Agricultural Market Integration Analysis

Because food prices are typically non-stationary, two unrelated price series that each trend over time can show a high correlation purely by accident — the classic spurious-regression problem. A high correlation therefore says little about whether the markets are economically linked. Cointegration analysis instead tests whether deviations from a long-run price relationship are stationary (mean-reverting), which is the proper statistical counterpart to the arbitrage idea that price gaps cannot persist beyond transfer costs. It also separates short-run divergence, which is normal, from the long-run tie that defines integration.

It measures the speed of adjustment back to the long-run price relationship. When a shock pushes the two markets' prices apart, the error-correction term captures how much of that gap is closed each period as traders arbitrage. A coefficient near minus one means almost the entire disequilibrium is eliminated quickly, indicating efficient, well-integrated markets; a small-magnitude coefficient means gaps persist and transmission is sluggish, pointing to high transaction costs, poor infrastructure, or barriers to trade. It is usually the single most policy-relevant number the analysis produces.

No, and conflating the two is a common error. Barrett and Li emphasise the difference between price equilibrium and genuine, tradability-based integration: prices can move together because the markets are linked by trade, but also because they respond to a common external driver (a shared exchange rate, a national policy, or world prices) without any goods flowing between them. Cointegration establishes a stable long-run price relationship; confirming that trade is the mechanism requires complementary evidence on trade flows, transfer costs, and the threshold band within which no arbitrage occurs.