I: Migration, Remittances, and Economic Transformation in Africa

Migration, Trade, and Development in Africa: Conceptual and Empirical Links

Michaella Vanore

Introduction

As in many other world regions, Africa is facing multiple transitions, three of which are deeply interconnected: those related to mobility, demography, and economic growth. Changing fertility patterns and residential mobility have already begun to shift the age structures of populations across the continent and the places in which those populations mature. These shifts have been accompanied by a growing middle class in many African countries, which, in turn, increase demand for specialized goods and services that require skilled— and semi-skilled professionals to supply them (Mburu-Ndoria, 2016) — highlighting the role of mobile traders in both goods and services who can stimulate and sustain economic growth. Such transitions are also embedded within a context of growing regional integration, where sub-regional communities have developed and encouraged free movement to further stimulate regional integration and, ideally, regional growth.

These transitions provide unique opportunities for — and build upon long-standing histories of — cross-border trade. The story of cross-border trade in Africa is intimately tied to mobility. Intra-continental migration — movement from one country to another within the same continent — has been an important trend within the recent past for much of the African region. While it is beyond the scope of this chapter to meaningfully describe historical and contemporary migration patterns for the entire continent, it is important to recognize the interactions between regional economic systems and regional migration patterns. Regions like East Africa have been shaped by trade systems that extended beyond African borders to include countries in the wider Indian Ocean coastal region. With technological shifts in the 1750s related to manufacturing capabilities and shipping that facilitated greater trade in goods and services, the East African coast became more entrenched in a regional system of exchange. Within this system, a growing number of migrants from Central Africa, Arabia, and Western India arrived in East Africa, many as part of trade enterprises (Gilbert, 2002). Similarly, countries in West Africa have been part of long-standing systems of exchange in which a variety of migrants, including semi-skilled short-term workers and petty traders, routinely cross state borders to provide services and goods. According to Adepoju (2007), West Africa functions as a single socioeconomic zone where social and ethno-cultural ties support continuous cross-border movement and exchange. In Southern Africa, historical integration of countries into what can be described as a ‘regional labour market’ under colonial rule entailed frequent cross-border moves of service providers. While the motivations for migration and the policy frameworks under which migration occurs have changed markedly over time, trade remains an important reason for mobility within the Southern Africa region. Informal and small-scale cross-border trade is particularly important between South Africa and its neighbouring countries, with surveys conducted in the early 2000s indicating a strong presence of informal cross-border traders among those making daily crossings between countries like Zimbabwe and South Africa (Peberdy & Crush, 2007).

Historical and contemporary migration trends are often interpreted from a development lens, and discussions of migration and trade in the African context are often connected to underdevelopment. African migration is often framed as a symptom of development failure, with poverty and desperation driving people away from a ‘home’ that they would otherwise not leave given adequate levels of development (Bakewell, 2008). Governments and civil society voices from countries of origin and destination alike support this discourse. Throughout the last decade, the European Union’s engagement with Africa on migration has reflected what is termed the ‘root causes’ doctrine. The doctrine draws a simple line between migration outcomes — actual mobility — and social, political, and economic contexts in the place of origin that inspire movement. The doctrine emphasizes the logic that migration would not occur if the ‘root causes’ are fixed — that migration can be prevented by addressing systematic underdevelopment (Carling & Talleraas, 2016). Within this doctrine — and other development paradigms that equate mobility with development failure — trade has been framed as a tool to suppress migration by boosting employment and increasing wages in communities of origin. The underlying logic is that trade liberalization, including through policy frameworks such as the Euro-Mediterranean Partnership, will increase exports from migrant countries of origin to ‘more developed’ regions, which, in turn, will support employment and economic growth in origin countries, and incline would-be migrants to remain in countries of origin (Campaniello, 2014).

Despite the ‘root causes’ discourse that typically views migration as a result of underdevelopment, increasing mobility within and beyond Africa likely reflects economic growth, social transformation, and the changes in mobility aspirations and the resources needed to realize them (Flahaux & de Haas, 2016). As this chapter argues, migration is itself an important contributor to economic growth, and its pro-trade potential supports the inclusive development agenda envisioned in the Sustainable Development Goals (SDGs). Trade is an explicit element of the SDGs, which frames increased participation in

Migration, trade, development in Africa 17 international trade regimes as one means to increase economic growth in developing countries and regions. By supporting the development of networks and offering market-specific knowledge that can lower the transaction costs associated with entering new markets, migrants can stimulate trade. They can also encourage businesses to innovate in processes and products by offering scarce skills and services, which can enable businesses to produce products demanded on international markets. For these effects to materialize, however, restrictions on intra-African mobility should be eased.

Migration, trade, and development are deeply connected in the African context, yet their interlinkages have not been fully conceptualized or modelled. This chapter explores the relationships among trade, migration, and development in the African context. It is framed around three broad questions: (1) what is specific about the relationship between migration and trade in the African context?, (2) how are trade and migration linked, and what may this imply for African trade flows?, and (3) what policies can support the relationships among trade, migration, and development in Africa? This chapter is exploratory in nature, as there is relatively limited scholarship and data on the interconnections among migration, trade, and the policy regimes that affect both. The first section of this chapter provides an in-depth assessment of how demography, mobility, and trade are evolving across African countries. The second section then explores the conceptual links between migration and trade, and identifies potential impacts of migration on trade flows on both the global and the African level. The third section briefly reviews policy initiatives that stimulate both migration and trade.

What is specific about migration and trade in the African context?

Migration and trade are embedded within specific contexts, where both regional and local processes and dynamics shape contemporary trends. To properly embed discussions among migration, trade, and development within the African context, this section explores how transitions related to mobility and demography relate to economic growth and the role of trade in stimulating and sustaining growth.

Contemporary migration trends

Over the course of the past 30 years, migration patterns within and beyond Africa have shifted, as too have continental and regional governance frameworks that (explicitly) encourage intraregional mobility as a step to greater regional integration. Compared to 1990 when the total stock of African migrants residing outside of their country of citizenship was approximately 20.3 million, by 2017 that number had reached 36.2 million. While in absolute numbers the difference may seem stark, the share of African emigrants within the total global migration stock has remained fairly consistent over

Figure 2.1 African emigrant stocks, 1990—2017, in absolute numbers and as a share of global stocks.

Source: Own calculations based on data from UNDESA (2017a).

time, representing 13.2% of the share of emigrants residing outside of their country of citizenship in 1990 and 14.1% in 2017 (see Figure 2.1).

Over the past 30 years, migration patterns have more markedly shifted in terms of destination, with growing shares of African emigrants residing outside of the African region. Compared to 1990 when over 85% of the emigrant stock in Africa was a citizen of another African country, the share in 2017 had dropped to 78.5%. In line with this shift, the share of African emigrants residing within Africa dropped, from a high of 65.9% in 1990 to 53.3% in 2017 (own calculations based on UNDESA, 2015). The geography of African migration has thus gradually changed, and the share of African migrants residing particularly in Asia and Europe has generally increased — although the share of Africans residing in Europe has declined in recent years. The regional distribution of the African emigrant stock is visualized in Figure 2.2.

The expansion of migrant populations outside of Africa has been mirrored by changes in intra-African migration dynamics. As other authors (e.g., Flahaux & de Haas, 2016) have noted, much intra-African migration in the post-colonial period has been characterized by rural-to-urban migration and the movement of populations from marginal, inland areas to more fertile agricultural areas and (peri-)urban areas along the coast. Intra-African mobility is thus strongly regional in nature. Since 1990, Eastern and Western Africa have consistently hosted the largest stocks of migrants from other African countries, although the share of the African emigrant stock residing in each sub-region has shifted substantially. In 1990, Eastern Africa was home to over 26% of the total global stock

Migration, trade, development in Africa 19 of Africans, a share that shrunk to 18.5% by 2017. The region has also hosted a dwindling share of intra-African migrants: whereas over 40% of all African emigrants residing in another Africa country lived in the Eastern Africa region in 1990, by 2017 that share had decreased to 34.7% (see Figure 2.3 below).

During the same time period, Western Africa emerged as a more important regional destination, with over 30% of the African emigrant stock residing in that region between 1990 and 2005. While the share of the total African emigrant stock residing in Western Africa dipped between 2010 and 2015, it remained an important regional destination: by 2017 over 31% of all

Figure 2.2 Share of African emigrant stock by world region, 1990—2017.

Source: Own calculations based on data from UNDESA (2017a).

Figure 2.3 Share of total African emigrant stock residing in Africa bv region, 1990-2017.

Source: Own calculations based on data from UNDESA (2017a).

intra-African migrants resided in Western Africa. Compared to Eastern and Western Africa, the other regions of Africa appeared to be relatively moderate players in intra-African mobility trends, yet increasing movements have occurred to the Southern Africa region. Following the end of apartheid and the gradual transition of the South African economy, greater mobility in the sub-region emerged around 2005; while Southern Africa hosted only 6.9% of intra-African migrants in 1990, by 2017 that share was 12.5% (own calculations based on data from UNDESA, 2015) (see Figure 2.4 below).

Observing migration trends on sub-regional level disguises local-level migration dynamics and underplays the importance of migration between neighbouring states. Much intra-African migration is among citizens of countries in the same sub-region and often involves movement across shared borders from landlocked to coastal countries (e.g., Burkina Faso-Côte d’Ivoire). Countries and cities may also act as migration magnets given their specialization in specific activities or industries, such as the oil industries in countries like Gabon and Libya (Flahaux & de Haas, 2016). Governance frameworks may also stimulate intra-regional mobility, particularly within regional economic communities (RECs), where free movement of persons protocols have been developed to further support regional integration. All but one of Africa’s RECs have developed free movement of persons protocols to ease intrarégional mobility, but their success in doing so has been stymied by limited convergence in national laws and in differences across members states in ratification and implementation of the protocol. For example, intrarégional migration within the Economic Community of West African States (ECOWAS) increased following the body’s adoption of the free movement protocol and its supplementary protocols, which the African Development Bank (2018) attributes to the ratification and implementation of the first phase of the protocol by all 15 member states. In contrast, the Southern African Development Community (SADC) experienced limited

Figure 2.4 Share of intra-African emigrant stock by destination region, 1990—2017.

Source: Own calculations based on data from UNDESA (2017a).

Migration, trade, development in Africa 21 changes in mobility following ratification of the free movement protocol due to limited domestic implementation among its member states. Intraregional migration dynamics should therefore be understood within the context of regional integration, but limited data on stocks and flows of migrants within sub-regions challenges further study (AFDB, 2019).

The given statistics are likely to underrepresent the true scale of intraAfrican mobility, as they are collected through population censuses, population registries, and nationally representative surveys that generally require an individual to be present in a foreign territory for three or more months consecutively to be considered a migrant. The statistics would also likely exclude irregular migrants or migrants who are not required to register their residence. Systematic estimates of short-term cross-border moves are unfortunately absent, which makes it challenging to understand the true scale of intra-African mobility, particularly among traders in goods and services who may be temporary movers.

Demographic changes, mobility, and trade

Migration and demography are intimately linked, as migration — in combination with fertility and mortality — is one of the prime phenomena that shapes the demographic profile of a population. Demographic shifts may also precipitate or accelerate migration, particularly given the intersection between demographic profiles and livelihood opportunities.

According to UN estimates, the population of Africa is expected to rise from approximately 1.2 billion people in 2015 to 1.7 billion by 2030, yet the annual rate of population expansion is expected to slow over time. Whereas the number of live births per African woman in the early 2000s was around 5, this number is expected to drop to 3.9 between 2025 and 2030 (UNDESA, 2017b). This seemingly modest change may translate into significant changes to the total dependency ratio. The ratio of the working-age population (aged 15—64) to the population aged 0—14 and 65 and over has shifted rapidly over time. As indicated in Figure 2.5, the dependency ratios of Middle, Eastern, and Western African countries have demonstrated similar trends for much of the observed period, whereas Southern and Northern Africa have strongly divergent dependency ratios starting in the mid-1980s. Dependency ratios across most of Africa’s regions are expected to converge around 2055, at which point the dependency ratio is predicted to be below 65 for all regions.

Across much of Africa, changes in fertility have also been accompanied by changes in population distribution between rural and urban areas. With an annual urban growth rate of 3.5% over the past 20 years, Africa has experienced an acute growth of urban areas, and by 2025, it is estimated that more than 50% of the continent’s population will live in urban areas. Much of the growth of (peri-)urban areas is likely to occur through migration, which may continue to contribute to rapid urbanization in countries like Ghana (Awumbila, 2014).

Figure 2.5 Total dependency ratio per African sub-region, 1955—2100.

Source: Own visualization based on UNDESA (2017c).

The combination of demographic shifts — particularly given the predicted effects on the dependency ratio — in combination with urban growth may create a unique opportunity for Africa to exploit a so-called ‘demographic dividend’. As the population of prime working-age individuals increases relative to the population of people who are less able to contribute to economic systems (e.g., children and the elderly), the labour force, its productivity capacity, and savings tend to rise. Simulations of the potential impacts of decreasing total fertility rates on per capital income and GDP growth suggest significant economic gains with decreasing fertility. For example, estimations by Bloom et al. (2014 as reported in Bloom, Kuhn, & Prettner, 2017) suggest that lowering fertility by meeting one-third of the unmet contraceptive need in Kenya, Nigeria, and Senegal would correspond to increased per capita income of 16%, 9%, and 12%, respectively, by 2030. Similarly, per capita GDP growth would be expected to rise with decreasing fertility. An increase of the ratio of the working-age population to the total population by 1% has been estimated to correspond to a 1.39% increase in the per capita GDP growth rate (Bloom & Canning, 2008, as reported in Bloom, Kuhn, & Prettner, 2017), an impressive elasticity in a region like Africa where total fertility is expected to decline so markedly.

Africa’s demographic transition may generate sizeable economic gains, contributing to an increase in the share of Africans considered to be ‘middle class’ and who can provide predictable sources of income for states to invest in public infrastructure and services like education. If rising incomes are used to support policies that encourage greater human capital gains, a likely outcome is that a higher share of Africans will receive specialized education and

Migration, trade, development in Africa 23 training in medium- and high-skilled functions. While some authors (e.g., Bloom, Kuhn, & Prettner, 2017) have identified education-support policies as one of the important ways African governments can generate a substantial demographic dividend, there is a risk that the functions and services for which the skilled are trained are absent in some local labour markets. Growing education levels and the absence of local labour markets that can absorb those skills are likely to increase emigration pressures in specific countries (Hatton & Williamson, 2011). Structural mismatches in the available supply of skills and economic sectors to employ skilled persons may be offset, at least temporarily, through greater mobility, specifically of service providers, which emphasizes the importance of developing trade policies that can facilitate such cross-border flows.

The transitions many Africa countries face are deeply intertwined, as shifts in population age structure influence the places people choose to settle and the activities they engage in. The growing population of young people in cities with greater disposable income are likely to demand a range of goods and services — some of which are not readily available in local markets. The growing demand for specialized and leisure goods and services is likely to stimulate cross-border trade (Mburu-Ndoria, 2016).

Trade and economic growth

Trade is deeply implicated in Africa’s economic growth, and if the predicted demographic dividend comes to fruition, the support of cross-border trade may become even more key to ensure continued growth. While figures of formal trade are readily available, assessing the complete scale of African trade is challenged by limited data on informal trade flows.

Across many countries of Africa, trade in goods and services is an important contributor to gross domestic product (GDP). According to World Bank (2017) data, trade (the sum of exports and imports in goods and services) constituted a significant share of GDP for many countries, from a low of 21% in Sudan to a high of 174% in the Seychelles. For most countries, trade in services constitutes a much smaller share of GDP, ranging from 2.58% in Sudan to 103% in the Seychelles. While the Seychelles stands out for the role of trade in its GDP, trade in services has an important role in many other countries, including large economies like Ghana, Morocco, and Rwanda, where the contribution to GDP is well above 20%. The net contribution of trade flows to GDP is considerable in many countries, but across the continent, the export of goods and services is much more important to GDP in some countries over others. For example, while the export of goods and services constituted 7.7% of Ethiopia’s GDP in 2017, it constituted a much larger share to the GDP of smaller economies such as the Republic of the Congo (94% of GDP), the Seychelles (81.2% of GDP), and Equatorial Guinea (57.4% of GDP). Among Africa’s largest economies, trade export nevertheless contributed an important share of GDP, such as 13.1% in Nigeria, 29.7% in

South Africa, and 15.8% in Egypt (World Bank, 2017). In contrast to migration flows, which are still strongly intra-African in direction, the greatest share of African exports is destined for countries outside of the continent. As of 2017, only 17% of all African exports were destined for another African country, yet intra-African trade flows involve more complex, added-value goods. As of 2014, around 41.9% of intra-African trade was in manufactured and processed goods, whereas only 14.8% of exports beyond Africa were manufactured goods (Songwe, 2018).

The reported trade volumes are likely to capture only formal trade, and the figures are likely to under-represent the contributions of trade to GDP. In general, informal work is much more prevalent in Africa than in other world regions, with over 85% of all employment (and 71.9% of non-agricultural employment) across the continent considered informal in 2016. The proportion of informal work varies considerably across the continent, however, with 40.2% of work (including agriculture) considered informal in South Africa compared to 92.4% in Western Africa. In all regions except Northern Africa, a greater share of women than men are considered to work informally (ILO, 2018). High rates of informal work in general correspond to high rates of informal cross-border trade (ICBT), which may be of equivalent volume as formal flows for certain products (e.g., food staples, livestock) in specific African countries. Measuring these flows is challenging given their nature: ICBT can involve informal traders who function outside the formal economy, formal firms that do not comply with trade regulations and duties, and formal firms that partially comply with trade regulations and duties but evade part of their obligations. Recent estimations of the volume of ICBT are generally restricted to specific trade corridors or border-crossing points and generally capture only specific dimensions of ICBT (e.g., trade from informal traders functioning outside the formal economy). Nevertheless, they signal the important volume and value of goods exchanged through ICBT. In Uganda, for example, ICBT was estimated to account for approximately 15% of total exports in 2015, and in Rwanda, 59% of exports in 2014 were classified as ICBT (Bouet, Pace, & Glauber, 2018).

Trade — whether occurring formally or informally — may contribute to GDP and to future economic resilience and the attainment of the SDGs. Trade in the context of global development is certainly not uncontested, and the SDGs in particular have received critique from states, civil society, and academia (see, e.g., Spangenberg, 2016) for unbalanced framing of trade as a positive development force despite its potential impacts on environmental sustainability and inequality. Despite this critique, trade has been identified as a force that can support more equitable and inclusive development in the African context, particularly if it stimulates investment in export diversification, vertically integrated industries on regional level, and further value chain development (Songwe, 2018). Trade, specifically ICBT, is also an important contributor to food security: many informal cross-border traders specialize in staple foods (Awumbila, 2014), and in West Africa, ICBT in staple foods was

Migration, trade, development in Africa 25 estimated to account for around one-third of total trade in the region in the mid-aughts (Bouet, Pace, & Glauber, 2018).

Increasing the volume of trade, particularly intra-African trade, has been an explicit strategy within the African Economic Community (AEC) as a way to support sustainable regional growth. While reducing barriers to trade through channels such as customs unions and free trade areas have been key elements of this strategy, promotion of free movement of persons within RECs has also been emphasized as a mechanism that can foster regional integration and economic growth. This emphasizes that migration can indeed strengthen both economic growth and trade, but it is important to more carefully identify the channels through which migration and trade are tied.

How are trade and migration linked?

Understanding more about how migration, demography, and trade are evolving in the African context is important in answering the question of how trade and migration are linked and what the implication of this relationship is for African trade. Both theoretical and empirical literature provide some preliminary answers. The movement of people and goods between any two given places is likely to be mutually reinforcing, and it is likely to be supported by some common factors (e.g., colonial ties) that stimulate exchange between two regions or countries. The channels through which trade affects migration — and vice versa — can therefore be conceptualized as direct or indirect. Trade may cause migration in the sense that people move for the purpose of trading, which may be considered a direct channel of influence. Indirectly, trade may mediate migration choices by affecting the context in which mobility decisions are made, including by influencing production sectors and wages. Similarly, migration may directly cause trade through investment in businesses that allow exchange of goods between countries of origin and destination. Indirectly, there are a range of ways in which migrants can mediate trade relationships: for example, by lowering transactions costs of trade, by promoting the concentration or reach of trade networks, or by having preferences for goods that are novel to the market in a country of destination or origin. This section surveys the theoretical and empirical literature that suggests either direct or indirect channels by which migration and trade may influence each other.

Ties between migration and trade: theoretical assumptions and empirical evidence

The relationship between migration and trade has been theorized in neoclassical economic theory, most notably from Mundell (1957), who proposed that labour and capital may be substitutes and that labour (as a factor of production) may be a substitute for trade. The premise is that countries will have an incentive to trade with one another when there is a difference inthe relative cost of production between countries, which reflect underlying differences in each country’s endowment of resources that shape factors of production (including labour). If goods could be exchanged without restrictions between countries with almost identical profiles (in terms of the goods they produce, the factors of production, technology, and infrastructure), commodity-price equalization and eventually factor-price equalization would occur. Constraints on trade would stimulate factors of production (including labour) to move instead, and restrictions on factor mobility would conversely stimulate trade (Mundell, 1957) — implying that international trade and labour mobility may act as substitutes and that trade — or lack thereof — could stimulate migration.

Given the strict theoretical assumptions under which trade and migration would be expected to act as substitutes, the theory has been contested and has inspired competing models of the conditions under which trade and migration act as substitutes or complements. Del Rio and Thorwarth (2009), for example, suggest that differences in technologies, differences in return to scale for production of differentiated goods, and heterogeneity of labour skills across countries, among other factors, can shift the complementarity/substi-tution effect between trade and migration. Under some circumstances, trade would be expected to stimulate migration — for example, in circumstances when one country produces a homogenous good with constant returns to scale, and another produces differentiated goods with increased returns to scale. Given free trade, a country may specialize in the sector with increasing returns to scale, supporting wage growth that then results in wage differentials between the two countries and motivates labour (as a factor of production) to move (Del Rio & Thorwarth, 2009). Alternatively, Anukoonwattaka and Heal (2014) argue that migration and trade may act as substitutes when countries differ only in factor endowments, which may lead to countries substituting both finished goods and low-skilled, labour-intensive goods from countries with lower factor costs (e.g., wages) in place of importing labour. Countries with higher factor costs may try to suppress the cost of factors of production by outsourcing services, indicating substitution between trade and migration. Trade and migration may instead act as complements when trade and the mobility of factors of production increase simultaneously. Further, while trade and migration may not be fully complementary, migration may nevertheless stimulate trade when migrant workers are complementary to local workers qua skill profiles and occupational choices, which may accelerate growth of export-related industries and therefore promote trade (Anukoonwattaka & Heal, 2014).

The advancement of theory on the relationship between migration and trade has been accompanied by a growth in empirical studies that model the impact of trade on migration flows and the impact of migration on bilateral import and export flows. As noted by Campaniello (2014), empirical studies on causal relationships between trade and migration are relatively scarce; of the small number of studies that have modelled this relationship, most have

Migration, trade, development in Africa 27 been limited to establishing correlation. Within such studies, trade is generally found to positively correlate with migration, at least in the short- and medium-term. A study on the impacts of trade on migration between Mexico and the United States between 1968 and 2004 found strong, positive correlations between trade volume and irregular migration (as proxied by number or border apprehensions). One estimation suggested that an increase of volume of trade between Mexico and the United States of USS1 billion would correspond to an increase in irregular migration of up to 71,000 people per year (Del Rio & Thorwarth, 2009). Campaniello (2014) similarly estimated the causal effect of trade on migration from the so-called ‘Mediterranean third countries’ (non-EU countries included in the Euro-Mediterranean partnership) to EU destination countries between 1970 and 2000. The results of different model specifications consistently predicted an elasticity between trade and migration exceeding 40%, which suggests that additional trade flows between the Mediterranean third countries and the EU would substantially accelerate migration to EU countries.

A larger number of studies have modelled the opposite relationship: that between migration and trade. Several empirical studies have calculated what can be termed the ‘immigrant elasticity of trade’ — essentially the percent change in import/export volumes given a 1% change in the stock of migrants in a given country of residence. A quantitative meta-review of 48 studies investigating the immigrant elasticity of trade in goods conducted by Gen? (2014) concluded that a 10% increase of the migrant stock from a particular origin country increased the volume of exports between the destination and origin country by 2.4% (on average) and the volume of imports between the origin and destination country by 2.9% on average. The average values disguise some important variations: elasticities appeared to decrease as migrant populations matured, tended to be smallest for homogenous goods (goods that are indistinguishable in quality and appearance between manufacturers in countries of origin and residence), and increased among higher-skilled migrants (Gen?, 2014).

Few of the studies reviewed by Gent: (2014) specifically modelled the emigrant elasticity of trade or trade elasticities for migration corridors beyond the high-income OECD region. A small number of past studies have estimated emigrant elasticities of trade and for non-OECD country sets, however. For example, Parsons (2012) estimated the influence of both immigrant and emigrant stocks on bilateral trade using panel data for the period between 1960 and 2000. The study found that emigration and immigration were both positively correlated with bilateral trade flows, and the emigrant elasticity of trade was generally higher than the immigrant elasticity. Any positive effect of either immigration or emigration on trade flows generally disappeared in panel regressions, however, suggesting that time-invariant factors related to specific trade partners that may not be captured in cross-sectional data drive the positive impacts of migration on trade flows. The association between migration and bilateral trade flows was found to vary significantly acrossregional partners, however, with both immigrants and emigrants found to exert statistically significant influence on aggregate bilateral trade flows of exports moving from countries in the global north to those in the global south (Parsons, 2012). Other scholarship focussed specifically on the role of emigrants in shaping African trade flows suggest other patterns. A 2012 study by Ehrhart et al. explored the immigrant elasticity of trade among 52 African origin countries and their 195 global commercial partners. Their estimations suggest that a 1% increase in the global stock of African migrants corresponded to an increase of exports from their countries of origin by 0.178% on average. The authors estimated that in the year 2010, such an elasticity would imply that every African emigrant would generate approximately USS2100 in additional exports per year for the origin country (Ehrhart et al., 2014).

Despite relatively consistent findings between studies on the positive association between migration and trade flows, several important questions remain open about the relationship between migration and trade in the African context given the focus and framing of previous studies. First, in modelling the influence of trade on migration, few studies have been able to establish causal links, and even among those that have, the specific channels that support the impact of trade and migration have not been rigorously examined — in part given the specific information required on the ties between specific pairs of trade partners. Second, in terms of modelling the influence of migration on trade, few empirical studies have explicitly modelled the immigrant or emigrant elasticity of trade between pairs of non-high-income OECD countries, and the role of intra-African migration on import and export volumes between African trading partners have not been extensively explored. Most studies have also focussed exclusively on the migrant elasticity of trade in goods only, and the impact of migration on service trade remains unclear. The studies also highlight that factors such as maturity of the immigrant stock, skill structure of the migrant population, and nature of trade goods mediate the effect of migration on trade. These factors suggest that migration can affect trade through different channels, the function of which may differ in the African context.

Channels of in fluence between migration and trade

Both theoretical and empirical literature have identified different channels through which migration and bilateral trade flows may influence each other. The channels through which trade may influence migration include: (1) increasing incomes, which can be used to fund migration, and (2) widening differences in labour conditions between countries of destination and origin.

One mechanism by which trade could stimulate migration relates to relaxation of capital constraints that would frustrate migration. Migration can be a costly process, particularly in the absence of transparent or regulated migration channels. A growing body of migration and development literature has addressed how economic growth and trade liberalization may spur rather

Migration, trade, development in Africa 29 than stymie migration because of their impacts on the human and financial capital resources that inform migration aspirations and enable migration journeys (see, e.g., de Haas, 2007). In a study among a sample of countries in the Euro-Mediterranean region, Faini and Venturini (2010) modelled the impact of increasing incomes on migration propensity, finding that migration propensities spiked with increasing incomes — at least until incomes approached convergence between countries of origin and destination.

The issue of relative differences between conditions in countries of origin and destination is central in neo-classical economic theories of migration (such as the Harris-Todaro model related to expected wage differentials), and indeed differences in labour market conditions between countries of origin and destination may affect migration choices — suggesting another channel through which trade can shape migration. Trade may encourage countries to specialize in specific sectors where factors of production, such as labour, are relatively more productive, which can support expansion of industry; higher demand for labour; and, in some cases, more competitive wages. Such conditions may attract migrant workers — a familiar scenario in the United States, where trade-related sector expansion corresponded to increased volumes of Mexican workers (Del Rio & Thorwarth, 2009). A third channel by which trade could stimulate migration is through worsening labour conditions in origin countries, which could increase pressures to migrate. For example, differences in production technologies between trading partners may increase the relative efficiency of production in one country at the expense of another. Based again on the relationship between the United States and Mexico, Del Rio and Thorwarth (2009) suggested that Mexico, with its labour-intensive agricultural production, could not compete with the United States, which is characterized by capital-intensive production. As a consequence, trade between the two countries compounded losses to the Mexican agricultural industry, which resulted in reduction of agricultural employment in Mexico and increased pressures to migrate for work to the United States. In how far these mechanisms apply to countries with more similar economies is uncertain, as these channels of influence between trade and migration have been observed primarily in contexts of global south to global north migration.

The channels through which migration may affect trade have received greater examination in the literature and include: (1) lowering of transaction costs, (2) facilitating nostalgia trade (the so-called immigrant-preference effect), (3) direct capital investment in businesses in the country of origin or ancestry, and (4) the development of migrant- or diaspora-owned enterprises. The extent to which these different channels support the growth of trade remains ambiguous, particularly in contexts where migrants move between countries in the same region.

Migrants and their descendants may support trade between their countries of destination and origin by leveraging their knowledge of market conditions, language skills, cultural norms, and business contacts/networks that facilitate the exchange of privileged information and business resources, which inturn lower transaction costs (Gould, 1994; White, 2007; Lewer & van den Berg, 2009). Gould (1994) proposed in his seminal contribution to the literature that ‘human-capital-type externalities’ (p. 302) generated by migrants could improve bilateral trade opportunities. He proposed that migrants are uniquely positioned to provide information on foreign markets to aspiring exporters and may act as network builders between contacts in countries of origin and residence, which helps businesses and investors overcome information asymmetries that increase the risks associated with new trade routes. These proposed mechanisms were tested using bilateral import/export flow data collected between 1970 and 1989 for the United States and 47 of its trade partners, and the analysis revealed that migrant information had a significant, positive influence on trade flows between the United States and migrant origin countries (Gould, 1994).

Specifically within the African context, networks may play a particularly pivotal role in supporting trade given the absence of strong, formal institutions that support and protect businesses engaging in trade. Networks, particularly those convened among co-ethnics or co-nationals, are often based on trust and a commitment of the community to enforce (implicit) agreements within that community. In settings where contracts are difficult to enforce given the absence of legal bodies with the mandate to do or so or where corruption can lead to opportunistic, selective enforcement, such informal networks may stand in for formal institutions by promoting adherence and cooperation (Rauch, 2001). The role of networks in facilitating African trade flows in weak institutional environments was modelled econometrically by Ehrhart et al. (2012). While the African emigrant stock in a particular destination country was generally found to increase trade flows between the destination and origin country, the effect was especially pronounced for countries with low-quality institutions (Ehrhart et al., 2012) — suggesting that migrants can have a pro-trade effect by substituting for formal, protective institutions.

Migrants may also stimulate trade by generating demand for new products and services, called the ‘immigrant preference effect’. This effect occurs when migrants retain preferences for goods/services that are more readily available in the origin country, eventually stimulating the importation of preferred goods from countries of origin to the residence country (Gould, 1994). If such an effect occurs, flows of imports from the origin to destination country would be expected to increase, but only when such goods are strongly differentiated between the two places. There is empirical evidence that such preference effects exist, particularly linked to migrants from lower-income countries. White (2007) modelled bilateral trade flows between the United States and 73 trading partners between 1980 and 2001, revealing that the influence of migrant stocks on bilateral import/export flows was most marked for specific countries. The strongest annual increases in trade flows were between the United States and China, Bangladesh, and Nigeria. The analysis revealed that migrants from lower-income countries supported greater growth in imports than exports, generating an average annual value

Migration, trade, development in Africa 31 of USS910 in exports to and up to US$2,967 in imports from the origin country.

While less directly related to trade as such, migrants may also strengthen trade flows by directly investing in businesses or by starting businesses that may eventually engage in trade activities. The investments made by migrants and their descendants — who are often referred to as diaspora regardless of their self-identification or felt ties — have been increasingly addressed in literature, with the term diaspora direct investments (DDIs) used to identify investments made from expatriate populations in enterprises (foreign direct investment) and in capital markets (e.g., stock and bond sales) in countries of origin or ancestry (Newland & Tanaka, 2010). Investments made from abroad may support businesses to enter trade systems, particularly if those investments enable scaling up and professionalization of products, processes, and services in such a way that would enable a business to meet complex regulatory requirements tied to trade. Tracking DDI and its influence on business growth is difficult, as many countries do not discern the national origin or ancestral ties of investors in businesses. There is some limited evidence that businesses that receive DDIs, particularly small and medium-size enterprises traditionally shut out of trade systems because of regulatory complexity and cost, are better able to engage in trade given their ties to the diaspora. A recent assessment of firms started by migrants and diaspora members (or supported by DDI from abroad) in 19 sub-Saharan African countries found that diaspora-linked firms were more likely to be involved in export than domestic firms without diaspora ties. Such businesses also signalled a higher export intensity, the share of exports in total sales, and were connected to a greater number and diversity of export markets. The study also found that diaspora-linked firms had a higher probability of exporting to countries in the global south — namely other African countries, South Asia, India, and China — but no higher probabilities of exporting to OECD countries than non-diaspora-linked firms. The finding may signal that diaspora investors or investors with a previous migration experience may be better able to leverage their knowledge of destination countries in the global south (Boly et al., 2014).

Taken together, the different strands of literature signal that migrants can have a pro-trade effect by stimulating businesses to offer goods to new markets, supporting networks that can assist businesses in entering international markets, and providing the capital businesses need to compete in new trade spaces. How well migration supports trade is likely to be conditioned by policy, however.

How can policy foster migration, trade, and development in the African context?

While the relationship is by no means linear, literature suggests compelling links among demography, migration, trade, and eventually economic growth and development. If indeed the demographic transition in Africa willsupport the growth of a young, economically active population, a potential demographic dividend may emerge that also enables greater expenditure in education; growing wages; and, in turn, growing supply of and demand for specialized goods and services. Higher incomes and education levels combined with potential mismatches in the supply and demand of (semi-)skilled youth in local labour markets may incentivize greater migration. Greater migration may increase trade flows, which, in turn, may support greater food security, stimulate product and process innovation, encourage the development of regional and global value chains, and eventually stimulate more inclusive economic growth.

In how far migration, demography, and trade converge in the ways sketched to support economic growth in Africa is likely to depend on the development of supportive policy regimes. The role of policy in supporting migration-related trade merits an individual chapter; this section therefore only shortly addresses some of the ways in which policies related to migration and trade can foster their mutual advantages.

Enhancing the pro-trade potential of migration

As addressed in ‘How are trade and migration linked?’, the pro-trade effect of migration is likely to reflect migrants’ capacities to leverage knowledge and networks to lower transaction costs, their demand for nostalgia goods and services, and their investments in enterprises that engage in international trade. Policies that aim to enhance the pro-trade potential of migration could address these mechanisms specifically or could facilitate mobility more generally by lowering barriers to migration — particularly among traders involved in ICBT and service provision.

Facilitating intra-African migration has been prioritized within the African policy agenda. The African Union’s Protocol to the Treaty Establishing the African Economic Community Relating to Free Movement of Persons, Rights of Residence and Right of Establishment (Free Movement of Persons Protocol), adopted in January 2018, reflects the centrality of free movement and labour migration in continental policy strategies. Enabling continental free movement has been a goal since 1991, and while the protocol’s adoption represents an important step forward in eliminating barriers to intra-African migration, its impacts remain uncertain (Dick & Schraven, 2019). As of July 2018, 32 of 55 AU member states had signed the Free Movement of Persons Protocol, but only one (Rwanda) had domestically ratified it. Signatories are clustered regionally: whereas 8 of 9 Middle African states, 11 of 17 in Western Africa, and 11 of 20 in East Africa have signed the protocol, only one country each in Northern and Southern Africa has signed (AU, 2018).

The Free Movement of Persons Protocol could have a significant impact on trade, as cross-border traders may face unclear visa requirements and stringent entry rules that make it challenging to access new markets. Firms could also benefit from more relaxed movement and stay regimes by making it

Migration, trade, development in Africa 33 easier for skilled and semi-skilled labourers to enter and be hired by businesses, particularly those in need of skill sets that are scarce locally. As noted above, however, there remain substantial challenges in harmonizing domestic regulation and legislation related to issues such as admission of workers, which diffuses potential positive impacts of both continent-wide and REC-specific free movement protocols on economic integration and trade.

Following the intuition that migration can support trade, the African Continental Free Trade Area Agreement (AfCFTA) was adopted in March 2018. The aim of the AfCFTA is to create a single African market — ultimately the largest free trade area in the world — through a 90% reduction of trade tariffs, elimination of non-tariff barriers (e.g., subsidies, import tariffs, technical barriers), and liberalization of service trade (Songwe, 2018). Liberalization of trade in services is an especially novel element of AfCFTA, which will likely create greater modalities for cross-border labour flows and channels for movement of services related to finance, distribution, and transport, which would be necessary to support greater regional integration and intra-African trade in goods (Mburu-Ndoria, 2016). The AfCFTA has inspired greater commitment from AU member states, with 44 of 55 countries signatory and 22 countries ratifying it, which is the minimum number needed to bring the agreement into effect. In addition to AU-level initiatives, regional cooperation bodies like the Common Market for Eastern and Southern Africa (COMESA), East Africa Community (EAC), ECOWAS, the West African Economic and Monetary Union (WAEMU), Economic Community of Central African Countries (ECCAS), and SADC have agreed to support services liberalization as part of regional integration efforts (Mburu-Ndoria, 2016), and the near-universal acceptance of free movement of persons protocols within RECs may also support trade in services. Such efforts reflect proactive commitments on sub-region level to increase mobility and circulation, which in some cases — such as ECOWAS, which concluded its own free movement protocol in 1979 — have been explicit strategies for economic integration (Dick & Schraven, 2019).

The extent to which regional integration efforts — particularly in the form of free movement protocols and the reduction of barriers to trade — have succeeded in supporting greater mobility and exchange among participating members is ambiguous. Given the recent acceptance of AfCFTA, there is limited evidence of its efficacy — particularly in terms of impacts on mobility of traders and service providers. Some research has suggested that other efforts to promote trade more generally (and by extension mobility of traders) through RECs have been underwhelming. For example, within the AEC, under which most RECs constitute 'pillars’, strategic aims include supporting economic growth through greater regional economic cooperation, which entailed creation of free trade areas and customs unions. The creation of RECs within the AEC has appeared to have modest impacts on intra-African trade, however. One study of intra-African trade intensity, which compared intensity two years prior and five years after the implementation of RECs,

suggested that most regions experienced minimal intra-regional trade gains through the creation of an REC. In the EAC region, the intensity of intra-regional trade had increased substantially five years after conclusion of the REC agreement, but the intra-regional intensity actually decreased within COMESA, suggesting mixed impacts of lower trade barriers through RECs on intra-regional trade intensity (de Melo & Tskikata, 2015, as cited in ADB, 2018).

Policies that support free movement and service trade liberalization — if effectively implemented — should accelerate intra-African trade and stimulate the production of added-value goods, but other types of policies can support the pro-trade potential of migration. Policies that encourage or facilitate mi-grants/diaspora members to invest in businesses that engage in trade by improving investment environments are one example. Countries such as Benin, Sierra Leone, and Uganda have developed diaspora institutions specifically to channel migrant and diaspora investments (Gamlen, 2014). Others, such as Nigeria and Ethiopia, have developed diaspora bonds to attract capital from their expatriate communities, which can be used to finance infrastructure improvements (e.g., transportation, energy networks, telecommunication) that, in turn, support manufacturing and enable trade. Other policies may indirectly support migrant investments by improving the private sector ecosystem. A recent assessment among Sierra Leonean emigrants in the United States, the UK, and Canada found that significant shares of the sampled population wanted to invest in enterprises and projects in Sierra Leone but did not do so due to limited infrastructure, difficulties in identifying supply chain partners, and difficulties in needed human capital resources in Sierra Leone capable of performing necessary functions for business growth (World Bank, 2015). Policies that support infrastructure development, stimulate the growth of business networks, and encourage skill mobility could play an instrumental role in diffusing such investment barriers.

Conclusions and ways forward

Migration, trade, and development are inherently tied, and, as this review suggests, better supporting their integration is necessary for Africa to meet the inclusive growth envisioned by the SDGs. Migration and trade are implied in a number of SDGs, most notably goal 8 (decent work and economic growth), goal 9 (industry, innovation, and infrastructure), and goal 17 (partnerships for the goals). Eased intra-African movement can play an important role in supporting businesses to acquire the knowledge and skills they need to innovate in products in processes, engage in technological upgrading, and boost productivity through diversification (target 8.2) by allowing greater mobility of workers and service providers. Easing cross-border movements and encouraging circularity may also stimulate the development of business and scientific networks among migrants and non-migrants, which can support businesses to acquire the knowledge and resources needed to innovate,

Migration, trade, development in Africa 35 professionalize, and enter international trade systems (ideas linked to targets 9.3, 9.4, and 9.a). Migrant resources — including social, human, and financial capital — can support development within the private sector through both investment and philanthropy (target 17.3), which can also increase export and import flows within and beyond Africa (target 17.11). For these benefits to materialize, however, trade and migration policies must be designed to be mutually complementary and forward-looking. From this perspective, demographic shifts must be taken into account, as the availability of the human capital needed to sustain business transitions will shift along with the structure of the population.

While this chapter suggests that there are compelling ties among migration, demography, trade, and inclusive development, there remain important gaps in knowledge regarding how these intersections function within Africa. In order to develop more sensitive policies and programmes that enhance these links, future research should provide more nuanced analysis of how intra-African migration affects intra-African trade, particularly trade in services. Research could also further disentangle the channels through which African migrants stimulate trade in goods and services between sets of African countries, and it could also explore whether the sophistication of exports, indicating innovation, increases with migration. The adoption of recent policy initiatives (e.g., the AU Free Movement in Persons Protocol, AfCFTA) that can intensify the intra-African mobility of traders in goods and services also invites future research that explores the impacts of these policies on intra-African trade flows and the implications of those flows for equitable economic growth.

References

Adepoju, A. (2007). “Creating a borderless West Africa: Constraints and prospects for intra-regional migration" in A. Pecaud and P. de Guchteneire (eds.). Migration without Borders: Essays on the Free Movement of People (pp. 161—174). New York: UNESCO & Berghahn Books.

African Development Bank group (AfDB). (2019). African economic outlook.

African Union. (2018). List of countries which have signed, ratified/acceded to the Protocol to the Treaty Establishing the African Economic Community Relating to Free Movement of Persons, Rights of Residence and Right of Establishment, https:// au.int/sites/default/files/treaties/36403-sl-PROTOCOL%20TO%20THE%20 TREATY%20ESTABLISHING%20THE%20AFRICAN%20ECONOMIC%20 COMMUNITY%20RELAT....pdf

Anukoonwattaka, W. & Heal, A. (2014). “Regional integration and labour mobility: Linking trade, migration and development.” Studies in Trade and Investment No. 81, United Nations Economic and Social Commissions for Asia and the Pacific (ESCAP): New York.

Awumbila, M. (2014). “Linkages between urbanization, rural-urban migration and poverty outcomes in Africa.” Background paper for the li'or/d Migration Report 2015. International Organisation for Migration: Geneva.

Bakewell, O. (2008). “‘Keeping them in their place’: The ambivalent relationship between development and migration in Africa.” Third World Quarterly, 29(7): 1341-1358. DOI: 10.1080/01436590802386492.

Bloom, D.E., Kuhn, M., & Prettner, K. (2017). “Africa’s prospects for enjoying a demographic dividend." Journal of Demographic Economics, 83: 63—76.

Boly, A., Coniglio, N.D., Prota, F., & Serie, A. (2014). “Diaspora investments and firm export performance in selected sub-Saharan African countries.” World Development. 59: 422-433.

Bouet, A., Pace, K., & Glauber, J. (2018). “Informal cross-border trade in Africa: How much? Why? And what impact?” International Food Policy Research Institute Discussion Paper No. 01783, Washington DC.

Campaniello, N. (2014). “The causal effect of trade on migration: Evidence from countries of the Euro-Mediterranean partnership." Labour Economics, 30: 223—233.

Carling, J. & Talleraas, C. (2016). “Root Causes and Drivers of Migration.” Oslo: Peace Research Institute Oslo (PRIO) 2016: 1-44.

Paper.

Del Rio, A.M. & Thorwarthe, S. (2009). “Tomatoes or tomato pickers? Free trade and migration between Mexico and the United States.” Journal of Applied Econometrics, 12(1): 109-135.

Dick, E. & Schraven, B. (2019). “Towards a borderless Africa? Regional organisations and free movement of persons in west and north-east Africa.” German Development Institute briefing paper 1/2019. DOI: 10.23661/bpl.2019.

Ehrhart, H., Le Goff, M., Rocher, E., & Singh. RJ. (2014). “Does migration foster exports? Evidence from Africa.” World Bank Policy Research Working Papers. DOI: 10.1596/1813-9450-6739.

Faini, R. & Venturini. A. (2010). “Development and migration: lessons from Southern Europe.” Migration and Culture. Emerald Group Publishing Limited. DOI: 10.1108/S1574-8715(2010)0000008011

Flahaux, M.-L. & de Haas, H. (2016). “African migration: Trends, patterns, drivers.” Comparative Migration Studies, 4(1): 1—25. DOI: 10.1186/s40878-015-0015-6.

Gamlen, A. (2014). “Diaspora institutions and diaspora governance.” International Migration Review, 48(1): 180-217. DOI: 10.1111/imre.l2136.

Gene, M. (2014). “The impact of migration on trade.” IZA World of Labour: 82. DOI: 10.15185/izawol.82.

Gilbert, E. (2002). “Coastal east Africa and the western Indian Ocean: Longdistance trade, empire, migration, and regional unity, 1750—1970.” The History Teacher, 36(1): 7-34.

Gould, D.M. (1994). “Immigrant links to the home country: Empirical implications for U.S. bilateral trade flows.” The Review of Economics and Statistics, 76(2): 302—316.

de Haas, H. (2007). “Turning the tide? Why development will not stop migration." Development and Change, 38(5): 819-841.

International Labour Organisation (ILO). (2018). Women and Men in the Informal Economy: A Statistical Picture, 3rd edition. ILO: Geneva.

Lewer, J.J. & van den Verg, H. (2009). “Does immigration stimulate international trade? Measuring the channels of influence.” The International Trade Journal, 23(2): 187-230.

Mburu-Ndoria, E. (2016). “Africa continental free trade area: Liberalizing trade in services for trade facilitation.” UNCTAD report on project “Strengthening Capacities of African Countries in Boosting Intra-African Trade." Retrieved from https:// unctad.org/meetings/en/SessionalDocuments/ditc-ted-Nairobi-24082015-mburu.pdf

Mundell, R.A. (1957). “International trade and factor mobility.” The American Economic Review, 47(3): 321-335.

Newland, K. & Tanaka, H. (2010). Mobilizing diaspora entrepreneurship for development. Migration Policy Institute: Washington, DC.

Parsons, C.R. (2012). “Do migrants really foster trade? The trade-migration nexus, a panel approach 1960-2000”. World Bank Policy Research Working Paper No. 6034. The World Bank: Washington, DC.

Peberdy, S. & Crush, J. (2007). “Histories, realities and negotiating free movement in southern Africa” in A. Pecaud and P. de Guchteneire (eds.). Migration without Borders: Essays on the Free Movement of People (pp. 175-198). New York: UNESCO & Berghahn Books.

Rauch, J.E. (2001). “Business and social networks in international trade." Journal of Economic Literature, 39: 1177—1203.

Songwe, V. (2018). “Intra-African trade: A path to economic diversification and inclusion." Boosting Trade and Investment: A New Agenda for Regional and International Engagement. Brookings Institution report. Retrived from https://www.brookings. edu/wp-content/uploads/2019/01/BLS18234_BRO_book_006.l_CH6.pdf

Spangenberg, J.H. (2016). “Hot air or comprehensive progress? A critical assessment of the SDGs." Sustainable Development, 25(4): 311-321.

United Nations, Department of Economic and Social Affairs (UNDESA). (2017a). Trends in International Migrant Stock: Migrants by Destination and Origin. United Nations database, POP/DB/MIG/Stock/Rev.2017.

United Nations, Department of Economic and Social Affairs (UNDESA). (2017b). World Population Prospects: The 2017 Revision. Volume II: Demographic Profiles. UNDESA Population Division, document number ST/ESA/SER.A/400.

United Nations, Department of Economic and Social Affairs (UNDESA). (2017c). World Population Prospects: The 2017 Revision. Custom data acquired via website.

White, R. (2007). “Immigrant-trade links, transplanted home bias and network effects.” Applied Economics, 39(7): 839-852. DOI: 10.1080/00036840500447849.

World Bank. (2015). “Sierra Leone diaspora investment and trade study: Executive summary.” World Bank document number 99768. World Bank: Washington, DC.

World Bank. (2017). Databank macroeconomic indicators, trade as % of GDP. Retrieved from https://data.worldbank.org/indicator/ne.trd.gnfs.zs (accessed March 2019).