<![CDATA[Previous coffee-market studies have linked domestic and international prices, but they have rarely examined long-run integration, short-run asymmetry, shock decomposition, predictive causality, and threshold adjustment in one empirical sequence for the North Sumatran robusta supply chain. This study addresses that gap by analysing farm-level or production-area prices (PFC), North Sumatra exporter-level prices (PEC), and world prices represented by London Robusta Futures (PWC) using 120 monthly observations from January 2016 to December 2025. The analysis applied unit-root testing, Johansen cointegration, VECM, FEVD, Granger causality, AECM, Wald tests, and Hansen-Seo threshold testing. All series were I(1), and the Johansen test identified two cointegrating equations, confirming long-run integration among farm, exporter, and world prices. VECM adjustment was significant for PFC (ECT = -0.426806; t = -5.07885) and PEC (ECT = -0.298370; t = -4.44265), while PWC behaved as the reference market. At the 10-month horizon, PWC shocks explained 51.15% of PFC variation and 40.50% of PEC variation, while PWC remained dominated by its own shocks (84.22%). Supporting four-market IRF evidence showed that the largest own-price responses occurred in the first period, were followed by sign reversals, and in several equations remained visible at period 10, indicating gradual rather than instantaneous stabilization. Granger causality was bidirectional between PFC and PEC and unidirectional from PWC to PEC. Short-run asymmetry was found mainly in the PFC-PEC relationship, whereas long-run asymmetry and threshold cointegration were not supported. These findings contribute to agricultural market-integration literature by showing that benchmark prices transmit risk to domestic actors through linear long-run adjustment and temporary short-run frictions. Policy should strengthen transparent price information, quality-based pricing, and farmer bargaining institutions.]]>
