Comparing the Roman and Han economies

[guy]

 

Here’s an interesting article showing the greater income inequality in the Han Dynasty than in the Roman Empire:

 

Quote

They found that the average income for people living during the Roman Empire was approximately 2.25 times the subsistence minimum and 1.88 for those living during the Han Dynasty. They also found that the top 1% of Romans earned approximately 19% of total income, while the top 1% in Han earned approximately 26%.

 

 

https://phys.org/news/2025-04-economic-inequality-roman-empire-han.html#google_vignette

 

Here is the academic article:

 

 

https://www.nature.com/articles/s41467-025-58581-0

Edited Saturday at 01:20 AM by guy

[caesar novus]

On 4/11/2025 at 3:03 PM, guy said:

Here is the academic article:

 

 

https://www.nature.com/articles/s41467-025-58581-0

That seems refreshingly non marxist, with clever metrics and handling of uncertainty. Lots of nice quotes, like Han fiscal system perhaps better in theory but Roman better in practice. I will try to entice further reading with a quote and a dramatic graph proposing the fatal difference of Han economy:

Quote

Roman imperial authorities allied themselves with, and co-opted into the administration, the provincial elites. The prosperity and the resilience to crises of communities rested primarily on the ability and willingness of local elites to contribute to the common welfare. Instead, the Han tried to weaken local political and economic elites who might have challenged the power of the ruling dynasty.

(I can't post their graph, but look for their final fig. 3, just before their massive references section comprising 30% of length)

Edited Sunday at 11:48 PM by caesar novus

[guy]

12 hours ago, caesar novus said:

(I can't post their graph, but look for their final fig. 3, just before their massive references section comprising 30% of length)

Interesting point they make.  Here’s the graph:

[caesar novus]

5 hours ago, guy said:

Interesting point they make.  Here’s the graph:

Thank you. It's a little hard to see the numbers on similar shaped charts there, but the income per capita is almost double for the Romans. Since they don't calibrate charts with a common zero point, the different offsets are disguised and charts can falsely look alike.

Edited Monday at 05:25 PM by caesar novus

[guidoLaMoto]

I haven't read the article yet, but at first glance at the graphs presented here, I'm highly suspicious that they can discern three significant figure  accuracy in collecting such remote data from such incomplete sources. ...I bet the only thing they are justified to conclude is that the ancient 10th decile group was relatively richer than the other 9 deciles compared to the modern USA.

[caesar novus]

6 hours ago, guidoLaMoto said:

highly suspicious that they can discern three significant figure  accuracy in collecting such remote data from such incomplete sources

They are labeled simulation results and seem to be precision of the model rather than claimed accuracy of actual history. Often they are ratios ranging only from 0 to 1 or 1 to 3, and a lot of granularity is happening in that range that you wouldn't want to discard. They have a section on uncertainty which is nice to be explicit about, but you may not find it convincing.

[guidoLaMoto]

Computer modeling is art, not science. The model only tells you what you told it to tell you. 

Anyone visiting this site, for instance, probably is already well aware that the Roman elite were filthy rich and the plebs were just getting by. Any attempt at quantitating that with any more precision is just an exercise in imagination..... Not only do we  have only very rough indications of average wages, but we don't even know what the population numbers really were.

[caesar novus]

I think the art was not to determine the exact Roman and then Han situation and then compare them, with that bottom line graph that shows a stunning difference. Rather they played around with similar estimation techniques which, if biased, might be biased in a consistent way and thus cancel out in a Rome/Han ratio. Even the mental exercise sheds sidelights, and seems more useful than admitting all is more or less uncertain.

Maybe I just have warm feelings for simulations, since I worked on optimizing electronic design thru evolution or monte carlo/annealing simulations. Injecting random mutations or whatever was fun and effective until the Phd folks invented better boring math approaches.

[guidoLaMoto]

It's been said that all computer models are wrong, but some are useful. They show us what, if the modeler is close to being right, to expect. They can serve as a basis for forming hypotheses that can then be tested. That's their biggest value.

JD Murray https://link.springer.com/book/10.1007/978-3-662-08542-4 has done some amazing modeling to reproduce animal coat patterns, for instance, just by varying the values of the constants in the model.

I'm just pointing out that it's absurd for the authors of that paper to assert single digit per cent differences between Roman and Han economics when their data is probably not more accurate than a 50% level at best...Eg- did a Roman laborer earn 1 or did he earn 2 sisterci a day?...and how many days did he work each year?....and what was the total value of all money in circulation? ...How do they know?

Edited 7 hours ago by guidoLaMoto