive, constant returns-to-scale technology, and bilateral imports are differentiated by region of origin using the Armington specification. The model incorporates five factors of production. Skilled/unskilled labour and capital are perfectly mobile, whereas land and natural resources are both sector specific with the former moving ?sluggishly? between productive sectors. In all factor markets, full employment is assumed (long-run equilibrium). However, the mobility of labour and capital can only occur within regions. GTAP allows divergences between regional investment and saving, but forces all existing capital within a region to move only across industries within that region. Finally, investment behaviour is characterized by a fictitious ?global bank?, which collects investment funds (savings) from each region and allocates them across regions according to a rate of return or a fixed investment share mechanism.The data used for the simulation is taken from the GTAP Data Base 7.1 which is a fully documented, publicly available global data base which contains complete bilateral trade information, transport, and protection linkages among regions for all GTAP commodities.


The first simulation examines the outcome of lifting the restrictions for skilled and unskilled migrants to the EU by 1% and estimates the impact on the main macroeconomic variables: growth, real investment, exports, imports, and welfare (decomposed into allocative efficiency, endowment, population, terms of trade, price of capital goods, remittances effect) in both labour-importing and exporting countries. Furthermore, the impact on wages is examined.The second simulation tests for the separate effect of an increased number of skilled migrants from MENA to the EU. This allows to test the hypothesis whether allowing only skilled workers to migrate by issuing special visas is more profitable compared to a general relaxation of migration restrictions. Thus, the shock consists of a 1% increase in skilled migrants from MENA to the EU.At last, several sensitivity analyses examine how vulnerable the results are to different parameter changes.

Findings and conclusions

Our main findings suggest that there is potential, but substantial income gain in the world GDP (as high as 56 Million USD) if labour movement restrictions were further relaxed in the EU. However, this potential gain is realised largely at the expense of the MENA countries. This finding is in line with the recent literature review which predicts a boost in GDP through greater efficiency in factor allocation.Throughout the research we have employed the CGE framework that enabled us to model labour flows with two different skills decomposition ? high and low. We have also taken into account the remittances and trade flows, land, capital and natural resources. In our first empirical exercise an increase of 1% in the number of MENA labour migrants (both high and low skilled) was examined European workers are expected to experience small wage declines ? both skilled and unskilled, while positive outcome is predicted by the model in the returns to land, capital and natural resources. These findings remain plausible in the context of the existing literature concerning native wages. The same holds reverse for MENA countries, while the wage increase for skilled workers (2.07%) is almost five times greater than that of unskilled workers (0.4%). The EU countries stand to gain in along multiple welfare components due to greater allocative and tax collection efficiency. MENA countries on the other hand are expected to gain from expanding terms of trade and remittances from the EU, while losing slightly in allocative efficiency, endowment, population effect and capital goods. Other macroeconomic variables (consumption, investment, government spending, exports and imports) in the EU 27 countries show around 0.5 % change, while consumption is growing in MENA by 0.24%, government spending increases by 0.3% and investment (0.43%) and trade (imports (0.16%) and exports (0.7%)) stand to decline. Overall, a 1% increase in migration from the MENA countries to the European Union leads to a decline in total GDP within the MENA region, whereas the EU as well as the whole world faces a rise in total output.Our second scenario tested an asymmetric shock of 1% increase only in skilled migration from MENA to the EU 27. The gains in the input factors (land, capital and resources) in the EU countries were less pronounced and the same applies to losses for MENA. However, the wage effect was more dramatic: EU skilled workers would expect lower wages among stronger competition, while unskilled workers would enjoy a wage raise of 0.2%. In MENA the scarce skilled workers would receive more than 2% wage increase, while their unskilled counterparts would suffer a wage decrease of 0.11%. Welfare gains are expected to be lower than in previous estimation due to smaller labour flows (only more skilled workers move) and overall GDP welfare would be around 31 million USD. All other economic measures retain their sign but dwindle in magnitude. GDP decomposition in the EU 27 shows smaller values as well and gains hover around 0.2%-0.3%, while in the MENA countries the effects are varied: greater consumption (0.2%) and government spending, less investment, exports and imports. Nevertheless it must be noted again that the numerical developments within the scope of simulation 2 are smaller than the ones caused by an increase in both skilled and unskilled migration to the EU.Finally, further simulations were tested in order to determine how robust the results are under different parameters.It appears that the EU labour market is able to absorb successfully even greater numbers of migrants without greatly worsening the welfare of residents than today which clearly coincides with the previous findings in the reviewed literature. Notably, the MENA countries are expected to lose in GDP, population, resource utilization, etc., while these losses are being only partially offset by remittances. These detrimental effects were not totally surprising and are present in literature. If it comes true, such scenario poses a serious challenge to the policy makers in the sending countries who are concerned with the state welfare. For example, a new policy might be aimed at encouraging return migration since it is supposed to mitigate the negative effects. Returnees would contribute to greater productivity through higher skills and know-how gained abroad.Labour participation rates of women and young workers are very low compared to other regions in the world. Our study due to the limitations of the available data was not able to investigate the effects of migration on these particular groups. Further research could provide some insight into the matter.The existing data refers to the EU and MENA countries before the on-going debt crisis that started in the end of 2007. While still in development the effects of the crisis are most likely to alter some of the conclusions in the study.