Chaos Monkey: Unraveling the Vulnerabilities of Today's Global Supply Chain

Chaos Monkey is a software tool (developed by Netflix engineers) that tests the resilience and recoverability of their system. “Chaos Monkey is responsible for randomly terminating instances in production to ensure that engineers implement their services to be resilient to instance failures.”

There's no supply chain equivalent (and I don’t mean running simulations), but maybe the UPS workers’ “near strike” and the “near closure” of the I-95 can serve as one. 

The UPS workers’ strike was averted, but during the discussions leading up to it, I found the anticipated impact of such a strike rather surprising:

“A potential strike by 340,000 unionized workers at United Parcel Service Inc. threatens to unravel progress in tackling two of the US economy’s biggest hurdles in decades: inflation and supply-chain disruptions. In the absence of a labor deal by Aug. 1, a walkout by members of the International Brotherhood of Teamsters would snarl shipments of the 19 million packages that UPS moves daily in the US. It would also enable competitors such as FedEx Corp. to raise prices to help throttle some of the parcels that would pour into its networks.”

Maybe the fear was more because of what happened the last time UPS had a major strike. 

“Shippers have more alternatives for sending packages than they did during the last UPS strike in 1997. At that time, the parcel industry moved about 16 million packages a day and UPS accounted for 70% of that volume. About 80% of packages were sent by businesses to other businesses.”

And while UPS may not be as significant a player in 2023 as it was in 1997, the firm continues to dominate the supply chain landscape if we look at the numbers. Last year, the company handled an average of 24.3 million packages worldwide each day, which translates to a staggering 6.2 billion packages annually. Based on ShipMatrix’s data, it is responsible for roughly 28% of the 75 million daily packages handled by the industry. Taking all this into account, UPS’s role in the global supply chain, and especially in the U.S., where it's ranked as the top private delivery service, is indeed paramount. The advent of e-commerce, driving more than half of all parcel deliveries to homes, has also underscored the pivotal role UPS plays. So with the strike looming, firms began to prepare: 

“Companies will likely have already increased inventory of key parts and supplies to reduce their shipping need to ride out a possible UPS strike, said Mike Skordeles head of US economics for Truist Wealth. ‘Businesses will essentially treat it like a hurricane,’ Skordeles said. ‘They’re preparing ahead of this, such as front-loading orders and transportation for critical components.’”

Just to set the record straight, the U.S. and global supply chain does have some leeway, but not nearly enough when a considerable chunk of the economy is at stake. Picture the scenario: The threat of a strike is like an ominous cloud hanging over us just when the parcel industry has started to recover from the peak volumes of the pandemic era. This was when stay-at-home consumers, flush with stimulus checks, opted to purchase goods rather than services.

Take Amazon, UPS's largest customer, for example. Amazon seemed unperturbed by the potential disruption. Over time, they’ve been discreetly reducing their dependence on third-party carriers by building their own delivery network. It seems as if they’ve been crafting a contingency plan for times like these. So while the rest of the industry might be fraught with worry, Amazon stands poised, seemingly unshaken by the storm that might be on the horizon. So clearly, some supply chains are more resilient than others. 

Not the First Stress Test

But this is not the first time that we’ve heard of a “minor” disruption possibly affecting the entire economy over the last few weeks. A similar threat hung over us when the I-95 bridge collapsed. And while the crisis was once again averted, the anticipated impact was huge: 

“The collapse of an elevated section of Interstate 95 in Philadelphia on Sunday could pressure the price of goods as the disaster impacts East-Coast supply chains, according to Transportation Secretary Pete Buttigieg. Speaking during a press conference at the collapse site Monday, Buttigieg described the impact on supply chains in the region. ‘Part of what goes into the cost of everything that we pay for in the store is the cost of shipping and if a route is disrupted, or if it’s longer, or if trucks have to wait, that finds its way into the cost of goods,’ he said.”

Why?

The closure of the I-95, a crucial artery in America's transportation system, could have had considerable implications. This highway carries an average of 160,000 vehicles daily, about 8% of which are trucks, according to Buttigieg. Therefore, a shutdown would not only disrupt the movement of goods but also significantly impact a substantial segment of America’s daily GDP, which traverses this route daily. 

In such a scenario, trucks would have to take lengthier, more expensive routes to reach their destinations, essentially increasing their operational costs. This increased expenditure burdens the freight carriers but can also potentially trickle down to consumers in the form of higher prices for goods. 

The example of the I-95 closure illustrates the dependence of U.S. supply chains on the highway system, highlighting its vulnerability to unforeseen disruptions. While the U.S. has a vast geographical spread, it heavily relies on the flexibility and economic feasibility that the highway system offers for transporting goods, especially over long distances.

Again, such interruptions, whether due to a road closure or a larger-scale infrastructure failure, underscore the need for enhancing resilience in the supply chain. 

While a single closure like the I-95 might not have a significant impact on its own, it serves as a stark reminder of the potential ripple effect caused by widespread disruptions in our interconnected supply chains. It emphasizes the importance of preparedness and adaptability in maintaining the steady flow of goods, which is integral to the stability of our economy.

Supply Chain Fragility

These are, of course, only two examples. We saw similar doom and gloom predictions when the Evergiven was stuck in the Suez Canal. So, the issue isn't restricted to the U.S. but has a global outreach.

The fragility of the global supply chain has many factors, some of which have been discussed here before: 

Just-In-Time Delivery: Many businesses operate on a "just-in-time" delivery system to minimize inventory costs. This means they rely heavily on items being delivered exactly when needed and don't have much buffer if there's a delay. A strike or road closure like the ones discussed above could easily disrupt this finely tuned system.

Globalization: Many goods and components are produced overseas and have to be shipped to the U.S., adding another layer of complexity and potential points of failure to the supply chain.

Lack of Resiliency: The global supply chain has been optimized for efficiency and cost-effectiveness but not necessarily for resilience. When disruptions occur, such as a pandemic, a natural disaster, or a strike, the system may struggle to adapt.

But these issues are not new. The notion of supply chain resilience has been top of mind for many firms since the pandemic. And some firms have been trying to address them. I would argue that some issues are structural, driven by the structure of the supply chain, rather than an action any single firm can take. 

Supply Chain Fragmentation

One of the most interesting recent papers on this topic, co-authored by Nikolay Osadchiy, Maximiliano Udenio, and Vishal Gaur, has shown that over the last 20 years, the global supply chain has become increasingly fragmented. 

The paper uses the FactSet/Revere data set of supply linkages to study the structure of the global production network during 2003-2018. At the individual firm level, the paper shows that the in-degree of firms increased at 2.7% per year (so firms source from more firms) and out-degree increased at 2.1% per year on average (so firms sell to more firms), showing evidence consistent with multi-sourcing and customer diversification.

The interesting aspect, however, is the network-level analysis. At the network level, the paper proposes a measure of integration and fragmentation that is based on a notion of communities in the production network. A community represents a set of firms that trade mostly connected with each other.

The paper's findings indicate that while firms in the network have increased their connections over time, the overall network has become more fragmented. The number of distinct communities within the network has risen from 34 to 60. Additionally, the concentration of these communities, as measured by the Herfindahl-Hirschman Index (HHI), has declined from 0.065 to 0.045.

What does this mean? Think of the global production network as a professional networking event. When the event starts, there are 34 distinct groups of people, each representing a community of firms with strong business ties. As the event progresses, people start to network and form new connections, similar to firms expanding their business relationships.

However, instead of these connections leading to a more unified network, the opposite happens. The number of distinct groups or "communities" increases from 34 to 60. This is like saying that even though individuals at the event are networking more, they are forming new, separate groups rather than merging into larger ones.

The Herfindahl-Hirschman Index (HHI) in this context measures how evenly distributed the firms are across these communities. A higher HHI would mean a few large communities dominate the network, while a lower HHI indicates that firms are more evenly spread out across a larger number of communities. The decline in HHI from 0.065 to 0.045 suggests that firms are spreading out more into smaller, distinct communities.

So, in essence, the paper shows that even though companies are forming more connections, the business network is becoming more fragmented, with firms spreading out into more and smaller communities.

The final result is that this fragmentation manifested in both the extended supply and customer networks, resulting in the emergence of firms with abnormally high betweenness centrality, which sit at the interfaces of communities.

What are the implications for our discussion of resilience:

“In a fragmented network, firms that link communities can be especially important. On the supplier side, such firms provide access to components and raw materials that are not typically used in a focal firm’s community and may be hard to substitute. On the customer side, such firms link the focal firm to the sources of demand from other communities which may be less predictable. … Such nodes may be more important for supply and demand risk management than the other nodes in a community.”

While roads and logistics companies are not discussed in this paper (the focus is on production firms), one can argue that I-95 and UPS are exactly the type of arcs that connect these communities. 

I know this may sound counterintuitive, but with more fragmentation, we shouldn't have high dependency on any single firm or road unless they are the interconnected arcs. So the additional fragmentation created makes the two following phenomena even more important: 

Interconnectedness of Supply Chain: The supply chain in the U.S. is a complex, interconnected web. A delay in one area can have a domino effect, causing delays elsewhere. For example, a delay in delivering raw materials could delay manufacturing, which could delay delivery to retailers, and so on.

Centralization: A lot of logistics is centralized in a few key companies, like UPS, FedEx, and Amazon. This means if one of these companies experiences a disruption, it can have a big impact. It's akin to having all your eggs in one basket.

The issue of supply chain fragility is a complex one with no easy solutions. It involves balancing efficiency and cost-effectiveness with resilience and flexibility. It's a challenge that businesses, policymakers, and labor groups continue to grapple with.

But these two recent examples should serve as warning signs. While there is a lot of investment in the new supply chain for semiconductors and batteries, it’s important to make sure we safeguard the supply chain around it, like transportation, roads, and logistics firms. we need to make sure we don’t over-rely on any single firm or country. Fragmentation is going to be even more pronounced over the next few years. We need to make sure that we don’t overly depend on a few firms that connect these emerging networks.