A holistic network model for supply chain analysis

Supply chain researchers are experiencing a conceptual and analytical paradox. They are asked to move beyond dyadic analyses and investigate larger network effects with only a limited analytical toolkit. This research proposes the use of bilinear mixed-modeling to holistically analyze supply chain phenomena. Through this approach, researchers are able to account for multiple supply chain relationships, higher-order dependencies among member firms, and simultaneously evaluate covariates from buyer and seller perspectives. The model is validated through the lens of a pervasive supply chain problem commonly referred to as the bullwhip effect. A sample of firms from the US apparel industry in 2004 is analyzed and then the findings are confirmed using data from 2005. In addition to validating the model through the presence of the bullwhip effect, the bilinear model illuminates variables such as advertising, price deals, inventory turnover, and inventory backlogs that exacerbate or diminish inventory differences between firms in a supply chain. The results extend research on supply networks and supply efficiency to a more holistic level and show that higher-order dependencies are important drivers of supply chain phenomena.