Productive Structure and Regional Economic Development in Cotopaxi, Ecuador: A Sectoral Analysis from 2001 to 2023

This study describes the evolution and interrelationships of the main productive sectors in the Province of Cotopaxi, Ecuador, over 2001–2023 (N=23 annual observations), using provincial accounts from the Central Bank of Ecuador (BCE). The available series spans a legacy 2001–2006 statistical regime and a later 2007–2023 regime with a different monetary basis; sector definitions were therefore harmonized, and each sector-indicator series was standardized separately within its source regime before correlation analysis. Two-tailed Pearson correlations were calculated for Gross Value Added (GVA), Gross Production, and Intermediate Consumption. Robustness was evaluated with detrended residual/partial correlations that control separate linear trends by source regime and with within-regime first differences that exclude the 2006–2007 transition. In standardized levels, the strongest correlations with total GVA were observed for services (r=0.957, p<0.001), commerce (r=0.953, p<0.001), and agriculture (r=0.933, p<0.001). Gross Production showed the strongest associations for commerce (r=0.959, p<0.001), agriculture (r=0.946, p<0.001), and services (r=0.888, p<0.001), while Intermediate Consumption was most strongly associated with agriculture (r=0.958, p<0.001), commerce (r=0.948, p<0.001), and industry (r=0.887, p<0.001). The robustness checks substantially attenuated several level relationships; for industry, signs of GVA association reversed from positive at the level to negative after detrending; meanwhile, commerce and services retained comparatively consistent short-run associations in annual changes in GVA and Gross Production. The results are therefore interpreted as descriptive associations rather than causal effects: Pearson correlation coefficients cannot, by construction, establish the direction or mechanism of influence between sectors. Beyond these sectoral findings, the central contribution of this study is methodological: it illustrates why detrended and first-differenced robustness checks are necessary when Pearson correlations are computed from data spanning incompatible statistical regimes, a common challenge in regional statistics for developing countries.