Do High‐Skill Immigrants trigger High‐Quality Trade?

• Published date: 01 July 2017
• Date: 01 July 2017
• DOI: http://doi.org/10.1111/twec.12431
• Author: Mauro Lanati,Giorgia Giovannetti

Do High-Skill Immigrants trigger

High-Quality Trade?

Giorgia Giovannetti

1,2

and Mauro Lanati

2

1

Department of Economics and Management, University of Firenze, Firenze, Italy and

2

Robert Schuman

Centre for Advanced Studies, European University Institute, Fiesole, Italy

1. INTRODUCTION

IN the last decades, the world economy witnessed an increase in international integration.

Exports rose from 21.6 per cent of world GDP in 1995 to 29.9 per cent in 2013 World

Bank, and foreign direct investment (inward stock) from 9.7 per cent to 34.3 per cent between

1995 and 2013.

1

Over the same period, international migrants increased substantially also: the

total number of migrants rose by 50.1 per cent between 1990 and 2013 and nowadays around

231 million people (approximately 3 per cent of the world’s population) are living in coun-

tries other than their countries of birth.

2

These trends are likely to be closely connected. Indeed, this rise in the stock of immigrants

has been often related to trade flows. Empirically, the possible bi-univocal relationship trig-

gered contrasting results and a lack of consensus on the direction of causation; while there

seems to be some agreement on the strong and significant correlation of the stock of immi-

grants in the receiving country and the amount of trade with their country of origin, particu-

larly evident for high-skilled migrants (see for instance Herander and Saavedra 2005;

Felbermayr and Jung 2009; Felbermayr and Toubal 2012). More specifically, international

migrants could enhance bilateral trade by lowering information costs and increasing demand

for goods from their source countries. There is a fairly established literature on the trade cost

channel of migration, that is the role the ethnic networks play in triggering trade by reducing

information costs: the empirical analysis was first confined to case studies level but, more

recently, the gravity approach has been extended to multi-country analysis.

3

The existing liter-

ature assumes that both imports and exports are symmetrically affected by improved informa-

tion while only imports from source country depend on migrants’ preferences. The preference

channel of migration hinges on the difference in tastes between immigrants and natives: it

reflects the tendency of immigrants to prefer goods they were consuming at home, since it

may take time for them to adjust their tastes to the destination country’s culture and usage.

Against this background, high-skilled migrants tend to impact more on trade because of lower

We are especially grateful to an anonymous referee for insightful suggestions. We also thank Alasdair

Smith (University of Sussex), Juan Dolado (EUI), Ramon Marimon (EUI), Luca De Benedictis (Univer-

sity of Macerata) and Giuseppe De Arcangelis (University of Rome –La Sapienza) for helpful com-

ments and feedback.

1

Data are from World Bank’s Development Indicators. Data on FDI are from UNCTAD (2014).

2

These figures on migrants are from UN (2013).

3

Information costs include the cost for acquiring information on distribution networks, the cost to find

the right component suppliers or the investors to find joint-venture partners. These costs also include

adapting exporters’ products to consumer preferences in a given country, adapting the marketing strate-

gies to a specific context and adjusting commodities to different qualitative and technical standards (see

Rauch 2001).

©2016 John Wiley & Sons Ltd 1345

The World Economy (2017)

doi: 10.1111/twec.12431

The World Economy

liquidity constraints and advantages in their human capital that imply lower costs. Indeed, on

the one hand better education is likely to help ethnic networks to boost their informat ional

advantages so that migrants may become more effective in promoting and facilitating interna-

tional trade; on the other hand, more educated migrants could, in principle, strengthen the

preference channel simply because they earn and possess higher incomes (see Felbermayr

et al. 2012). Building on this literature, this paper tests whether, and to what extent, the rela-

tionship between ethnic networks and trade varies with product quality. More precisely, we

investigate how the pro-trade effect of immigrants varies with the quality of traded products

over the period 1995–2000. To our knowledge, the link between product quality and the pro-

trade elasticity of ethnic networks has not yet been explored in the literature. Existing studies

mainly focus on the variation of the pro-trade effect of ethnic networks due to different levels

of goods’ heterogeneity, following the methodology adopted –among others –by Rauch and

Trindade (2002). We adopt a similar approach by classifying traded goods according to their

quality level and separately estimating pro-trade elasticity of ethnic networks for each sub-

group.

We expect that the relationship between product quality and pro-trade elasticity of ethnic

networks will be highly dependent on the North–South specialisation across varieties. On the

supply side, since emerging and OECD economies specialise in different bundles of varieties

(see Schott 2004; Fontagn

e et al. 2008), ethnic networks will mostly facilitate imports of

those varieties for which their countries of origin have a comparative advantage. On the

demand side, given difference in the market positioning of the various exporters on their dif-

ferent destination markets (see Fontagn

e et al. 2008), we expect ethnic networks to be more

effective in promoting exports to their homeland of those varieties for which there is rela-

tively higher demand. Furthermore, our hypothesis is that the relationship between product

quality and pro-trade elasticity of ethnic networks will largely depend on the composition of

the stock of immigrants by skill level. Given their lower liquidity constraints and advantages

in human capital, the pro-trade effect of high-skilled migrants is likely to affect relatively

more high-quality goods. Our empirical analysis aims at testing these conjectures.

We follow the trade–migration literature by incorporating the stock of immigrants (whole

stock and high skill) into an augmented gravity model. The proxy for ethnic networks is

included in a general gravity expression derived from a supply-side Ricardian model of trade



alaEaton and Kortum (2002) which predicts the North–South specialisation across quality

segments that emerge from the findings of Schott (2004) and Fontagn

e et al. (2008). Along

with the other standard trade cost proxies in the econometric specification, we also include

HS-6 digit industry fixed effects which enable to capture non-observable industry specific

characteristics. As in Gould (1994), we distinguish between imports and exports and we sepa-

rately analyse the effect on differentiated products. Our data set covers approximately 30 per

cent of the global stock of migrants: the extensive country coverage –177 countries of origin

and 19 OECD destinations –attenuates the sample selection bias due to the specific choice of

the countries entering the analysis and it allows to exploit differences between countries at

different income levels. Following Fontagn

e et al. (2008), we utilise unit values as a proxy

for product quality. As in Hallak (2006), we assume that all cross-country variation in export

unit values can be attributed to differences in quality.

Five main results stand out: (i) as the sample expands to include emerging economies, the

pro-trade effect of immigrants decreases significantly; (ii) for all levels of income and for any

level of quality the high-skilled ethnic networks have a stronger impact on trade; (iii) the pro-

export effects are always larger: this seems to suggests that the trade cost channel of

©2016 John Wiley & Sons Ltd

1346 G. GIOVANNETTI AND M. LANATI

migration is likely to dominate the preference channel; (iv) the trend of the pro-trade effect of

immigrants over quality is driven by the North–South specialisation across varieties for both

supply and demand; and (v) these trends apply regardless the degree of product differentiation

based on the classification proposed by Rauch (1999).

2. LITERATURE REVIEW

a. Trade–Migration Link

Since the seminal contribution of Gould (1994), several papers using different samples, time

coverage and econometric techniques have investigated the relationships between bilateral trade

and migration flows to find a strong and significant empirical correlation between the stock of

immigrants in the destination country and the volume of trade with their country of origin (see

Parsons and Winters 2014). As mentioned above, the underlying idea is that migrants have a

comparative advantage in conveying reliable information on markets which are very different

from the host country. These could be the origin countries but also countries which are similar to

the origin in terms of religion, culture and structure of the society. The majority of the contribu-

tions study the pro-trade effect of immigrants into a single country, while relatively few papers

focus on a multicountry analysis. With the recent availability of more and better data on migrant

stocks, some studies also exploit the regional distribution of immigrants and look at the bilateral

trade relationship between regions (or provinces) and foreign countries.

4

Three main stylised facts emerge from the literature: (i) the trade–migration link appears

stronger for differentiated goods than for homogeneous commodities; (ii) the effect of immi-

grants on imports is typically estimated to be larger than the one on exports; and (iii) there is

ample evidence of a stronger pro-trade effect for high-skilled migrants.

1. The first stylised fact implies greater import ance of ethnic networks in reducing infor-

mation costs for more differentiated goods. This rather intuitive statement has been

tested empirically mostly by dividing the spectrum of traded goods into three broad sub-

classes that differ with respect to the degree of differentiation according to the classifi-

cation proposed by Rauch (1999).

5

By running a gravity model separately for each

aggregated group, Rauch and Trindade (2002) estimate separate elasticities of trade with

4

Genc et al. (2012) analyse the distribution of immigration elasticities of imports and exports across 48 stud-

ies that yielded 300 observations: they report the meta-modal elasticities of immigrants which are, respectively,

0.12 for exports and 0.15 for imports. Among the main contributions on a single-country analysis of the pro-

trade effect of immigrants we cite Dunlevy and Hutchinson (1999) for USA, Head and Ries (1998) for Canada,

Tai (2009) for Switzerland and Girma and Yu (2002) for UK. The most important articles on a multicountry

analysis are Felbermayr and Jung (2009), Egger et al. (2012), Felbermayr and Toubal (2012), Aleksynska and

Peri (2014), Ehrhart et al. (2014). Lastly, the most influential papers that study the bilateral trade relationship

between regions (or provinces) and foreign countries are Herander and Saavedra (2005) for USA, Wagner

et al. (2002) for Canada, Bratti et al. (2014) for Italy, Peri and Requena-Silvente (2010) for Spain, Combes

et al. (2005) and Briant et al. (2014) for France.

5

Peri and Requena-Silvente (2010) and Aleksynska and Peri (2014) use Broda and Weinstein (2006)

classification to characterise the degree of differentiability of traded products according to their elasticity

of substitution across varieties. Although Peri and Requena-Silvente (2010) and Aleksynska and Peri

(2014) use a different classification of goods to characterise the degree of differentiability of products,

they follow the same procedure of grouping these products into three broad categories: highly differenti-

ated, moderately differentiated and less differentiated.

©2016 John Wiley & Sons Ltd

HIGH SKILL MIGRANTS VERSUS HIGH QUALITY TRADE 1347