Estimating Long-Run Price Relationship with Structural Change of Unknown Timing: An Application to the Japanese Pork Market

Linear regression is often used to measure economic relationships in order to understand the underlying structure and obtain forecasts. Clearly, a major advantage of a linear model is its simplicity, increasing the likelihood of obtaining a robust estimate of the parameter that is not overly sensitive to model specification and choice of numerical method. Within the framework of linear regression, this article discusses the new structural change paradigm and illustrates the procedures via an analysis of structural breaks in the retail-farm price relationship in the Japanese pork market. By design, an empirical study of structural change inevitably requires time series data. While regression involving nonstationary time series may be spurious. This article examined structural breaks in the retail-farm price relationship in the Japanese pork market using the endogenous break date estimation procedures of recent studies. Four breaks were identified for the period from 1967 to 2008, reflecting changes in production costs, the state of the economy, and trade regime.