Structural Complementarity, Bottlenecks, and Total Factor Productivity in Mexico (2000–2025): A Matrix and Critical Political Economy Approach
Year: 2026
Type: preprint
Abstract: Traditional modeling of regional economic growth and Total Factor Productivity (TFP) in emerging economies has relied on continuous neoclassical specifications assuming unitary elasticity of substitution. This paper demonstrates the analytical insufficiency of such frameworks to explain persistent subnational divergence across Mexico, where localized human capital accumulation fails to mechanically translate into proportional technical efficiency gains. Engaging in critical dialogue with evolutionary and radical political economy, we examine the fundamental epistemological tensions underlying aggregate capital and TFP accounting, formalizing a multidimensional matrix architecture that synthesizes Kremer’s (1993) O-Ring production theory, Liebig’s Law of the Minimum, input-output shock propagation through intersectoral networks (Acemoglu et al., 2012), and power asymmetries across Global Value Chains (Kapeller et al., 2024; Durand & Milberg, 2020). Utilizing the Adjusted Mazziotta-Pareto Index (AMPI), internal horizontal disparities across complex labor skills and critical infrastructure vectors are penalized non-compensatorily. Coupled Leontief inverse matrices derive sector-specific co-requirement structures, projecting realized factor absorption. Furthermore, applying local projections (Jordà, 2005) across five historical macroeconomic disruptions, we derive normative resilience matrices via Moore-Penrose pseudoinverses, auditing spatial mismatch through Mahalanobis distances and cosine similarity. Panel threshold regressions (Hansen, 1999) empirically validate the bottleneck hypothesis: the marginal return on technical skills is statistically indistinguishable from zero in states below the critical infrastructure threshold (γ 1 = 92.40), switching to a highly significant regime once surpassed (θ 2 = 0.428, p < 0.001). Two-step System GMM dynamic panel estimations with Windmeijer (2005) correction confirm structural supermodularity (β 3 = 0.015, p < 0.001) and significant penalties from Mahalanobis mismatch (β 4 = -0.142, p < 0.001). We deconstruct the nearshoring phenomenon through the lenses of capital’s spatial fix (Harvey, 2001), the maquiladora regime and labor market segmentation (Delgado Wise & Cypher, 2007; Marini, 1973), as well as biophysical boundaries and ecologically unequal exchange (Althouse et al., 2023; Obeng-Odoom, 2025). Finally, we categorize Mexico’s 32 states into four operational quadrants, contrasting conventional technocratic policy prescriptions with sovereign planning strategies for regional development. JEL Classification : O47, R11, C23, D57, B51.
Source: Research Square
Institution: Universidad Autónoma de Nuevo León
Language: English
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7233596542
