An incomplete model

Scott Alexander at his new substack site, Astral Codex Ten, solicited book reviews from his readers and has been posting the quality ones.

I was particularly interested in The Collapse Of Complex Societies: Finalist #13 in the Book Review Contest.  I am interested in history, economic development and in societal development, particularly in terms of technology and productivity.  The review is of The Collapse of Complex Societies by Joseph Tainter.  

I have had this book in a stack waiting to be read for at least five years.  It looked very interesting with a lot of interesting quantified history.  And the thesis seemed somewhat plausible but about which I had doubts.  Mostly, I felt that the quality of the data was not sufficient to bear the weight of the conclusions.  

The review starts off - 

Joseph Tainter’s explanation for why complex societies collapse in one sentence: the collapse of a society is a response to declining marginal returns on investment in complexity.

This was a crystalizing summary and captured another dimension of my skepticism.  Yes, complex societies can continue to grow in productivity until the marginal returns on investment in complexity start to collapse.

That is true to a point but does not take into account S-curves and S-curve transitions.  If you look at businesses in sectors, it is easy to see that many companies can become quite successful and profitable at their founding and during their early development.  Eventually, though, other competitors enter the market, attracted by the first company's healthy profit margins.  The second entrant (or third or fourth) hits on better processes or technology and begins to gain market share at the expense of the first company.  They are all climbing the S-curve.

Eventually they all hit a commoditization ceiling where survival is dependent on a very strong and enduring brand or regulatory capture or the capacity to constantly but cheaply improve technology and processes.  Playing in commodity markets is reasonably wretched.  

Click to enlarge.

Tainter stops there.  Reaching the top of the productivity S-curve is perilous given the returns on investment are so slim and business (or state) is so complex.

But what happens in the real world is that companies find a way to transition from one S-curve to another.  

Click to enlarge.

You can accept the thin commodity margins at the top of the curve, or you can make a fundamental transition.  It is hard, risky, and uncertain, but it is the only escape.  Many try.  Many fail.  A few achieve the transition and begin once again to ascend the high profit margins of the mid-S-curve.

Coca-Cola is no longer a beverage producer.  They are a brand manager.  3M is no longer a mining company.  Colgate is no longer a soap manufacturer.  Transitioning from one S-curve to another is hard but the only way to achieve continuing excess profitability.

S-curve transitioning isn't really integrated into Tainter's model.

I'll read it anyway, but perhaps it will transition a little lower in the stack.