The J-curve effect and US agricultural and industrial trade

This paper examines the J-curve hypothesis for US agricultural and manufact ured goods, using the Shiller lag model. The results support the J-curve ef fect for agricultural goods, but not for manufactured goods. These findings explain why many studies in the literature fail to support the J-curve phe nomenon. There are two explanations for these findings: (1) the aggregation bias of data that combine both agricultural and manufactured goods and (2) the country under study is often an industrial nation like the US or Japan with a high proportion of manufactured goods in both exports and imports.