Industrial Structure Similarity and Urban Growth in the Chengdu–Chongqing Region: Exploratory Evidence and Spatial-Unit Sensitivity

This study presents an exploratory assessment of industrial-composition similarity and subsequent economic growth in the Chengdu–Chongqing region during 2000–2020, with an emphasis on inferential uncertainty and spatial-unit comparability rather than on presumed growth effects. We calculate the initial three-sector cosine similarity relative to Chengdu and Chongqing across four five-year periods and reconstruct real GDP growth from annual comparable-price indices for 15 study units. Neither the eight HC3-robust linear models nor the eight similarity-threshold tests are significant at the 5% level after the relevant multiple-testing adjustment. An isolated 2010–2015 tertiary-share-change threshold remains significant after the original Holm correction (adjusted p = 0.0350) but is no longer significant when Chongqing municipality is excluded (bootstrap p = 0.8990) or replaced by a fixed-name nine-district proxy (bootstrap p = 0.7830). These findings demonstrate the sensitivity of exploratory nonlinear results to small cross-sectional samples and non-equivalent administrative units. Three-sector similarity describes compositional resemblance not intercity competition, coordination, or causal spillovers. The paper contributes a dual-reference descriptive comparison, transparent reporting of null results, and a spatial-scale sensitivity diagnostic; it does not establish a similarity-driven growth effect or a competition-to-coordination transition.