In the present research we focus on the issue of growth convergence at both country and regional levels within the EU boundaries. We build on the theoretical and empirical literature on club convergence that points to the role of heterogeneity in initial factor endowments and structural factors in determining multiple long-run growth paths, and use techniques recently proposed in the literature to first assess
convergence/divergence in real per-capita incomes and productivity, and then identify endogenously convergence clubs: the mixture and semiparametric models, and the Phillip and Sul (2007) nonlinear factor model. We then employ alternative methods to assess empirically the factors conditioning club convergence, the strength of club membership, the transitional dynamics and interclub mobility. From our results we derive theory-based policy implications for convergence/divergence within EU. Our research is relevant for understanding the resilience of the integration process to adverse shocks and for highlighting the factors that withhold growth convergence within the EU boundaries, in order to implement the appropriate policies to tackle a multispeed Europe. A special focus will be on the link between migration, and the ensuing cultural diversity, growth and convergence. Among the various mechanisms providing a growth enhancing role for migration and diversity, the analysis will center on their risk sharing implications.
This research is innovative with respect to the current literature in several ways.
In part A, the research will use mixture models to estimates the equilibrium shape of the productivity distribution going beyond the mere description of the dynamics of the distribution function shape, with a view to addressing the policy implications underlying such dynamics. We expect to contribute to this literature as follows. First, we shall focus on regional data rather than country data, and discuss productivity issues rather than income inequality. Second, we shall pay attention to the comprehension of the macroeconomic forces that might have shaped the productivity structure and evolution of clusters of contemporary Europe, a point left behind in the literature.
In Part B we shall contribute to the club convergence literature that employs the PS methodology by considering all 28 EU countries and focusing on the evolutions occurred in more recent years, following the developments of the Euro-area sovereign-bond crisis. We are particularly interested in assessing the resilience of the ongoing integration process to adverse shocks by measuring weather distance across countries and multiplicity of equilibria have increased. Secondly, following Bartkowska and Riedl (2012), Lyncker and Thoennessn (2017), we shall look at the role of initial conditions and structural factors in predicting club membership. Given the low number of observations (countries rather than regions) compared to the possible predictors, we shall not resort to traditional ordered logit estimations, but rather employ regularized techniques as the LASSO (least absolute shrinkage and selection operator) for Logit estimations, that impose constraints on the estimation to deal with sparsity. This will allow to consider the variables drawn from theory and typically considered important in the formation of clubs, (capital, human capital, initial income), as well as a number of additional drivers related to structural factors (i.e. institutional quality, global competitiveness, etc.). At the same time, this methodologies allow to deal with the issue of collinearity, which is likely to arise, and the variable selection problem. To the best of our knowledge, there are no other attempts in the literature to test the conditioning factors of club convergence by means of regularised techniques. From the computation of the marginal probabilities, we shall gain important insights on the impact of each predictor for club membership, with important implications for the appropriate structural reforms to put in place for income convergence.
This research will bring about some innovations also with respect to part C, dealing with the effects of migration on growth and productivity convergence.
As mentioned above, there is a conspicuous amount of literature dealing with the effects of migration and, to a lesser extent, of diversity on growth, but to the best of our knowledge no contributions have directly dealt with the impact of migration, and diversity, on convergence (in income or productivity). A particular focus will be on the effects of migration and diversity on fostering risk sharing in recipient economies and, consequently, induce more investment and more growth.
This is the new avenue we would like to explore, and to do so we will have to enrich the above mentioned empirical models with suitable measures of migration and of diversity as explanatory variables. The choice is simple as far as migration is concerned (stock of migrants, characterized along several possible socio-economic dimensions), but less so in the case of diversity.
Among other measures of diversity that have been found to affect economic performance, we may acknowledge a polarization index, which is such that it reaches its maximum when the population splits in two equally sized groups. Ager and Bruckner (2013) compare this to the previous parameter finding opposites effects, with polarization potentially capturing the scope for conflicts among groups. On the other hand, Bove and Elia (2017) find positive effects of both measures on GDP per capita, particularly in developing countries. Rodriguez-Pose and Hardy (2015) use instead Theil entropy indices, which are expected to be more sensitive to the distribution and variety of groups, exploiting both professed ethnicity and birthplace as diversity criteria: their evidence suggests that both measures positively correlate to entrepreneurship in England and Wales, although birthplace diversity is more relevant to knowledge-intense entrepreneurship. Finally, another interesting criterion relies on genetic diversity, which Ashraf and Galor (2013) find to have a world-wide hump-shaped effect on development, but would also appear to be significantly related to long run economic performance in the US (Ager and Bruckner, 2013).
In our research, we will also explore the different implications of such varied measures on the phenomenon we wish to analyse.