The Global Supply Chain Pressure Index and the “Smooth Lines” Fallacy
What a fed index capturing the state of 'the supply chain' might be getting wrong
Variations of the sentence “the pandemic and the blockage of the Suez canal by the Ever Given brought the world’s supply chains to light” are, generally, annoying, although there is admittedly some truth to them. Shipping was not particularly on the news agenda several years ago. Running a search for news articles in 2019 that contain the phrase “Global Shipping” pulls up 1830 results. Running the same search for subsequent years sees that figure more than double 4800 in 2021, and rise to 5260 for 2023. In the last few years, Bloomberg have hired dedicated logistics reporters, and other news outlets are starting to consider supply chains as its own beat rather than just a dull branch of business news.
Economists are taking note too. The multiple overlapping crises of pandemic, climate breakdown and war have brought about a kind of reckoning with what the economy is actually composed of, and how mainstream economists pick up the constant fight against inflation. In 2021, Isabella Weber contended that crises and commodity shocks can accelerate price gouging on foundational commodities (gas, oil, wheat) that form the basic building blocks of other commodities, and that these increased prices lead to greater inflation further down the supply chain (she called this “seller’s inflation”, and crucially, it is enabled in situations of supply chain bottlenecks). Her suggestion to combat inflation not solely through monetary policy but by imposing strategic price caps on critical things was roundly laughed at in 2021, but has since rapidly become adopted. Similarly, in 2022 Zoltan Poszar (of the now-collapsed Credit Suisse) wrote in the introduction to a highly influential note to clients that “commodities are real resources (food, energy, metals), and resource inequality cannot be addressed by QE [quantitative easing]”. Basically, economies (and the economists that theorise them) started coming down to earth again, and in many cases, that earth turned out to be the sea.
Amidst all this, the US federal reserve launched the Global Supply Chain Pressure Index (GSCPI), pulling data from manufacturing and transportation sources to get a single figure on how much ‘pressure’ the supply chain is under. It’s updated monthly, and uses standard deviation: if the line goes up, then things are pressured, and if it goes down, things are easing up. Like many indexes, the GSCPI is primarily an index of indexes. The components of the index are:
Freight cost data, which is pulled from other indexes. This includes the Baltic Dry Index, which is pretty industry standard, and the Harpex index – which isn’t. It also takes into account airfreight and trucking indexes, including US Bureau of Labor statistics.
“Purchase Manager Indexes”, which are just surveys of various senior executives, ostensibly to get a sense of the temperature of the market. This would cover things like backlogs and current actual delivery times.
For a supply chain index that’s “global”, there are some pretty big omissions here. Trucking data, for example – the leading form of internal commodity transportation in much of the world – seems absent, perhaps in part because reliable global trucking data simply doesn’t exist. It shouldn’t come as much of a surprise that the ambition of the “Global Supply Chain Pressure Index” is something more like “ease with which commodities are moving towards or away from the US”.
In an instructional video about the uses of the GSCPI, the Fed state that:
Although it claims to not be a forecasting tool, it is based on several private indexes that are essentially marketed as forecasting tools, and I don’t think the Fed has any power to stop people writing articles with advice on “how you can use it to predict and manage supply chain pressures.” It describes itself as more of a “nowcasting” tool – using recent data to estimate, with more confidence, economic indicators that tend to take a while to come out, like GDP or inflation. But the function of nowcasting and forecasting tools is, for all intents and purposes, pretty much the same: to build up a temporal picture that guides people with capital on what to do, right now, with that capital. As our friend Gary Zhang notes, “You could say that the financial market to which we are habituated is largely a game of self-fulfilling prophecy, an ourobouros narrative of its own making and unmaking. Its players compete to predict a future which is – by the laws of their own models – unpredictable. As sociologist David MacKenzie puts it, the tools of contemporary finance form ‘an engine, not a camera’.”
In finance, for all intents and purposes, the horsepower of the engine is basically determined by how powerful and reliable everyone thinks it is. Accordingly, the Fed take time to validate the index through noting the extent to which it correlates with past supply chain shocks and crises:
“For instance, the index surged in 2011 following the Sendai earthquake and resulting tsunami. Japanese car manufacturing and its exports to foreign markets were severely impacted. The same year, flooding around Bangkok, Thailand, disrupted global supply chains in the automotive and electronics sectors, particularly hard drives, leading to shortages among global computer manufacturers. The index rose again in 2018 and 2019 following United States-China trade disputes, inciting several large manufacturers and retailers to revise their manufacturing and procurement strategies, but mostly absorbing higher costs.”
This seems to be less of a method of studying “inflationary events” guided by the GSCPI, and more of an exercise in attempting to validate the veracity of the index itself. Testing other shocks on it though, it doesn’t always work out: in August 2005 Hurricane Katrina knocked the port of South Louisiana – the US’ 5th largest –out of use. Gulf Coast oil production fell by 88%, also causing fuel prices to spike. At a time you’d expect the index to rocket, it barely moves.
In their pitch for the GSCPI, the video explainer also notes that “conventional measures tend to focus on specific elements of the global supply chain rather than as one interconnected whole”. What we’d caution here is that logistical systems are really not as much of a harmonious ‘interconnected whole’ as people think, but are instead characterised by the management of all kinds of strategic frictions, stoppages and artificial scarcities. These are decoupled from landside production (but highly reminiscent of their often counterintuitive profit-motive logics).
A criticism of the supply chain index would be an extension of the argument we made in a piece we wrote last year, which is to assume that circulation is a logical extension of production and transportation systems exist to diligently move things smoothly from A to B. Because of how it conflates data on production and circulation, the GSCPI draws a field of equivalence across those things, assuming that transportation acts as a subsidiary to the best interests of manufacturing.
Maybe we could call this something like the “diagram of smooth lines” fallacy. If you search for ‘logistics’ in google images you inevitably get stuff like this:
When the reality it is something more like this:
That last picture shows oil ships waiting to unload crude in Singapore in June 2020, during a period of contango: when futures (prices for forward delivery dates) are higher than nearer delivery dates, meaning it becomes more profitable for oil movers to simply tell the ships to stop for a while and become floating storage facilities. And what senior executive is going to tell market researchers that they are doing that – the same market researchers that compile the surveys that go into the GSCPI – when doing so would risk dropping the futures price of the oil they’re betting on.
This example poses a problem for the assumptions of the architects of the GSCPI and critical logistics scholars alike: that disruptions, decelerations and bottlenecks in logistical systems rise out of either design inefficiencies (bad planning) or social disruptions (strikes and blockades), and that this negatively impacts profit and accumulation. Micheal Simpson explored this in a 2018 paper called ‘The annihilation of time by space: Pluri-temporal strategies of capitalist circulation’, and I’m going to end on it:
“These explanations for delays or gridlock in supply chains are not incorrect, but taken together what they miss are the instances in which slowing down circulation occurs intentionally and in the interest of capital accumulation. In other words, slow circulation is not always a result of accidents, oversights, or disruptions – rather, there are moments at which decelerating speed is an outcome consistent with the internal logic of capitalism.”
Notes:
Interestingly, the World Bank also released a very similar index focusing more solely on shipping, called the ‘Global Supply Chain Stress Index’, which it seems never really took off.
If you’re enjoying this, tell your friends! Also, some personal news: Miriam has wrapped up the first year of her PhD looking at how ‘green’ regulations are accelerating the datafication of maritime capitalism, the move of shipping functions from ship to shore, and where this leaves crews. Jacob got funding to also go and do a PhD at LSE, looking at ongoing speculation and securitisation of shipping lanes in the North Pole that don’t exist yet, but are looking viable because of melting sea ice, so expect a lot of writing on that over the coming years.