At the end of our last episode, our proto-capitalist tax baron has collected enough food from enough places to build an urban settlement, and has in the process created a situation of wealth imbalance: he has more than he needs, everyone else has slightly less but not so much less that they’ll die off.
And in so doing, our proto-capitalist has invented a whole new social problem; underclasses. Eventually, some bright spark always asks himself why the feller in the middle gets to take some of what he worked to create; the answer is lads with swords. The bright spark realises that there’s a lot more farmers than there are people in the city: depending on tech level and hydrogeography the rural population are about 98-80% in pre-industrial societies.  So Sparky gets some plough-shares and beats them into swords, but after the revolution he finds himself with a city to feed and they’re all looking at him and the only way to fix that is…
And so you create centralised/hierarchical state models. There’s two ways to run an effective society; educate everyone to the extent that they truly understand both ‘enlightened self-interest’ and ‘your right to swing your fist ends at my nose’, or educate a small number of people in joined-up thinking and then indoctrinate the rest with the principle of taking orders from the few. Humans en masse act like herd animals anyway; guess which of these is easier to implement in a low-bandwidth society? Subsistence farmers really don’t have time to screw around with theoretical models of governance.
Once they’ve figured out boats, these groupings of proto-capitalists start getting involved in serious trading with each other. Once you start that process, you also initiate capital competition; different capital reserve administrators start competing with each other directly, and thus you invent market capitalism. It’s intrinsically statist, because no-one who doesn’t have a hierarchical, violence-maintained state can hold together long enough to be a competitor. What determines who wins?
What do you know? How fast do you learn?
In any given civilisation conflict, a variety of factors. Look at the differences between what happened to Attic Greece during the bronze age, compared with what happened when the Medes and the Persians met Leonidas and his lads, with the career of Alexander with Genghiz Khan. Numbers make a difference, efficiency of hierarchy makes a difference, speed of movement makes a difference, geography makes a huge difference, but in the long run a pattern does emerge.
From fairly early on the determinant seems to be ‘the guy who can write’. Why? Because centralised capital hierarchies which have writing can disseminate any good idea that is had in any part of their hegemony faster and more accurately than a pre-literate competitor can. The idea of iron-working, for example gets to the capitalist; the capitalist says “That’s a good idea! I shall tell all my bronze smiths to start learning about iron. I can spread the word fast, because I have writing; I can ensure the technique is accurately reproduced, because I have writing. I can afford it, because I have capital. Everyone will do what they’re told because I have an army.” The idea spreads throughout your political hegemony. The guy over the hill is still only making iron knives for one village.
Time and again, from Mesopotamia and Europe through the meso-American civilisations, from the Parsees and the Kyrgyz down past the Ganges delta to Siam, you see one system evolving, over and over and over again. The Chinese start getting organised pretty early, as do the Koreans. In each place the exact path, rate of development and shape of society is different but everywhere that starts to tax develops bureaucracy like mushrooms in the dark. With writing comes efficient information flow, and with that comes the capacity to accumulate ever more capital reserves and centralise their control ever more effectively.
Centralised and hierarchical states are the best competitors at distributing knowledge in a low-bandwidth environment. When travel is hard and information is scarce, the more centralised your control systems, the more efficient your learning curve. The more you know, the better you learn, the faster you get the word out; these things self-reinforce to place a culture firmly on the path towards modern, industrial capitalism of the kind everyone is familiar with from the rhetoric of the Cold War.
* * *
So the answer to the question the left doesn’t like to ask is this: centralised/hierarchical capitalist models are spectacularly effective at generating and disseminating technological progress in a low-bandwidth, high-risk environment. They’re better at that than any other model which has so far been tried, anywhere. As a result, civilisations which centralise effectively and document effectively scale better than any other option at low tech levels. Britain’s success at large scale administration owes as much to Domesday Book as to the Navy.
But the story, inevitably, doesn’t end there. The left may not like the question: but when you follow through the implications of history, you realise very quickly that the right cannot accept the answer.
 The ordering is chronological. Lower-tech societies need more farmers per urban head, therefore in earlier urbanising cultures, the percentage of rural people is higher.