Technological Complexity: The Next Black Swan?

A comment in Friday’s Financial Times by John Thornhill caught my eye, although not for the reason you might think. His piece was mostly a warning about how our overreliance on technology could spark the next global crisis. The article elaborated on a number of cyberwar and cybersecurity risks that make our modern society increasingly vulnerable to any number of attacks. And sure enough – would you believe it! – just two days later it appears that some Russians successfully hacked into the U.S. Treasury and Commerce Departments, and possibly also a number of national-security related agencies. And this not even a week after warnings from the NSA and FireEye, the US cybersecurity company that was itself infiltrated.

Newsworthy? Indeed. A potentially serious problem? Very possibly. But truth be told, I tend to lose interest whenever the conversation shifts to an “us versus them” narrative. One does not need to be a historian to understand that when it comes to geopolitical conflicts…well, let’s just say that the U.S. government does not always (or even often) occupy the moral high ground. What does interest me about Thornhill’s editorial is the more general argument he makes about technology, complexity, and vulnerability. Even he states that it is not only state-on-state cyberconflict that is alarming, but the internet’s “systemic instability” due to its “unnervingly flimsy” governance.

Thanks to Naseem Taleb, we have come to understand a Black Swan as an entirely unpredictable or unforeseen event that often has devastating national or global effects. Yet as Thornhill correctly points out, the risk of a future crisis due to a tech meltdown of some sort, while potentially unforeseen post hoc, is most certainly foreseeable. Something unforeseen yet eminently foreseeable (the Covid-19 pandemic, anyone?) is what Nouriel Roubini and perhaps others refer to as “White Swans.” And if we consider how the research of archaeologist Joseph Tainter relates to the aforementioned threat, the latter seems a more accurate characterization.

In his best-known work, The Collapse of Complex Societies, Tainter identifies more than a dozen cases of ancient civilizations that underwent a rapid collapse. What might that have to do with technology and contemporary crises? In his book he develops an innovative model (focusing primarily on the Romans, the Maya, and the Chaco culture) to support his claim that what the cases of civilizational collapse all have in common is that they all succumbed to complexity. Too much complexity, in other words, is what invariably did them in.

The degree of complexity of these ancient societies was of course, at least in some respects, orders of magnitude less than of contemporary global society. But that distinction is unimportant, since it is the standards of the time that matter. Tainter’s point is that whenever societies succeed, they grow. And as they grow, they necessarily become complex. Towns, for example, require more developed infrastructure, communications networks, systems of laws, and administration – as well as much else. And cities much more than towns, and so on. The problem (see Diagram) is that while there are immense social gains from increased complexity – much having to do with efficiency improvements – the marginal gains diminish over time as society grows yet more complex. 

Another way of saying this is that once people’s critical needs are met – and here there is no need to get distracted with what exactly these are – continued benefits from greater complexity progressively diminish in importance. Which kind of makes sense. The opposite, unfortunately, happens with costs. The more complex a society, the more uncertainty and unpredictability related to bottlenecks, feedback loops, and increasingly unanticipated challenges. Because of this, according to Tainter, problems grow increasingly more difficult to solve. As a result, exponentially rising costs over time become inexorable. 

The complexity threshold depicted on the diagram cannot literally be observed. Since social complexity – never mind its benefits and costs – would be devilishly difficult to measure, the critical limit shown is a conceptual expedient. Nor is it even necessarily “critical,” since it is conceivable that a society could function beyond the threshold – that is, in the range where the marginal social cost of complexity exceeded the marginal benefit. In fact, I would submit that it is quite likely that the United States – and probably a good number of other advanced economies – have already exceeded the complexity threshold.

What it might mean, for example, is that modern-day bureaucracies are too inconvenient (when not indeed harmful) than they are worth. The same could be said of the global financial system, which primarily served a useful, even indispensable, function not too long ago. Many other examples come to mind – food production and distribution, the electrical grid, and the source of Thornhill’s anxiety – the internet. In every case, growing complexity creates new, more complex and potentially catastrophic problems that require ever-increasing layers of society and classes of workers to solve.

I believe it is a compelling thesis, and ancient civilizations might serve as useful cautionary tales, if we cared to learn from them. You might object that historians know history – and precisely for this reason lack the chops to prognosticate future events. That may be so, but given the dismal forecasting record for which economists are well known, it might be time to give other specialists a chance to predict the next “Swan” – whatever its color.