Gillan Tett’s latest offering in the Financial Times discusses the woes that have befallen various major companies that find themselves exposed as a result of having extended supply chains that have Japan-based manufacturing as an important part. She correctly depicts this as a symptom of a much larger problem, of having pushed the idea of wringing out production costs too far. But perhaps due to space constraints, she fails to draw out the most important conclusion: just as with financial engineering, management incentives favored ignoring risk, and the resulting blow ups were predictable.
Tett tells us the Japan-related disruptions are merely the most visible symptom of a widespread pathology:
Last year, the Business Continuity Institute published a survey of companies which suggested that three-quarters had experienced production hiccups in their supply chain in the previous 12 months, due to unexpected surprises (ranging from weather to health issues to earthquakes). A quarter said that problems were getting worse. “In the highly competitive, global market of recent years, many businesses implemented cost-saving strategies to maintain profit margins, including just-in-time deliveries of critical resources and components,” says Nick Wildgoose, supply chain product manager at Zurich. “[But] some of those earlier savings are becoming operational weak links, especially in extended supply chains.”
It isn’t all that hard to understand that stressing efficiency at all cost comes at the expense of safety. Engineers will tell you that efforts that are pro-safety involve various forms of buffers and redundancy and are thus costly. During the early days of the crisis, commentators often discussed the implications of Richard Bookstaber’s book A Demon of Our Own Design, in which he argued that systems that lacked breaks between various processes were tightly coupled, which meant that a failure at one point in the process would propagate through the entire system, as if one transformer failing would bring down an entire electrical grid.
Anyone who has been on the periphery of manufacturing can see that all its fads can easily have pushed too many companies into failure-prone systems. Just in time inventory was a reversal of the historical propensity of manufactures to carry a lot of inventory to make life easier for managers. That practice in isolation might not be a bad thing. But over the years, many manufacturers have also concentrated on limiting the number of vendors to give them more bargaining leverage with them and squeeze them on costs. That increases dependence on suppliers while also increasing the riskiness of their operations. It has finally become fashionable to work with vendors or offshored factories in countries with low labor costs like China, Bangladesh, and Vietnam. Long transit times also increases business risk.
Now of course, like traders, top manager have every reason not to be terribly worried about long term risks. The prototype of the profitable but risky trading strategies that Nicholas Nassim Taleb likes to deride is that they work just fine on a day to day basis but blow up catastrophically periodically. And those blowups are predictable. But as long as they aren’t likely to happen every year or every other year, decision-makers have huge incentives that increase risk as long as the blow-up risk is not all that imminent (I am waiting for a quant to devise an optimal blow up metric as a covert trading strategy tool). So we should also regard the fact that business managers have acted exactly like reckless traders to be an expected outcome.
Tett also tells us that the business leaders aren’t quite sure what to do:
The good news, as Mr Wildgoose notes, is that corporate awareness of this problem is rising: many companies have started to develop mitigation strategies to diversify their supply chains.
The bad news, as the World Economic Forum recently warned, is that many of the vulnerabilities in these chains remain unaddressed and are poorly understood.
The lack of understanding doesn’t simply mean they’ll have a wee bit of study needed before they act. Bookstaber warned that tightly coupled systems were indeed difficult to contend with and that risk mitigation efforts that look logical typically make matters worse. His example was the emergency measures taken when a valve got stuck at Three Mile Island. The efforts to ameliorate the situation led to the meltdown. The only way to make a tightly coupled system less risky is to de-couple it, which will also make it less efficient on an ongoing basis. Management incentives will be to try to do everything but increase costs in a meaningful way, which means they are likely to try precisely the sort of interventions that Bookstaber warns against.
The only good news is that few manufacturers are so fundamental to the economy that we cannot afford them to take some hits from short-sighted decisions. But until a CEO bites the dust over this sort of error, we are unlikely to see a serious rethinking on this front. The fact that this sort of error is common means it will be treated as more legitimate than it ought to be.