[en] This paper revisits Shift-Share and Constant Market Share Analysis by introducing a maximum entropy prior to replace ad hoc weighting schemes, assuming a uniform distribution of sectors. The resulting decomposition is additive and exact over discrete time periods, eliminating the need for a residual term. Both composition and competitiveness effects are derived from deviations relative to neutral expected growth.
Applying the method to global manufacturing data (1995–2019) for 76 countries at current and constant prices, the analysis distinguishes developed and developing countries trajectories. Adjustments for home bias and demand structure deepen the insight. Compared with traditional approaches, this probabilistic SSA confirms competitiveness as the main driver of industrial growth, with methodological implications for structural analysis and policy.
Disciplines :
Production, distribution & supply chain management Quantitative methods in economics & management
