The great supply-chain massacre: better data and business intelligence needed
Leading up to the 2008 global financial crisis, a few prescient voices warned of potentially catastrophic systemic instability.
In a famous 2005 speech, Raghuram G Rajan cautioned that although structural and technological changes meant that the financial system was theoretically diversifying risk better than before, it might, in practice, concentrate risk. He was mocked; former US Treasury Secretary Larry Summers was not alone in thinking him a “Luddite”.
This episode comes to mind because of the widespread shortages emerging around the world. Markets for gas, truck drivers, CO² (extraordinarily), toys, computer chips and more have been affected. Will these supply shocks prove a temporary disruption as the global economy recovers from Covid-19? Or are we witnessing a meltdown of the global production system? What would be the supply-chain equivalent of leading central banks’ interventions to prevent a financial collapse in 2008?
The parallels between today’s supply shocks and the 2008 financial shocks are striking. Before each crisis, the prevailing assumption was that decentralised markets would provide adequate resilience, whether by spreading financial risks or by ensuring a diversity of alternative supplies.
In the energy sector, there has been a steady shift away from national self-sufficiency toward reliance on global markets.
The EU started the “liberalisation” process in 2008, enabling new competition in gas and electricity, intended to be an EUwide market. Although some had expressed concerns about the implications for security of supply, policymakers pressed ahead with legislation to entrust European economies’ energy imports to global markets.
Most analysts – and policymakers – failed to anticipate that the global markets for gas and many other commodities would have bottlenecks or gatekeepers. The supposed diversification of supply resulting from liberalisation frequently seems to be illusory.
For products, including semiconductors or CO² (a fertiliser by-product) for food processing, supplies have become more concentrated. Splitting global production chains into more specialised links over several decades has led to close correlations between supply shocks in different industries, as with fertiliser and food or semiconductors and cars.
In addition, some shortages (such as those of truck drivers and shipping containers, or fuel in the UK) directly affect the logistics connecting the links in supply chains. As a result, vulnerabilities have rapidly become mutually reinforcing and self-amplifying.
The global production system’s highly specialised, just-in-time design delivered substantial benefits, but its weaknesses are now evidently greater.
So, how should policymakers think about this lack of system resilience, and what can be done to counter it?
Northwestern University’s Benjamin Golub has shown that queuing theory offers insights into how a small change in a well-functioning system (such as cutting two supermarket checkout lanes down to one) can lead to huge increases in wait times. Conversely, introducing a little slack into a system adds a lot of resilience.
Likewise, the classic cobweb model shows how time lags can destabilise markets and trigger large fluctuations in demand and supply. If demand is less responsive than supply to price signals, and expectations about the future prove incorrect, then a delay in suppliers’ responses drives volatility. Economist W Brian Arthur’s famous El Farol bar problem, which combines decisions made over time with the need to form expectations, also produces an unstable outcome.
So there are mental models for understanding the current problem. The challenge is how to restore stability and ease the shortages. A top priority is to have better data and business intelligence in government.
Even after 30 years of globalisation, there is astonishingly little detailed, publicly available information on product flows in global supply chains. Ministries need to restore the kind of engineering-based industry knowledge that was more common back when industrial policy was considered a key government function.
But in the short term, decentralised markets and price signals are the problem, not the solution. Governments will need to step in – whether by deploying soldiers to drive fuel tankers or providing production subsidies – to mitigate some shortages.
When the immediate supply concerns abate, firms and policymakers must consider what kind of insurance or slack they should build into the production system over the longer term.