Transnational ties and performance of immigrant entrepreneurs: the role of home-country conditions

Abstract

This study contributes to the recent empirical literature on the performance of transnational immigrants' firms by investigating the effect of transnational ties on the firm's growth. In addition to the effect of the ties, the paper shows that home country's institutional and socio-economic characteristics and country-specific entrepreneurial factors have a crucial role in shaping the ties–performance relationship. The evidence from a sample of immigrant-owned firms in the Italian information and communications technology (ICT) sector in the period 2000–2010 confirmed the relevance of the proposed model and helped in understanding a potential channel of improvements in immigrant firms' performance through transnational ties. Our results show the limited relevance of a direct, or linear, impact of ties on the growth of sales in immigrant-run firms in the ICT sector, but support the crucial moderating role of home-country characteristics on the ties–performance relationship.

Keywords::

• transnational entrepreneurship
• ethnic business ties
• immigrants' firms' performance
• ICT industry
• Italy

Notes

^1. A more positive picture of transnational entrepreneurs is presented by Miera (2008). She describes some advantages of Polish entrepreneurs in Berlin, who have access to transnational networks, over traditional Turkish IEs. As Polish entrepreneurs are highly mobile, constantly circulating between Poland and Germany, they are able to identify the differences in prices and income and conduct cross-border trade. Also, the immigrant-owned firms in the construction and cleaning sectors can rely on a growing pool of transnational workers, who circulate back and forth between those two countries, thus making them more likely to open branches in both countries.

^2. We should note that they use expression ‘diaspora international entrepreneurs’ instead of transnational entrepreneurship, but the meaning of the former is the same as the latter.

^3. On the other hand, by focusing on developing and emerging economies, we are able to investigate the possible impact of socio-economic environment on possible activation of transnational ties, as in those countries those characteristics remain more heterogeneous than in the case of the developed economies.

^4. We follow Coad and Rao (2008, 634) in adopting sales as dependent variable as sales growth is a particularly meaningful indicator of performance. We also rely on sales instead of employees (as in the case of Henrekson and Johansson 2010) because a productivity effect has been detected by a number of papers on the Italian economy. As far as the variable AGE is concerned, its use of the linear contribution is driven by our adoption of the Jovanović model, whereas the exclusion of its quadratic term is driven by the results of a t-test that rejects the joint significance of the linear and quadratic variable coefficients.

^5. A description of variables included in model (3) is given in Section 4.2 and Table A1.

^6. This approach corrects the bias due to the presence of a fixed effect. We have also tested our empirical specification by a panel dynamic model to consider endogeneity (see Arellano and Bover 1995 and Blundell and Bond 1998 for econometric reference, and Yang and Huang 2005 for an empirical application to firm growth). Based on standard evaluation tests (Sargan and AR1–AR2 tests), in this paper we prefer to rely on the results of fixed effects panel model.

^7. A t-test of the mean value of major variables has been performed by comparing the early and late respondents, i.e. mean values for the first 70 (about 15% of total sample) and the last 70 interviewed companies. These tests do not show any significant difference in the mean values of the following variables: sale growth, number of ties, average tie tenure, average firm age and employees. We have found moderately significant differences in the CEO education. The inclusion of this variable in the estimated equation help controlling for the sample heterogeneity.

^8. We thank one referee for suggesting us to explore this issue.