Eduardo A. Haddad, Mahmoud Arbouch (p), Zaynab El Mourtaji, Ayat Lammalemi

Carlos Azzoni

Since the beginning of the century, Morocco has experienced a relatively long period of stable growth, with potential growth moving from around 4% in the first decade to 3% in the second. Such a decline in potential growth, which would probably be the result of an overall growth model that is significantly losing steam, is likely to become more pronounced after the recent economic crises. At the regional level, the state of growth potential is more contrasted with roughly half of the regions generating more than 80% of the national wealth, and only one region concentrating its third. This growth polarization resulted mainly from two key economic policies. On the one hand, Morocco has made a remarkable effort in providing public investment, which resulted in GDP investment rates constantly exceeding 30%, allowing thus the country and particularly its northern part (until recently), to be equipped with good quantity and quality infrastructure (High-ways, Ports, Airports, High-Speed Railways …) necessary to enhance Morocco’s growth potential and initiate a real structural change. On the other hand, Morocco declined its development strategy in many sectoral plans for the industry, agriculture, and tourism … with a major industrial plan to increase the value-added of the manufacturing sector and enable export-led growth in manufacturing. This paper aims to quantify the main drivers of changes in Morocco’s regional value-added using a Structural Decomposition Analysis. To do so, we differentiated six components that affect this change in regional value-added, namely: value-added coefficients, industrial structure, domestic absorption, and foreign exports. We rely on two sets of interregional input-output tables for 2014 and 2019, comprising 12 regions and 20 sectors. The main results show that regional Growth in Morocco remains driven by domestic consumption, with slight contributions from the remaining components.

Joseph Bahati Mukulu (p)

Mahmoud Arbouch

The topic of convergence has aroused great interest from many researchers. For a long time, the literature on economic convergence asserted that less developed countries (regions) should grow faster to catch-up with wealthy countries. Echoing scholars, globalization and technology transfer are among the drivers of convergence. However, for some decades, scholars contend the divergence between countries and that there is a club convergence phenomenon. The latter claim stresses that within the same group, countries converge while groups diverge. This raises the question of the conditions of convergence. We investigate the role of globalization in the convergence/divergence of African countries. This choice comes from the claim that different regions may have their own production frontier. We assume that the degree of openness will contribute to labour productivity growth and, therefore, will promote convergence. We extend the Kumar and Russel (2002), Henderson and Russel (2005), and Badunenko, Henderson and Houssa (2014) approaches by integrating the globalization intensity in the analysis of convergence. We rely on a panel of 41 countries over 19 years (2001 to 2019). We use the DEA production frontier methodology to assess the technical efficiency and compute the Malmquist Index of Productivity (MPI) that allows us to decompose labour productivity into its components. Our findings reveal that globalization is a source of labour productivity growth in Africa. Therefore, ignoring globalization when analyzing convergence biases conclusions by overstating the role of physical capital accumulation and understating that of human capital accumulation. We compute the Kernel Density Estimation to check whether African countries converge and plot regression lines to grasp the contribution of each component to convergence. Our results show a polarization in Africa since the distribution of labour productivity is bimodal: technological progress and human capital accumulation are the sources of divergence and polarization of African economies, and technological catch-up (efficiency change), physical capital accumulation and globalization intensity change are the drivers of convergence in Africa. This study has the merit of using a “holistic” measure of globalization that integrates different dimensions of the globalization concept. However, this study has some limits, including (1) the lack of a depth analysis per sector to grasp how the spillover effects from globalization disseminate across sectors; (2) the period analysis is short (19 years). Accordingly, getting inside the black box of productivity by analyzing different sectors over sub-periods and extending the analysis to a long period can yield interesting insights.

Zongnan Hu (p), Shougeng Hu

Joseph Bahati Mukulu

This study examines the impact of the structure of cooperative networks between cities on the development of a region and individual cities, utilizing concepts of constraint and competition from structural hole theory. The study employs a multi-agent model constructed from two sets of networks - one conceptual, and one real and complex - to test the hypothesis. The study contributes a new perspective on the interplay between cooperation and competition in urban networks, providing a tool for assessing the impact of urban network structure on development. The model has been completed and tested on a simple set of networks. To further validate the findings, a collection of 7875 train schedules from China was utilized to construct a more realistic and complex cooperative network, testing whether the theory holds for complex urban networks.

Carlos Azzoni (p), Gustavo Castro

Zongnan Hu

Objective
This study deals with regional skill convergence in private manufacturing in Brazil.

Topic of the contribution
We show evidence that the dynamics of convergence of skills in manufacturing depend on spatial and non-spatial variables and the type of industry. The regionalization method utilized can serve as a guide to finding ways to regionalize the sector.

Method and data
We use a hierarchical cluster method with contiguity constraint and a quintile analysis to identify spatial heterogeneities, in addition to ML techniques. The baseline and the heterogeneities were estimated using spatial econometrics. We use microdata of official statistics of formal employment.

Main Results
We show that social skills converge faster than motor and cognitive skills in manufacturing in the Brazilian regions. When analyzing the top 25% of skills, we found that cognitive skills converge faster than social and motor skills. The evidence shows that cognitive and motor skills converge faster for capital-intensive and non-agribusiness than labor-intensive; the other sectors converge faster in social skills. Less industrialized regions converge within their clusters, but more industrialized regions are not. Moving from skills to formal education, we found convergence of schooling levels for all clusters, meaning that regions converge on schooling within their clubs. There is a mismatch between the required skills (labor demand) and formal education (labor supply): firms in the regions are demanding the same skills, although workers are becoming more educated.

Keywords: manufacturing industry, skills, spatial econometrics, machine learning

JEL: C33, C38, J24