Value Chains: Production Upstreamness and Downstreamness Revisited

Measures devised to quantify production upstreamness and downstreamness in value chains have been introduced and used increasingly in recent years. While the constructs embodied in these measures are meaningful, this paper identifies an overlooked implementation problem. Specifically, the algorithms have been applied either as though the underlying data represent flows, or under the assumption of a one-to-one correspondence between industries and primary products. Implementation data, however, are drawn from modern input-output accounting frameworks that recognize secondary production explicitly. Although they describe the use of commodities by industries, published Use matrices unadjusted are not conventional flows matrices because they do not identify the industries from which commodities originate. Neither do they represent commodity destinations. We demonstrate logical inconsistencies that arise from treating the Use matrix as a flow matrix, provide correct flow matrix formulations for upstreamness and downstreamness calculations, and present comparisons among empirical results from the correct and incorrect formulations.