From Comparative Advantage to Forecast Regimes: Integrating Structure and Dynamics in Export Growth

This study examines whether structural comparative advantage translates into predictable exportgrowth. Using 40 HS-level Vietnamese export series over 2005–2024, we integrate triangulatedrevealed comparative advantage measures (RCA, RSCA, and gravity-adjusted indices) with adiagnostics-driven time-series framework combining ARIMA, ARIMAX, and ARIMA–GARCHmodels under rolling-origin validation. The findings reveal three distinct forecast regimes—structure-dominated, mixed, and shock-dominated. Manufacturing sectors display low forecastdispersion (MAPE < 3%) and persistent stochastic drift, while commodity-based sectors exhibit highvolatility and wide error bands. Methodologically, the study bridges static specialization metrics anddynamic forecasting through a unified regime-classification protocol. Theoretically, it interpretsARIMA drift as a reduced-form signal of cumulative capability formation and structural upgrading.Empirically, it provides the first systematic forecast-regime mapping of Vietnamese HS-levelexports. Practically, it translates statistical regimes into differentiated policy strategies for capacityexpansion, upgrading, diversification, and resilience. By linking comparative advantage to stochasticgrowth dynamics, the paper advances trade theory beyond static specialization and offers anoperational framework for export strategy under uncertainty.