Understanding International Milk Price Relationships

Abstract

Price studies have been extensively investigated in agricultural economics literature. In the grain market, it seems that more information is available in terms of price behavior and relationships across markets. In the dairy market, on the other hand, price information is more limited. There are studies related to milk price relationships, but none explore the short-run and long-run relationship between milk prices across countries. This analysis aims to fill this gap in the literature. Insights from this work, which may help private companies and policymakers on how milk prices behave across the world, are provided. Data are offered by the IFCN Dairy research network, located in Kiel, Germany, and includes monthly price data from the United States, Brazil, Germany, the Netherlands, Russia, South Africa, India, China, and New Zealand. A Vector Error Correction (VEC) model is used to summarize the long-run relationship among prices from these nine countries. Contemporaneous innovations are modeled as an acyclic graph using recently developed algorithms from the machine literature. Forecast error variance decompositions are also estimated to visualize the relationships among variables in the system. As for validation, out-of-sample forecasts from the estimated Vector Autoregression (VAR) and VEC models are performed and the latter performed better. This forecasting exercise also supports the imposition of a low number of cointegrating vectors, which coheres well with trace tests and information criteria. The United States, New Zealand, and the IFCN milk price play an important role in the international dairy market.