Cultural Gravity Effects among Migrants: A Comparative Analysis of the EU15

abstract

This article introduces cultural gravity as a concept that serves to better disentangle the direction and magnitude of the effects from migration, which is controversial in recent literature. The aim is to test for cultural gravity effects on both the geographic concentration and human capital productivity of immigrants in the EU15 countries. Operationally, we proceed to construct an empirical cultural gravity measure and test it with the use of a composite cross-sectional database, comprising, inter alia, the World Value Survey and Eurostat Census data. After an initial exploration of relevant cultural data by means of multivariate statistical analysis, we present an extended formulation of a gravity model approached through logistic regression methods and a three-stage least-squares estimation. Our results clearly demonstrate the existence of a cultural gravity effect among immigrants. Finally, an interesting finding is that cultural gravity also plays a significant role in the context of the Culture-Based Development (CBD) growth model.

Key words::

• immigration
• gravity
• cultural distance
• human capital
• cultural capital
• productivity

JEL codes:

• Z10
• O31
• R12

Notes

Since the 1990s, gravity models, able to accommodate these notions, have appeared more prominently in economic textbooks (e.g., Feenstra ). Indeed, Ravenstein () pioneered gravity models in the context of migration in the United Kingdom during the nineteenth century. Yet, later on, gravity became largely associated in the economics literature with geographic distance and trade, and it gained much popularity through a series of empirical studies that used the latter approach (see, e.g., Cornish ; Anderson and van Wincoop ; Bergstrand and Egger ). For a detailed review of the development of gravity models and their application, see Anderson ().

The existing literature on the effect of networking (or the existence of ethnic/social networks) on migration has highlighted the role of social and cultural elements in migration (see, e.g., Kloosterman, Leun, and Rath ). The mechanism as to why social networking originates however, is rather obscure. Tubadji and Nijkamp (forthcoming) provide an interpretation of social networks in the spirit of Axelrod’s understanding of culture as a coordination game determinant. Therefore, the authors interpret culture as a lattice within which a cultural percolation of new ideas is taking place, in which culture affects local productivity through the innovation channel. However, the present article includes both a quantitative and qualitative dimension of the cultural factor (i.e. both a direction [cultural distance] and a magnitude [type of open or closed cultural milieu]), which is a step further in elaborating our understanding of the intricate mechanism of cultural impacts on local productivity and makes this impact rationally predictable.

There are multiple studies regarding the importance of cultural distance (see, e.g., Kogut and Singh ; Grinblatt and Keloharju ; Guo ; Tihanyi, Griffith, and Russell ; Lee, Shenkar, and Li ; Lucey and Zhang ), which demonstrate that different behaviors in financial investment and cooperation between localities depend on their cultural similarity or difference. The research on cultural distance is also backed by the stream of literature that investigates how culture determines investment decisions through the cultural home bias (Duru and Reeb ; Chan, Covrig, and Ng ; Nijkamp, Gheasi, and Rietveld ) or through various other cultural effects on investment decisions (see, among others, Guo ; Tihanyi et al. ; Lee et al. ; Lucey and Zhang ). It is also interesting that the gravity model in relation to the spatial distribution of foreign direct investment (FDI) has recently found original applications (Head and Ries ) and touches also on cultural distance (for an example, see Diyarbakirlioglu ). The findings in the FDI literature however, are also ambivalent in that cultural distance is sometimes found to influence FDI positively, while it is also known to affect negatively risk evaluation through a home bias or at best arrives at the inexplicability of the irrationality of the investors’ risk-taking behavior (see, e.g., Tadesse and Shukralla ).

CBD assumes that indeed the cultural gravity shapes the economic growth process in a locality rather than just a marginal or incremental part of it. This is due to the assumption that cultural gravity is moved by culture, which in its essence is a proto-institution (i.e., the matter shaping all institutions in a locality and creating an initial conditions setting for any particular economic momentum). For more details on culture as a proto-institutions, see Tubadji ().

From the perspective of a locality B, since local culture changes slowly in terms of actual distance, the only way to intervene and regulate the cultural gravity effect is through the regulation of the magnitude of the cultural capital vector between openness and closedness of the local milieu.

For a recent review on migration models, see Baycan and Nijkamp ().

Productive capital is defined here in terms of labor and economic capital, as shown in Baycan and Nijkamp ().

Differences between interlocal and intralocal applications of the model cannot be handled with the available data due to the low number of observations. We could incorporate country dummy variables capturing such differences, but this would increase the number of regressors and decrease further the already low degrees of freedom for our estimation. For a further explanation of the inter- and intralocal application of the model, see Tubadji ().

Constructing an index requires intelligible behavior and clear interpretation of this quantitative measure. In the case of the Cultural Gravity Index, the measure has two parts: milieu and distance. The first is targeted at indicating a direction, the second at representing the quantity of the effect. When considered as continuous variables, in a sense, both milieu and distance approximate the same phenomena: cultural capital. So, if both measures are taken as continuous variables in constructing the Cultural Gravity Index, they would be carrying correlated information on the quantity of this cultural capital they approximate. By using a discrete measure for the openness of the milieu, however, we manage to obtain information on another aspect of the characteristics of the local cultural capital—the direction of the effect that can be expected, allowing us to think of culture in terms of a treatment that occurs when a certain tipping point of openness is achieved (for more on culture and tipping points for development, see Tubadji and Nijkamp ).

CBD normally expands the first equation of this model with a variable on investment in human capital as a reflection of the endogeneity of economic growth (Tubadji , ; Tubadji and Nijkamp ). However, such data are not available for the data set of interest, so we stick to the standard neoclassical paradigm.

The HI used throughout this article is measured as the sum of the squares of shares of nationals and shares of each immigrant group present locally.

Therefore, the HI results are not recorded in a table here.