In his recent book End Times, Peter Turchin tries to lay out his story about social, political, and economic change for a popular audience. The book is given added immediacy by its publisher’s subtitle: “Elites, Counter-Elites, and the Path of Political Disintegration”. I know Turchin’s work quite well, and I know his underlying model well enough, and I think that End Times fails in this objective. I’ll come back to why this is later on, but first a bit of background on the underlying model and its intellectual history.
Turchin was an evolutionary biologist by background, and his early hypothesis, which brought him into the area of history and social sciences, was that you might see similar patterns in population pressures in human societies that you see when looking at animal groups. It’s not an unreasonable hypothesis.
(Photo: Andrew Curry. CC NC-BY-SA 4.0)
In his first book which detailed the model, Secular Cycles (2009), written with Sergey A. Nefodov, they described these as ‘structural-demographic cycles’, which drew on the work of the historical sociologist Jack Goldstone. They explored this idea across eight time periods, all pre-industrial, across four geographies. They included, for example, both Plantagenet England and Tudor-Stuart England, and Muscovite and Romanov Russia.
Secular cycles
The model in Secular Cycles is explained fairly clearly (p. 33), and I’m going to summarise it here.
The ‘cycle’ includes a long ‘integrative’ phase and a long ‘disintegrative’ phase. The integrative phase falls into two parts: an expansion phase, which involves growth, and a ‘stagflation’ or ‘compression’ phase. The disintegrative phase is messier, but has a ‘crisis’ phase involving state breakdown and a depression phase, which can also just be a stage of ‘intercycle’ stagnation before the next growth phase starts.
(Source: Peter Turchin and Sergey A. Nefodov. ‘Secular Cycles’, p.33)
Long? We’re talking generational timescales—this whole cycle can take between 100 to 300 years, depending on how easy it is to replace your elites.
Wages down, assets up
The narrative of the original model goes like this:
- Initially, workers (or peasants) are making a reasonable living. But that means that population increases—more children are born, people are better fed and less likely to die. So the population starts to approach carrying capacity.
- As population reaches carrying capacity, there are more people (so wages are squeezed) but increased demand for food and land)—so wages are doubly squeezed.
- But at the same time, the asset values of landowners, and their returns from food production, both increase.
- This shifts demand patterns towards things that the wealthy want to pay for, so you also see new industries start to emerge.
- You also see increasing numbers of the elite, and increasing competition for elite position, which creates instability. In this ‘stagflation’ phase inequality increases at all levels of society; extraction intensifies.
- One of the effects of this is that the state becomes more unstable. It finds it harder to raise money, for example.
- Typically there are food riots and patchy rebellions somewhere around here. Some people and families will fall out of the elite because they can’t maintain the level of expense that is required. But the outcome as the system moves into crisis phase depends on the cohesion of the elites.
- The elites can maintain the immiserised status of the immiserised wage-earners for a long time, provided the elites don’t split. And: there are patterns here.
- And every so often, elites do split, and members of the elite then choose to ally with others in society, typically because they fear that outcomes will be worse if they don’t.
So, for example, when the venture capitalist Nick Hanauer warned other wealthy people that “the pitchforks are coming”, he is an early sign of a potentially fracturing elite. Climate change may also have the same effect.
Explanatory power
This model has reasonable explanatory power, but (since it was tested with pre-Industrial Revolution data and narratives) it left open the question of whether the model would work in societies that were better able to cushion themselves against adversity. Later work identified that it did still apply.
And at some point, Turchin started referring to this application of statistical analysis of social data and trends as “cliodynamics”. This is how he defines cliodynamics in The End Times:
It uses the methods of data science, treating the historical record, compiled by generations of historians, as Big Data. It employs mathematical models to trace the intricate web of in-tractions between the different “moving parts” of the complex social systems that are our societies. Most importantly, cliodynamics uses the scientific method, in which alternative theories are subject to empirical tests with data (p.xii).
Modern times
Turchin’s work on both the US Antebellum period and 20th century America also suggested that the model has reasonable value in terms of projecting increases or declines in levels of social tension. (The relevant article is here).
Even if carrying capacity is not always a constraint in more complex societies, institutional and regime factors can have the same effect. For example: In the 1970s, in the UK and the US, wages reached their highest share of national income relative to profits—and we’re still living in the aftermath of the political and economic strategies deployed by capital to bring wages back down again.
That 2013 paper has this version of Structural Demographics in it, which at least explains some of the relationships and data points that sit within the structural-demographic model.
(Peter Turchin, ‘Modeling social pressures towards instability’, Cliodynamics, 2013.)
Different systems
Associated with this work is an institute (The Cliodynamics Institute) and an open access journal, and a large database that supports their associated modelling work.
But in short, the reason that I think that the model is interesting is that it combines two different systems that matter in human societies. The first is about economics, there are familiar long wave models (such as Kondratiev) that look at this in detail. (I wrote about these here).and are, if anything, becoming more important. The second is about status—humans are status-driven creatures, as we learn in different ways from Rene Girard and Michael Marmot. And on their own neither system really explains our current politics properly.
But, in End Times, it is hard to get a handle on this underlying model. Peter Turchin takes his thinking about long waves of political, social, and economic change, or at least some of it, and packages it up, as if it is one of those Big Airport Books That Explains What Is Going On In The World. (Or more specifically, In America).
In brief: a first section talks about elites and their over-production in the US, how that leads to crisis, and tries to contextualise that. A second section, also about the US, looks at the elements that create instability. A third section explores crisis, its aftermath, and possible futures. And an Appendix contains some background to the Cliodynamics approach (this might be subtitled, “I answer my critics and deal with some misconceptions.)
(Photo: Andrew Curry. CC NC-BY-SA 4.0)
Mistakes and traps
In his enthusiasm to popularise, I think Turchin has made some mistakes and fallen into some traps. The mistake is that —on my reading—I don’t see anywhere in the book where he explains the underlying model in any coherent form, narratively or diagrammatically. At worse, this could have gone into one of the three chapters in the Appendix section.
I thought I might be being careless in my reading here, but this also frustrated Will Davies in his New Statesman review of the book. If you’re not familiar with the history of Turchin’s work, there seems to be less here than meets the eye.
The first trap is about claiming “prediction” rather than “projection” or “pattern”, even while putting cavets around it. It is noticeable to me that Carlota Perez, who has a similarly useful model about the interaction between finance and technology innovation, always insists that her model is not a law, and can only really be described as a heuristic.
Psychohistory
This sometimes reminds me of Isaac Asimov’s “psychohistory”, and it’s hard to read the rationale for the (let’s face it, invented neologism) of “cliodynamics” without being reminded of Asimov. Turchin does discuss Asimov and psychohistory in the chapter on “the new science of history.”[1]
And as it happens, in one of the chapters in the Appendix, Turchin has a reasonably honest assessment of the limits of prediction that is hard to disagree with:
What Asimov did not know is that even if you can ignore such things as individual free will, you still run (up) against very strict limits to predictability. When components of a complex system interact nonlinearly, the resulting dynamics become effectively unpredictable, even if they are entirely deterministic… But this limitation on predictability is, paradoxically, a source of optimism. It means that human individuals are not as powerless as Asimov imagined them. Exercising one’s free will can have major consequences at the macro level, just like a butterfly fluttering its wings can affect the course of a hurricane. (pp. 250-251)
Underlying patterns
The second trap, and perhaps this was at the suggestion of an editor or a publisher, is the personal vignettes or anecdotes that litter the book. Perhaps they are supposed to humanise the story. What they actually do is to trivialise the underlying patterns in the model. (And if you need to talk about the travails of blue collar America, there are plenty of credible sources that have already brought these into the light in ways that combine insight and data, as there are, also, for the super-rich.) Even if most hypotheses start life as anecdote, the anecdotes don’t help the story.
The third trap is to use metaphors that don’t help to transmit the story. The mechanism by which the asset-based wealth of the elites increases as real wage levels decline is referred to in the book as “the wealth pump”. But there’s almost nothing pump-like about the mechanism, which is actually just a form of extraction. It’s more like a rake-off.
In short, I think he has over-claimed and over-simplified. I was struck by this while reading a couple of reviews. The economist Branko Milanovic reviewed it on his newsletter, and took away from the book that it was all about inequality. This is a reasonable proxy for falling real wages, just about, but it misses the parts of the argument about cycles or elite behaviour and income.
‘The science of history’
In the New Statesman, William Davies hated it (metered paywall). “Scathing” would be the right word here:
Rather like Deep Thought, the super-computer in The Hitchhiker’s Guide to the Galaxy, CrisisDB and MPF (Turchin’s model appear to have been fed the most fundamental, existential questions of politics, and answered with a banality: moderate redistribution or barbarism)… This is TED-Talk-lit that so vastly over-promises that the under-delivery is baked in from the start. The insistence that nothing can escape statistical model-building is ultimately a power grab, a determination to own and to control, not just the future but also adjacent scholarly disciplines.
I think Turchin also overclaims by claiming, as he does in an appendix, that this is “the new science of history”, even though the word “Cliodynamics” also hints at this. When Turchin was profiled in The Atlantic a while ago, he felt the need to write a long response on the Cliodynamics blog identifying ways in which he felt he had been misrepresented. The blog post was headlined, in quotes, “The Mad Prophet of Connecticut”. He probably had been misrepresented, a little bit, but when you set yourself up in this way misrepresentation is inevitable. And, for heavens’ sake, The Atlantic is one of the last bastions of the great American reporting tradition, with fact-checkers and proper editors and everything.
Historians are sometime suspicious of Turchin’s work. This may be also because of the claims of “the new science”, etc. In the blog post mentioned immediately above, he’s more circumspect:
We need academic historians to use their expertise to continue amassing the knowledge about different aspects of past societies… History can exist (and has existed) without Cliodynamics, but Cliodynamics cannot exist without History.
Hanging together
In his review, Will Davies points us towards other long wave theorists, like Arrighi, Piketty, and Eric Hobsbawm, who might be thought to cover similar ground. I’d say that Turchin’s model complements these.
Just as Perez adds a building block to previous models of long-run technology innovation (in her case the switch half-way through from finance to production capital, and the crash that is necessary to orchestrate this), I think Turchin adds a valuable building block about the way elites mostly hang together, even while they are busy reconstituting themselves.
It’s just a shame that End Times manages to bury this in the desire to be popular and contemporary.
Footnote
[1] Psychohistory is defined like this in Wikipedia: “Psychohistory is a fictional science in Isaac Asimov‘s Foundationuniverse which combines history, sociology, and mathematical statistics to make general predictions about the future behavior of very large groups of people, such as the Galactic Empire.”
—-
A version of this article is also published on my Just Two Things Newsletter.
Teaser photo credit: Four Horsemen of the Apocalypse, woodcut print from the Apocalypse of Albrecht Dürer (1497–1498), Staatliche Kunsthalle Karlsruhe. By http://www.wga.hu/html/d/durer/2/12/2apocaly/index.html, Public Domain, https://commons.wikimedia.org/w/index.php?curid=95226
