Everything is a Network: Trade

1992: Where we begin

If we are thinking deeply about network structure, it makes sense to look at how things have evolved over time. Fortunately, we have reasonable data for this going back to 1992. In the relatively simplistic network illustrated here, aiming to capture the majority of 'meaningful' activity, all goods trade flows (services are not included in this analysis) over the size of $20bn US current dollars between countries are illustrated, with those flows as the network 'edges' and trading nations as the network 'nodes'. Edge thickness and node size correspond to bilateral flow and total trade amounts respectively. Node colours are determined by the Louvain clustering algorithm, which groups countries into communities based on their trade interactions. The logic here is simple: total trade within the group is high, and total trade between groups is relatively low. Countries that trade heavily with each other are likely to end up in the same community.

So, 1992. First - note how small the network actually is. Second - note the primary players: the US, Germany and Japan. Third - look how neatly communities are formed. Trade in the early 90s is heavily regional with the three primary players acting as bridges between communities.

Network scientists will tell you that a 'hub-and-spoke' system is one in which the majority of nodes link directly to a few central nodes. It is clear that we have this, but it is a dynamic materialising at a regional level.

2000: The US at centre of the universe

Jump forward to the new millennium and note how concentrated trade flows to and from the US have become. The network itself has grown with more trading activity, but that has come with an economic rebalancing: China's (and other Asian counterparts) growth as a trading nation at the expense of Japan, Germany's decline relative to the likes of Italy, France and Spain.

We also see how only two communities are now formed - freshly Euro'd Europe vs the US and the rest of the trading world. Even then, it is clear of the reliance of all major trading partners on the US. Hub-and-spoke? Check.

2008: GFC, among other things

A further 8 years and the trading system has expanded remarkably both in participation and volume terms. Most prominent is the surge in Asian nations and the rise in the 'BRICs' as key players. China, at this point, is as critical a global trading partner as Germany, and a bigger Asia hub than Japan. And it has done so both by expanding bilateral trade with the US and by its heavy involvement with the new additions to the trade network.

But note how much more fragmented this system has become. Already by 2008 it is clear that regional trading divides are back in a more complex form. Western Europe clustered together, Germany as a gateway for growing Central and Eastern European nations, and the US as an Americas hub.

2018 - 2024: The consolidation

In the decade that followed, fewer new players entered the 'meaingful' trade network we constructed, but existing growth patterns became more entrenched. Regional disparities between a growing Asia and a consolidating Europe are most obvious, and the flip from a clear multipolar trading system dominated by the US, China and Germany to one that, at minimum appears bipolar (US vs China) and at maximum appears multipolar on a much broader basis.

One thing that should be obvious is how dynamic this network is. The dramatic shifts in the clustering algorithm from year-to-year illustrate it more clearly than can be justified in words. But it makes sense. Events including 2018's US-China trade war fragmented a system increasingly defined by the trading flow between those two nations alone. COVID-led breakdowns in supply chains and (primarily Western) policy around the Russian invasion of Ukraine further have further forced a re-routing in trading activity.

Hub-and-spoke? It just isn't that simple.