Climate Change and Adaptation in Global Supply-Chain Networks

Do buyers drop suppliers when the weather turns unexpectedly hot?

Nora M. C. Pankratz and Christoph M. Schiller examine how physical climate exposure at supplier locations travels through global production networks in "Climate Change and Adaptation in Global Supply-Chain Networks".

They combine supplier and customer relationships from FactSet Revere with local temperature records, flood footprints and long-run temperature projections, in over 70 countries between 2003 and 2017.

Their central measure compares the heat and flood days a supplier actually experienced during a relationship with the level the customer could reasonably have expected before it started.

Their main conclusions include:

• Heat at supplier locations reduces the operating income of suppliers and of their customers, so exposure sitting one step upstream reaches the accounts of firms whose own sites were never touched.
• When realised exposure at a supplier exceeds what the customer could have expected at the outset, the customer becomes around 7% more likely to terminate that supplier relationship.
• The response strengthens as the deviation grows larger and as it repeats over successive years, rather than reacting to a single unusual season.
• The effect decreases with country-level climate adaptation, so identical weather signals carry different consequences depending on how well the country in which the supplier operates has adapted.
• New suppliers show lower expected and lower realised heat exposure than the firms they replace, although their projected future exposure is similar.
• Flood exposure at supplier locations produces similar results, which points to a mechanism driven by disruption beyond what the buyer anticipated rather than by temperature alone.

This article shows mapping suppliers' sites against realised heat and flood days, and against the expectations held when contracts were signed, points out the relationships most likely to break.

The authors note that the regions most exposed to climate change are also the least developed, so substitution may weaken them further and hinder their own adaptation capacity.

The sample covers listed suppliers and customers only, so smaller unlisted suppliers, often the most exposed, sit outside it. Measured effects are therefore likely to be a lower bound.