The degree of spatial market integration in European Community agricultural product markets is investigated. The methodology is based on co‐integration analysis. A two‐step testing procedure is applied to test the existence of the ‘Law of One Price’ which should be in force if these markets are truly integrated. The results show that market integration has not been achieved in a number of cases even when MCAs are subtracted from the price series used.

No abstract

No abstract

Co-integration analysis is used in a study of the advertising and sales relationship using the Lydia Pinkham data set. The series are shown to have a valid long-run relationship while Granger-causality runs in both directions. The latter is found by using a causality test involving the cointegration restrictions which seem to constitute a crucial part of such tests in the case of co-integrated variables. A comparison with previous models shows that forecasting co-integrated series is more accurate with errorcorrection systems, especially in the case of multi-step forecasting.

The recent testing approaches of the 'Law of One Price' which are based on co-integration analysis are modified by incorporating the seasonal components of the agricultural price series into the testing procedure. Application of the modified testing approach to the soft wheat market of five European Union member stales produces mixed results as some of the markets turn out to be integrated while ill some cases a unified market cannot be assumed, Theseresults differ in some cases from those obtained by co-integration tests which ignorc seasonal unit roots.!', 1999 Elsevier Science B.V. All rights reserved.