Diversification in the Manufacturing Industry of Ecuador: Grouping Factors by K-means Clustering

Abstract

This study aims to show the state of industrial diversification in manufacturing in Ecuador and to establish clusters that enable to differentiate the state between cantons that share similar industrial characteristics. The three-digit categorization of the International Standard Industrial Classification (ISIC) is used, including a total of 136 sectors in 221 cantons at the national level in 2019. In addition, the Shannon and Simpson indices were calculated as a diversification measure of the manufacturing industry, explicitly allowing us to quantify both the breadth and dominance structure of industrial activities across cantons. These indices, combined with Gross Value Added, schooling and the Environmental Promotion Tax, support the application of a multivariate K-means clustering approach. The method is particularly applicable for distinguishing territorial patterns of industrial structure, as it enables the identification of groups of cantons with similar diversification profiles and economic characteristics. The novelty of this study lies in integrating ecological-based diversity measures with clustering techniques to characterise manufacturing heterogeneity at a sub-national level, an approach not previously applied in the Ecuadorian context. This methodological framework is used to detect the optimal number of clusters, in order to increase the heterogeneity between groups and improve the analyses for each of them. Once the homogeneous clusters were established, the results showed that the cantons of Guayaquil and Quito have a supremacy over the rest of the regions, in terms of diversification of their manufacturing industry. Finally, more than 70% of the cantons do not show significant industrial diversification according to the estimated indices.