Revisiting a 2017 Prediction

Making structural-demographic theory more quantitative

As readers of this Substack now, back in 2010 I published prediction that the United States will experience a period of heightened social and political instability during the 2020s (Political instability may be a contributor in the coming decade). Some years later, several people challenged me to make this prediction more quantitative, because quantitative predictions are a much harder test for a theory.

In response, nine years ago, in January 2017, I published a more precise prediction on my blog:

A Quantitative Prediction for Political Violence in the 2020s

Following a suggestion that David Andolfatto (@dandolfa) made on Twitter, I used the quantitative metric that was established in my 2012 JPR article — the number of instability events per 5 years. Here’s the relevant figure from the article:

For the 2020s, I proposed that the expectation from the theory is that there will be more than 100 events per 5 years.

Note that this was not a prophecy, but a scientific prediction (see Scientific Prediction ≠ Prophecy). What this means is that there is no certainty that the prediction will be proven correct in the future — it will only happen if the theory is correct and no unexpected event happens that could derail the prediction. The purpose of making scientific predictions is not to know the future, but to empirically test theories.

Last Fall I published a series of posts on the structural-demographic state of America and one of them updated the data in the US political violence database (see The Structural-Demographic State of America IV: Political Violence Trends). As the updated data series showed, my 2017 prediction was fulfilled, as the latest quinquennium (5-year period) just exceeded 100 events:

Source: USPVDB

This result, therefore, provides additional support for the structural-demographic theory (SDT) in general, and to its specific application to the contemporary USA. Although at this point it’s clear pretty much to everybody that America is in crisis; indeed, in a revolutionary crisis. Still, making the theory more quantitative is where we need to go. Furthermore, the future always brings surprises. No scientific theory is absolutely correct, and this is especially true about a young field, like Cliodynamics. An interesting new observation that I made after revisiting the 2017 prediction is that the current set of instability events is dominated by individual-instigated political violence. I discuss possible reasons for this development in this post: The Structural-Demographic State of America V: Why Individual-Instigated Political Violence Dominates the Current Peak.

As a final note, my research team is currently working on replicating the 2010 prediction. Our plan is to adapt the general SDT to twelve new case-studies, each involving a contemporary state. Then we will publish predictions from these country-specific models for the next decades. I am sure there will be many surprises, and my expectation is that the SDT will be substantially improved in the process.