As the pause on US reciprocal tariffs reaches its end and practitioners think about what comes next, they might do well to remember the story of car maker Ford during the Covid-19 pandemic.
In the summer of 2021, amid the supply-chain disruption of Covid lockdowns, the car maker halted manufacturing at several of its US plants because it was running out of silicon chips.
Ford uses lots. Each F-150 truck — the manufacturer’s most popular model — contains about 1,400 semiconductors. Chips control fuel injection, the air conditioning, airbag control, and so on.
Ford’s experience was commonplace across the industry. Output slumped by millions of units a year into 2023. The shortages, meanwhile, led to slowdowns in 168 other types of businesses, according to research by Goldman Sachs at the time.
The pandemic showed that supply shocks can propagate through the economy in surprising ways. Some experts suggest today’s yo-yo tariffs could cause something similar again.
Complexity
Table of Contents
Specialists in so-called complexity economics have warned of this sort of thing for years.
Leaders in the field such as Doyne Farmer look at economies as complex networks, and draw on the ideas of chaos theory to model interactions between the players.
The economy is to all intents and purposes a chaotic system
Robert Hillman, Neuron Capital
During the UK’s first Covid lockdown, Farmer and colleagues at the University of Oxford wrote a paper that suggested opening certain industries before others would cause economic output to shrink rather than grow. To lift restrictions on manufacturing, for example, risked a surge in demand for critical “input” items that would exhaust supply and create bottlenecks in non-manufacturing sectors.
“First-order shocks can be translated into overall impacts in highly nonlinear ways,” Farmer and his fellow researchers wrote. “We even find cases where smaller aggregate shocks can lead to larger economic impacts as a result of unbalanced supply.”
More recently, in May, a group of researchers in Austria ran a computer simulation of global trade and found that as many as 17 fewer container ships a day could dock at the ports of Long Beach and Los Angeles if tariffs were fixed at levels that effectively cut all US-China trade.
At the same time, a subsequent reduction could trigger a rebound phase of “catch-up” trading that would clog the port complex with an additional 19 ships daily, the simulation showed.
“The situation is similar to what happened during the pandemic,” the Austrian team wrote in a research note. “Chinese ports closed during the lockdowns and afterwards supply chains couldn’t accommodate the resurgence in demand, leading to shortages.”
Much of what matters in such shocks, it seems, is hard to foresee.
People change their buying behaviour, or change jobs, or retire. Companies change suppliers. Traders move to new shipping routes. Businesses adopt new technology, such as artificial intelligence.
Chip shortages in 2021 resulted from factory closures but also from a surge in demand as pandemic home workers rushed to buy laptops and other digital devices.
The complexity economists say this makes the task of modelling supply shocks more like forecasting climate change or, for that matter, how a virus will spread. Behavioural adaptations alter the course of the future as it plays out.
“The economy is to all intents and purposes a chaotic system,” writes Robert Hillman, founder of quant research firm Neuron Capital, in a recent blogpost. “Small changes can lead to large differences in outcomes.” This is the so-called ‘butterfly effect’ of chaos theory at work.
Best efforts
However hard predicting the fallout from supply disruptions may be, economists and market practitioners may have to do their best.
The Financial Times reported in June that European ports were experiencing the longest delays since the pandemic. Logistics companies blame US tariffs, which have forced shippers to change trade routing.
What, then, are the options? Many of the complexity economists favour the use of so-called agent-based models. The models run virtual simulations using thousands, even millions, of digital avatars coded to reflect the actions of companies or individuals.
Practitioners might argue about how well the models can or can’t forecast the future. But, either way, they provide a valuable laboratory to test scenarios. Economists can model behaviour such as truck drivers moving to jobs as parcel delivery drivers, which turned out to be a real-world problem in the UK during the pandemic, Hillman says.
Peter Klimek, director of the Supply Chain Intelligence Institute Austria, which ran the trade and shipping simulations – using an agent-based model – says the models offer a “perfect test bed” to evaluate the ripple effects of tariffs.
Of course, their users should exercise humility, he says. Economists know little today about how far each producer in a value chain will pass along the cost of tariffs to the next. That said, as more data becomes available, such details will become clearer, and the models more reliable.
Tariff wars will, no doubt, bring more surprises. Agent-based models may offer a way to explore what those surprises could be.
