Did Industrial and Export Complexity Drive Regional Economic Growth in Brazil?

This repository contains the data necessary to reproduce the results of the manuscript available at https://arxiv.org/abs/2312.07469.Title: Did Industrial and Export Complexity Drive Regional Economic Growth in Brazil?Abstract: Research on productive structures has shown that economic complexity conditions economic growth. However, little is known about which type of complexity, e.g., export or industrial complexity, matters more for regional economic growth in a large emerging country like Brazil. Brazil exports natural resources and agricultural goods, but a large share of the employment derives from services, non-tradables, and within-country manufacturing trade. Here, we use a large dataset on Brazil's formal labor market, including approximately 100 million workers and 581 industries, to reveal the patterns of export complexity, industrial complexity, and economic growth of 558 micro-regions between 2003 and 2019. Our results show that export complexity is more evenly spread than industrial complexity. Only a few -- mainly developed urban places -- have comparative advantages in sophisticated services. Regressions show that a region’s industrial complexity is a significant predictor for 3-year growth prospects, but export complexity is not. Moreover, economic complexity in neighboring regions is significantly associated with economic growth. The results show export complexity does not appropriately depict Brazil's knowledge base and growth opportunities. Instead, promoting the sophistication of the heterogeneous regional industrial structures and development spillovers is a key to growth. This study demonstrates that industrial complexity, which accounts for all employment sectors, provides a more accurate basis for designing effective and inclusive industrial policies in emerging economies like Brazil, compared to export-based complexity.******DATA DESCRIPTION*****(1) neighboring_microrregions.xlsx<br>If two different microregions' codes are in the same row, they are neighbors.<br>(2) data_by_microregion_and_year.xlsx<br>The indices of the data are microregion and year. The microregions has a code (cd_micro), a name in Portuguese (cn_micro) and a Federal Unity name (nm_uf).- IndECI: The Industry-based Complexity Index (Equation 8a) of the respective microregion and year<br>- IndECI_neighbors: The average of IndECI of the neighbors (Equation 12) of the respective microregion and year- ECI: The Export Economic Complexity Index (Equation 3) of the respective microregion and year<br>- ECI_neighbors: The average of ECI of the neighbors (Equation 12) of the respective microregion and year- GDPpc: the real GDP per capita (at 2010 prices) of the respective microregion and year population: the population of the respective microregion and year<br>- growth_2: the growth of real GDP per capital in the following 2 years of the respective microregion and year- growth_3: the growth of real GDP per capital in the following 3 years of the respective - microregion and year<br>- growth_4: the growth of real GDP per capital in the following 4 years of the respective microregion and year- corr_ICI_density: the correlations between the industry-density (Equation 9) and industry complexity index (Equation 8b) of the respective microregion and year<br>- corr_PCI_density: the correlations between the the product-density (Equation 9) and product complexity index of the respective microregion and year<br>(3) data_by_year.xlsxThe index of data is the year.<br>- Moran's I ECI: the Morans'I (Equation 10) of ECI across Brazilian microregions in the respective year- Moran's I IndECI: the Morans'I (Equation 10) of IndECI across Brazilian microregions in the respective year<br>- skewness ECI: the skewness (Equation 11) of ECI across Brazilian microregions in the respective year- skewness IndECI: the skewness (Equation 11) of IndECI across Brazilian microregions in the respective year******FIGURE DESCRIPTION*****Figure 2A: simple correlation between measures of data (2)<br>Figure 2B: simple correlation between measures of data (2)Figure 2C: simple spatial distribution of measures of data (2)<br>Figure 2D: simple spatial distribution of measures of data (2)<br>Figure 3A: simple correlation between measures of data (2)Figure 3B: simple correlation between measures of data (2)<br>Figure 3C: plot of time series in data (3)Figure 3D: plot of time series in data (3)Figure 4: result of regressions based on data (2)<br>Table 2: the ranking of measures of data (2)<br>Table 3: regressions based on data (2)