Tuesday, September 11, 2012

Myths of ancient China


There is a line of analysis that I see a lot in the press, that goes like this: For thousands of years, China economically dominated the world. Therefore, China's rapid growth is just a case of reversion to the mean; we should expect China to go back to being the same percent of world GDP that it always was before.

Now, I fully agree with the idea that China will return to being a bigger portion of world GDP. That's just "conditional convergence". It's a function of the Solow model. There is no good reason for 20% of the world's populace to produce less than 20% of the world's output in the very long run.

BUT, that being said, the people who make this claim seem to get their history a bit wrong. For example, here is a graph used by Deutsche Bank and PWC to make the point:


And here is one used by Michael Cembalest of JP Morgan:



Finally, here is one used by The Economist:


The source for all of these, if I'm not mistaken, is the esteemed economic historian Angus Maddison. But I have three problems with the use of this data in these charts. The first two have to do with how Angus Maddison's data gets put into those charts. The third has to do with the methodology used by Angus Maddison.

Angus Maddison's methodology, if I'm not mistaken, is something like this:

1. Assume that pretty much everyone prior to the 1700s was a farmer.

2. Very carefully estimate agricultural productivity and population density in ancient times.

3. Multiple population by agricultural productivity to get GDP.

So, here is the first problem with those charts: They leave out many of the major farming regions of the old world. For example, in the Roman Empire, the most productive agricultural region was Egypt. In fact, the Eastern Roman Empire had much higher agricultural output than the Western Roman Empire (France/Italy/England/Spain), which motivated it to periodically try to secede. 

Leaving out these regions skews the ratio wildly in China and India's favor. Including the entire Roman Empire would make the numbers for 100 AD look wildly different. For example, at its height, the Roman Empire, including all those Eastern Roman territories, had between 29 and 39 percent of the world's population. China's Han Dynasty, which was contemporary with the Romans, had only 26% of the world's population. Therefore, by agricultural output alone, Rome probably produced more GDP than China. Also, there was the entire Islamic Caliphate, which was a big deal from 700 - 1200 AD.

The second problem I have with those charts, and the conclusion drawn from them, is that they do not account for changes in the distribution of world population. In 100 AD, Germany was a sparsely inhabited forest. North America was mostly wilderness. China, in contrast, was densely settled. Today, things are different. The clearing of West Europe's forests in the early 2nd Millennium AD is why West Europe is now a densely settled population center instead of a backwater. Similarly, the colonization of North and South America, and the waves of immigration to the Western Hemisphere, has permanently altered the global population balance. China, despite going on a breeding binge that made it one of the world's most densely (over)populated regions, represents less than 20% of the world's population now, as opposed to over 25% in ancient times. As China ages, this percentage will shrink further.

In other words, China will not automatically return to its old percentage of world GDP, since the Western Hemisphere and West Europe are now on the map.

Now, on to my third problem: I think Angus Maddison may be doing things wrong. I realize this is a rather presumptuous thing to say, but I think it's true. Specifically, the assumption that GDP before 1700 was proportional to agricultural productivity seems to me not to be a good one. The reason is that even in a non-industrial society, there is a potentially huge source of GDP increases: trade. Remember, in a world where output is mostly in the form of commodities (i.e. no increasing returns to scale), the old Ricardian theory of trade makes a lot of sense. Stable ancient empires that could act as free trade zones were probably capable of dramatically increasing their per capita GDP beyond the base provided by the productivity of their land.

This is the finding of Ian Morris in Why the West Rules For Now. He constructs a "social development index" that includes things like urbanization and military capabilities, and probably correlates with an ancient region's per capita GDP (it is hard to build cities and make war without producing stuff). He finds dramatic changes in this social development index over the course of the Roman Empire; at its height, Rome seems to have been extremely rich, but a couple centuries earlier or later it was desperately poor. Morris corroborates this index with data on shipwrecks, lead poisoning, and other things that would tend to correlate with output. Basically, Rome saw huge fluctuations in per capita GDP. But it is unlikely that Rome's agricultural productivity changed much over this time. Instead, what probably happened was the rise and fall of cross-Mediterranean trade.

If trade could make Rome dramatically richer, and its absence could make Rome dramatically poorer, then Maddison's data set is wrong. Just because most people in 100 AD were farmers does not mean that most people were subsistence farmers. And frankly, I'm not sure how people use Maddison's data set without noticing this fact.

So anyway, to sum up: These graphs don't tell us as much as people think they do about how world GDP was distributed in the past. So don't use them as a guide to the future.

Update: Some people have been commenting "What about the 'other' parts? Couldn't that include the other parts of the Roman Empire?" Sure, it might, though the charts don't make it clear (look at Chart 3; if "other" = Rome, why isn't 1 AD the "golden age of the West"?). And does the residual include those portions of the world where records were not well-kept? Does it include areas that relied on animal herding? Does it include the Americas properly? And why does Chart 2 (the one by Michael Cembalest) not include "other"? In general, these charts just seem to present the data - such as it is - in a very misleading way.

61 comments:

  1. Noah! Did you miss me? Didn't you promise me a post on the economic value of tolerance?

    Let's see what Mill had to say about China...

    "China — a nation of much talent, and, in some respects, even wisdom, owing to the rare good fortune of having been provided at an early period with a particularly good set of customs, the work, in some measure, of men to whom even the most enlightened European must accord, under certain limitations, the title of sages and philosophers. They are remarkable, too, in the excellence of their apparatus for impressing, as far as possible, the best wisdom they possess upon every mind in the community, and securing that those who have appropriated most of it shall occupy the posts of honour and power. Surely the people who did this have discovered the secret of human progressiveness, and must have kept themselves steadily at the head of the movement of the world. On the contrary, they have become stationary—have remained so for thousands of years; and if they are ever to be farther improved, it must be by foreigners. They have succeeded beyond all hope in what English philanthropists are so industriously working at—in making a people all alike, all governing their thoughts and conduct by the same maxims and rules; and these are the fruits. The modern rĂ©gime of public opinion is, in an unorganized form, what the Chinese educational and political systems are in an organized; and unless individuality shall be able successfully to assert itself against this yoke, Europe, notwithstanding its noble antecedents and its professed Christianity, will tend to become another China." - J.S. Mill, On Liberty.

    We, as a society, benefit from having the liberty to choose who we exchange with. This is true in the private sector...so why isn't it true in the public sector?

    If you wouldn't be happy spending your taxes in the public sector...then that means that you would want to spend my taxes in addition to your own. The problem is that doing so would remove my perspective from the public sector.

    "And that standard, express or tacit, is to desire nothing strongly. Its ideal of character is to be without any marked character; to maim by compression, like a Chinese lady's foot, every part of human nature which stands out prominently, and tends to make the person markedly dissimilar in outline to commonplace humanity." - J.S. Mill, On Liberty.

    Why do you want to bind my feet? Do you think it would make me look sexy?

    If you removed the perspectives of everybody you disagreed with...then all our resources would be allocated according to your perspective. But who do you think Budddha's parable of the blind men and the elephant is referring to? Here's a hint...it's not referring to the people that would tolerate allowing others to spend their time/money according to their perspectives.

    If I think you would benefit from something...like reading what others have written on tax choice...Choose Where Your Taxes Go...then I should solely rely on persuasion to encourage you to spend your time according to my perspective. If I forced you to read my blog...then that would partially destroy your perspective. We, as a society, harm ourselves when we disintegrate perspectives.

    ReplyDelete
    Replies
    1. Anthony12:55 AM

      Long winded condescending trolls are so much more infuriating than the drive-by insult mongers. Good job.

      Delete
    2. I liked the John Stuart Mill quote, though. I wonder how right he was about China at the time.

      Delete
    3. Anthony2:02 AM

      China had just gotten out of the Taiping rebellion when he wrote that, so I would guess "not at all"

      Delete
    4. Noah, the other day I was debating tax choice with a friend. As he was arguing against the value of perspectives...I noticed that a large spider was descending from the ceiling directly above his head. There he was arguing that my perspective shouldn't matter in the public sector...yet from my perspective he was about to be attacked by a spider. The irony was tasty...but I finally decided to warn him and he jumped out of the way in the nick of time.

      You've stated before that you have no desire to censor comments. Why though? Why not just delete my posts? I've been blocked from a gazillion other sites...so why not just do the same thing?

      Around 2500 years ago Cyrus the Great banished slavery and promoted religious tolerance. He was responsible for freeing the Israelites from their Babylonian captors. Why did he do so? Because it was the moral thing to do? After his rule ended then things quickly reverted and it was only recently that we again banished slavery and promoted religious tolerance. Why did we do so? Because it was the moral thing to do.

      I'm not saying that it's not moral to protect liberty...what I'm saying is that history shows that a moral argument for liberty is not enough. What is needed is a consequentialist defense of liberty. There are plenty of them out there...but if any were truly effective then we would have liberty in the public sector. You would spend your taxes on the public goods that you value and I would spend my taxes on the public goods that I value.

      So why not share a consequentialist argument for liberty? Why should you have the freedom to choose how you spend your time/money in the private sector? Why should you have the freedom to express your individuality? How do we, as a society, benefit from liberty?

      Glad you liked the Mill quote...sure he was right. Human flourishing depends on how we respond to scarcity and how we respond to scarcity depends on how well we tolerate people who think outside the box. If people are forced to conform...then how can they come up with new and innovative uses for limited resources? If I limit who you exchange perspectives with...then how can you build on ideas and concepts that you are never exposed to?

      We can't stand on the shoulders of giants if the giants have had their feet bound. Chairman Mao closed China off to foreign investment and Deng Xiaoping did the opposite. Clearly, based on the disparity of outcomes, you don't limit who people are allowed to exchange with. Yet, we prevent taxpayers from choosing who they exchange with in the public sector.

      Delete
  2. Great post! Thoroughly enjoy a good debunking, especially of assumptions that are conveyed surreptitiously in the form of easily-citable data.

    ReplyDelete
  3. When did China start to pull ahead of the Romans in terms of technology? That may have been a contributor as well. I've also read arguments that the Romans were vastly more productive in terms of mining than Han China during the equivalent period.

    ReplyDelete
    Replies
    1. According to Ian Morris, the two started converging in the mid-200s and the switch happened in the 400s.

      Yes, many people think Rome was substantially richer than Han China. Which makes these graphs even sillier.

      Delete
    2. Was it Vespasian who was worried about the price of Chinese silk and the drain on Roman gold reserves? (Vespasian would bitch about that kind of thing so I'm probably off by a hundred years or so.) Rome's big export at the time was glass, but I'm not sure if much it made it back over the silk route.

      P.S. You make a good point about Egypt. It was a rich prize with its wheat production. I assume that production included Libyan and Tunisian wheat as well, as they were in the old Alexandrian division.

      Delete
  4. But Noah, what about scientific and technological advancement? Get this:

    One of the most surprisingly fascinating books is, "The Future for Investors" (2005), by famed University of Pennsylvania, Wharton, finance professor Jeremy Siegel. In it he talks about the great importance for the advancement of mankind of preservability and accessibility of information, and this regards, fascinatingly, China (this will be very surprising and contradict much of what's quoted from Little above):

    Until the onset of the industrial revolution, productivity and population growth inched ahead extremely slowly. In fact, productivity backtracked as often as it moved forward. Discoveries and inventions were made, but many were lost to the next generation. For example Rome of A.D. 100 is said to have had a better infrastructure (roads, sewage systems, and water distribution) than many European cities of the 1800s. (page 207)

    "...for a period of seven or eight centuries, Chinese civilization was by most standards the most advanced on Earth [quoting Michael Hart]" China's dominance was clearly facilitated by its ability to record and transmit information [the Chinese invented paper]. (page 209)

    However:

    "The Ming rulers rejected anything that disturbed the status quo...many books of knowledge vanished during the Ming Dynasty...Charles Jones, an economics professor at Stanford University, wrote, "China came within a hair's breadth of industrializing in the 14th century, yet in 1600 their technological backwardness was apparent to most visitors; by the 19th century the Chinese themselves found it intolerable" (pages 209-10)

    Politics was enormous – this shows the great power that good democracy can have.

    But could you imagine, China industrializing in the 1300's. Cars and airplanes in the 1400's! Computers in China at the time of the Renaissance!

    ReplyDelete
    Replies
    1. Anonymous7:48 AM

      there was that little episode of book burning in Alexandria by which Christians set back the West what, 1000 years, or so many have written?

      Delete
    2. Well, it's really interesting. The Great Chinese Stagnation of 1400-1900 is one of the big mysteries of all time.

      However, I'd say that the Great European Stagnation is an even bigger mystery. Rome was the richest state in the world, and Greece had the best math and science. They were way ahead of China, Iran, and India. And then when Rome collapsed, Europe lapsed into a stagnation that lasted over 1000 years - twice as long as China's stagnation, at a much lower level of wealth. Now THAT is scary...

      Delete
    3. Anonymous11:56 AM

      There's a whole discipline dedicated to such mystery...it's called history.
      And by the way, Europe did regressed in terms of government, population and life expectancy but not in math, science or military science.

      Delete
    4. There's a whole discipline dedicated to such mystery...it's called history.

      You're a poet and didn't know it!

      And by the way, Europe did regressed in terms of government, population and life expectancy but not in math, science or military science.

      Sure, but the same is true of China's stagnation. Ideas need institutions to be implemented for the benefit of the populace.

      Delete
    5. I think saying that China was a "hair's breadth" of industrializing in the 14th century is stretching it. There's no question that they were a relatively rich, productive society by 14th century standards, but they also lacked virtually all of the technologies necessary to make industrialization possible aside from possibly water and wind-powered mills. They also lacked the incentives to create and use more advanced machinery, something that didn't change until the late 19th century.

      I've read that excellent Morris book. One interesting thing he points out is that as late as 1880, the cost of a single steam pump for a mine was the same as roughly six hundred Chinese laborers. I suspect the ratio was even worse decades or centuries before then, so they never had an incentive to substitute machines for labor (unlike the British and particularly Londoners, who had some of the highest wages in the world in the late 17th/early 18th century).

      Delete
    6. That's a good point. Rice culture is more efficient than wheat culture, so China could have a much higher population per acre of cultivated land. Labor was always relatively cheap and that hinders productivity. The industrial revolution was driven largely by high wages and labor shortages. Otherwise, why bother?

      Delete
    7. You could have all the workers in the world to substitute for industrialization, they still can't duplicate a locomotive, or armored steamships. If they had the scientific and technological ability, there was still plenty of incentive for China to industrialize.

      Delete
    8. The problem is that in order to get to a steam locomotive, you need to build efficient steam engines first, and that's an iterative process that usually starts at much more inefficient and crude steam engines if you're the first society to develop steam power.

      The first steam engines sold commercially in England, for example, were the Newcomben Engines used to pump water out of mines. They were so inefficient and expensive that they were only used to pump water out of mines for nearly six decades, until James Watt produced a much more useful engine for a market that had adopted steam power in the past and was willing to use better variants. The only reason the British mines kept using the older engines was because British wages were high enough to make it worthwhile.

      China had much more and cheaper labor than the British. If someone came up with the equivalent of the Newcomben Engine (or introduced it to China in the early 18th century), odds are it would never be widely adopted. In fact, I suspect they would never even get that far - there was a predecessor engine even before the Newcomben Engine that was never widely adopted for commercial use (Thomas Savery's engine), but which received support because (again) high British labor costs made it worthwhile to experiment with them.

      Delete
  5. Can one talk about pre-1800 productivity without talking about the Malthusian model?

    ReplyDelete
    Replies
    1. I think people like Maddison think that the Malthusian model makes it very easy to talk about productivity, since by that model it's just a function of land. I'd like to complicate the picture by bringing trade into the picture.

      Delete
  6. Using GDP as a measure of economic output is fishy enough even now when we can measure it quite directly. Estimating it as a derivative of historical population data is bunk. If we could really measure global GDP I am sure that there would be huge divergences from the data here.

    That said, I think there is some mean reversion going on, because I think there is in the very long run a good deal of correlation between population and output. The difference was technology and the industrial revolutions, but that difference has not lasted forever.

    ReplyDelete
    Replies
    1. Yep.

      BUT, the mean has changed. North America and South America are on the map now. And the distribution of arable land has changed massively too. China will never again reach 25% of world population (barring really weird and dramatic global events), and probably not even 20%.

      Delete
    2. I'm not sure that we really know what the norm was in the Americas. There is a lot of evidence that tends to suggest that the pre-Columbian Americas were far more widely populated than is recognised.

      http://www.theatlantic.com/magazine/archive/2002/03/1491/302445/ (Long read, but one of the most interesting things I've read for a long time)

      This would be another reason why the graphs in question are even wronger than you suggest.

      Delete
    3. Good point! I didn't even know about that...

      Delete
    4. I would strongly, strongly recommend both of Charles Mann's books, 1491 and 1493.

      One particularly fascinating bit of it was the section on human influence on the Amazon basin rainforests, something that has been turning up in archaeology done there in recent years.

      Delete
  7. The legend on Cembalest's graph describes Turkey and Iran as "non-Asian". Are his numbers comparably accurate?

    ReplyDelete
  8. the graph police say
    those are really crappy graphs; couldn't you have taken 5 minutes to re draw them into something intelligble for your blog ? you clearly spent a lot of time writing this post, yet, for some reason, you didn't feel that it was necessary to spend any time to make the data presentable.
    typical economist: empirical facts arn't actually that important, or at least, not as impoortant as clever arguments

    ReplyDelete
    Replies
    1. Those were the graphs released by two well-respected analyst reports and one major business magazine article.

      If I altered the graphs before criticizing them, I would hardly be taken seriously, now, would I?

      Delete
  9. Anonymous11:59 AM

    Hey Noah,

    For what it's worth, there are weather related explanations for both dips in European Middle Ages agriculture and indeed, even some dips in Roman times that you attribute to fluctuations of trade.

    Worth noting that trade inside Ancient Empires didn't really look like trade now in terms of economic ripple effects. Wars and slavery raids also upset the calculations.

    Finally, there's good reason to believe that once you get that trade explains much less than developments in energy consumption. As you seem to like investigating the murky assumptions of economics, it's a good one to look at, because economics likes to deify trade on remarkably shaky evidence...

    ReplyDelete
    Replies
    1. Anonymous12:00 PM

      Bad English there. "good reason to believe that once you get to the Middle Ages, trade explains..."

      Delete
    2. Finally, there's good reason to believe that once you get that trade explains much less than developments in energy consumption.

      Those are highly endogenous to each other... ;)

      Delete
    3. Rome was heavily trade oriented. They ran a network across the Mediterranean and up to England. Rome traded with India through the Red Sea and with China through the silk route. The Italian heartland was long dependent on imports of north African wheat and the senatorial class consisted of wealthy multinationals with interests spanning the empire. We have wars too, and Roman slavery was not chattel slavery like modern slavery in that slaves could own property and buy their own freedom. Nero even granted them standing in court which is probably why he got such bad press.

      Delete
  10. I actually thought the way Maddison constructed his world estimates was not by looking at agricultural productivity, but looking at urbanization, and ascribing higher output to more urbanized countries.

    ReplyDelete
  11. I think your first critique of the charts may be incorrect, at least as regard the first and third charts. The third chart only shows GDP for eight countries, but they don't add to 100%: there's a large residual. (It's even larger than it looks, since the vertical axis only goees up to 80%.) And the first chart's percentages for China appear to match the third's. They also match the numbers you get from calculating China's share of the world's GNP directly from Maddison's figures.

    ReplyDelete
    Replies
    1. You could still argue, of course, that even if the numbers are right the graphs are misleading since they give the impression that China and India were the only economic superpowers from 1-1500 AD. Perhaps that was your original point and I misunderstood, in which case my apologies (and forget I said anything).

      Delete
    2. Nope. Check the difference between the residuals on the three charts.

      Delete
    3. I did. The residuals for the first and third charts are very close to each other, once you take into account that the vertical axis on the third chart only goes up to 80%. The residuals on the second charet are substantially smaller, but I didn't defend that chart.

      Delete
  12. Noah:
    "So, here is the first problem with those charts: They leave out many of the major farming regions of the old world."

    No, they don't! They're included in "Other" (Chart 1), "Non-Asian Ancient Civilizations (Chart 2), or are simply the residual (Chart 3).

    Noah:
    "Also, there was the entire Islamic Caliphate, which was a big deal from 700 - 1200 AD."

    This is surely reflected in the estimates for "Other" in 1000 AD etc as noted above.

    Noah:
    "For example, in the Roman Empire, the most productive agricultural region was Egypt. In fact, the Eastern Roman Empire had much higher agricultural output than the Western Roman Empire (France/Italy/England/Spain), which motivated it to periodically try to secede."

    I'm not sure about the "agricultural" claim but the most populous part of the Roman Empire in 1 AD was probably a tie between what is now modern Italy and Asia Minor (8 million each). The eastern part of the empire didn't really surpass the west until the time of Constantine by which point Asia Minor was by far the most populous part of the empire and Egypt may have been as populous as Italy. But the time of Constantine is not depicted on those charts.

    Noah:
    "Leaving out these regions skews the ratio wildly in China and India's favor. Including the entire Roman Empire would make the numbers for 100 AD look wildly different. For example, at its height, the Roman Empire, including all those Eastern Roman territories, had between 29 and 39 percent of the world's population. China's Han Dynasty, which was contemporary with the Romans, had only 26% of the world's population. Therefore, by agricultural output alone, Rome probably produced more GDP than China."

    Here your major problem is relying on Wikipedia as a source of reliable information. Those Roman Empire population estimates (65 to 88 million) supposedly come from a biography of Marcus Aurelius by Frank McLynn and are far higher than most scholarly estimates which range from 46.9 million in 1 AD according to C. J. Russell (1958) to 60 million in the 2nd century AD according to Walter Scheidel (2006). Maddison (2006) estimated the Roman Empire's population was 48 million in 1 AD (he estimates the world population was 226 million) which is almost identical to Russell's estimate.

    I've seen estimates of the Roman population as high as 130 million but I've also read accounts of Bigfoot sightings, UFO abductions and pots of gold found at the end of rainbows.

    In any case if you sum up Maddison's estimates for Eastern and Western Europe (which does not include the former USSR), and North Africa and Iraq and Turkey for 1 AD that comes to about a quarter of the world's GDP. One could make some adjustments by subtracting parts of Northern Europe and adding in the Eastern Mediteranean but based on his numbers I think that would all roughly balance out. This would probably be pretty close to the Roman Empire's share of world GDP at peak (2nd century AD).

    P.S. His estimates of GDP per capita range from a low of $400 (1990 International Geary-Khamis dollars) in most parts of the world in 1 AD, to $450 in China, $600 in Egypt to $809 in Italy.

    ReplyDelete
    Replies
    1. Hey! I was hoping someone would come and argue! ;)

      "So, here is the first problem with those charts: They leave out many of the major farming regions of the old world."

      No, they don't! They're included in "Other" (Chart 1), "Non-Asian Ancient Civilizations (Chart 2), or are simply the residual (Chart 3).


      Please note the enormous discrepancy between the three charts as to the size of "other". You will find strong support for my contention... ;)

      Here your major problem is relying on Wikipedia as a source of reliable information. Those Roman Empire population estimates (65 to 88 million) supposedly come from a biography of Marcus Aurelius by Frank McLynn and are far higher than most scholarly estimates which range from 46.9 million in 1 AD according to C. J. Russell (1958) to 60 million in the 2nd century AD according to Walter Scheidel (2006). Maddison (2006) estimated the Roman Empire's population was 48 million in 1 AD (he estimates the world population was 226 million) which is almost identical to Russell's estimate.

      And how is ancient China's population estimated? ;)

      Delete
    2. "Please note the enormous discrepancy between the three charts as to the size of "other". You will find strong support for my contention..."

      There's nothing wrong with Chart 3. The residual is 31.4% in 1 AD. Similarly Chart 1's "Other' is 27.5% of world GDP in 1 AD. (I don't like the fact that Chart 1 implies detail where there is none however.) But you are correct about Chart 2. There's 11-12% of world's GDP in 1 AD simply missing. None of these regions are major (Algeria is the most important, followed by Mexico, both about 0.85%) but collectively they do add up.

      "And how is ancient China's population estimated?"

      The 26.0% figure for the Han Dynasty of course comes from Maddison. But I'm not sure what your point is. I trust an economist's estimate over a biographer's. The Roman Empire never had 39.5% of the world's population.

      Delete
    3. There's nothing wrong with Chart 3.

      How do you know that? You don't know that. The residual is unlabeled. Which parts of that are part of the Roman Empire? Does that residual include Iran, Africa, etc.? Does it include the herding economies of Central Asia? In a world where everyone is a farmer, huge herds of sheep, horses, etc. are worth a lot.

      The 26.0% figure for the Han Dynasty of course comes from Maddison.

      OK, and where does he get that?

      For the Roman Empire, use the low figure on Wikipedia; it's 29%. That's still higher than Maddison's figure for the Han Dynasty.

      Here, read up:
      http://en.wikipedia.org/wiki/Classical_demography

      Also, none of this answers my critique of Maddison's methodology, which seems to be at odds with archaeologically observed wealth levels.

      Delete
    4. "How do you know that? You don't know that. The residual is unlabeled. Which parts of that are part of the Roman Empire? Does that residual include Iran, Africa, etc.? Does it include the herding economies of Central Asia? In a world where everyone is a farmer, huge herds of sheep, horses, etc. are worth a lot."

      I know it because I can do the arithmetic. Assuming Maddison's estimates are correct all of that is included in there.

      (The 26.0% figure for the Han Dynasty of course comes from Maddison.)
      "OK, and where does he get that?"

      Hmmm, Maddison seemingly gets China's population in 2 AD from the Book of Han:

      http://en.wikipedia.org/wiki/Book_of_Han

      However, John D. Durand's estimates put the Han Dynasty population even higher, at 74 million instead of 60 million.

      And there's nothing in the Wikipedia section on Classical demography that alters my perception of the accuracy of Maddison's estimates for the Roman Empire's population. Of course the population of Rome surely was higher at its peak (2nd century AD) than in 1 AD and Scheidel puts that figure at 60 million. But then the world's population was probably higher then as well. So it seems very likely that China had a larger population than the Roman Empire in the first few centuries AD.

      "Also, none of this answers my critique of Maddison's methodology, which seems to be at odds with archaeologically observed wealth levels."

      True, but I have my own doubts. Agricultural productivity may not be a very good proxy for pre-1700 GDP per capita. I think that is a very good point.

      Delete
    5. I know it because I can do the arithmetic. Assuming Maddison's estimates are correct all of that is included in there.

      Mmm, uh huh. Garbage in, garbage out, as they say...

      Hmmm, Maddison seemingly gets China's population in 2 AD from the Book of Han

      Yep. DEFINITELY more reliable than the biography of whoever.

      However, John D. Durand's estimates put the Han Dynasty population even higher, at 74 million instead of 60 million.

      This should give you an idea of the uncertainty that comes baked into these figures.

      Note that you are very eager to accept without question the Han population numbers, but you instantly distrust and discount the higher numbers for the Roman Empire. Interesting...

      Delete
    6. "Note that you are very eager to accept without question the Han population numbers, but you instantly distrust and discount the higher numbers for the Roman Empire. Interesting..."

      Well, where do those higher numbers come from? I haven't read McLynn's biography of Marcus Aurelius, but I suppose he must have gotten them from somewhere else. All of the "estimates" from 65 million to 130 million seem to be pulled out of thin air, don't they?

      Delete
    7. Well, where do those higher numbers come from? I haven't read McLynn's biography of Marcus Aurelius, but I suppose he must have gotten them from somewhere else. All of the "estimates" from 65 million to 130 million seem to be pulled out of thin air, don't they?

      Yep. As with the Han estimates.

      Delete
  13. Anonymous9:39 PM

    Your argument about trade is interesting, but isn't there a Jensen's inequality problem? I would have thought that traded goods mostly affected the lives of the wealthy, and the wealth distribution of in the Roman empire was extremely skewed, no?

    It seems plausible to me to say that, until pretty recently, the median person in any society spent most of their income on food. In which case, China's higher agricultural productivity would have made a big difference...

    ReplyDelete
  14. The charts are somewhat believable, but I've been wondering how you'd measure GDP in that era. The authors seem to have accepted a food based metric, maybe kilo-calories or people-rations.

    I'm not sure this accounts for the surplus implied by the size of each empire. I'd be tempted to use a power curve and assume high levels of wealth at the imperial level with a rapid fall off to the peasantry. Still, the integrals under these curves would be meaningful given the wealth of the Chinese and Roman aristocracies.

    ReplyDelete
  15. Passing By5:05 AM

    "Just because most people in 100 AD were farmers does not mean that most people were subsistence farmers"

    Actually, I think it does.

    Subsistence farming is producing just enough food and fiber for one's own household, or a bit more. Suppose 90% of the populaton are farmers. Then each farm household produces for 1.1 households on average ... subsistence farming. Or said diffferently, if 90% of households each produce for two households, then you have an immediate glut.

    ReplyDelete
  16. Maddison's estimates of Chinese agricultural productivity are based on studies by Dwight Perkins in 1969, who assumed 285 kg of food per capita from the Song to the Qing. Chinese scholars have since gone to the archives to try to reconstruct per capita agricultural production. Below, just as one example, is estimates from Zhengzheng and Wang Xingping in 2000

    Period Per capita(kg)

    Han 496.5
    AOF 483
    Early Tang 628
    Late Tang 579.5
    Song 661.5
    Yuan 676
    Ming 777
    Qing 770.5

    I haven't read Why the West Rules for Now. The only frame of reference I would have is debates by people comparing Roman and Han technology with Morris often used as a source on Chinese technology by nonspecialists. The general impression I get is that the those citing Morris don't have a clue when it comes to Han technology. One debater, who was a grad student in economics, cited Morris as putting agricultural productivity at 300 kg per capita in the 14th century Yangtze Valley. If that is an accurate citation, then Ian Morris is using outdated secondary sources (since he assuredly is not using primary sources as he is not an historian of China). The above numbers I posted has Ming Dynasty agricultural productivity at 777 kg per capita, and Guo Songyi puts it at 850 kg for the average peasant.

    I really can't nor desire to go into any thorough comparison between the two empires. I will note, however, that in many ways Han agriculture technology was more advanced. For instance, the rotary winnowing fan, which was a mechanical device used to separate chaff from grain through the use of artificial wind by means of a crank handle. This technology would not be imported to Europe until the 1700s. The Chinese had the multitube seed drill in the second century BCE, whereas Europeans didn't have a seed drill of any kind until the 16th century CE.

    The Han also used animal power much more effectively than the Romans. Whereas the Romans used oxen on their farms, the Han used horses, which are twice as productive. Rome was also probably more dependent on slave labor (indeed, they used slaves like farm animals) than any other contemporary civilization. The huge urban populations in Italy were supported largely by tribute, which could be sustained by conquest. That is one of the reasons the Roman Empire went into an almost continuous decline after it reached its peak territorial extent in the early second century CE.

    As for the Census: the estimates of Han cities are considered accurate, because they were planned in a grid, separated by wards, with a specific number of houses per ward. The countryside was not planned in a grid, so it was possible to undercount there. Generally speaking, Chinese censuses are usually to much more likely undercount than overcount population.

    ReplyDelete
    Replies
    1. The Roman Empire didn't go into a "continuous decline" after the early 2nd century. Large segments of it, particularly North Africa and western Asia Minor (where Constantinople was built) were extremely prosperous in the 4th century CE, and other areas went up and down between the 2nd century and 4th century. It wasn't until the very late 4th century/early 5th Century that the western Roman Empire went into a death spiral, and that coincided with a period.

      Delete
    2. You are half right--the economic decline of the Roman Empire actually began in the 1st century CE. This is corroborated by declining lead production as well as a decline in shipwrecks (which indicates significantly less activity on the seas).

      http://en.wikipedia.org/wiki/File:Mediterranean_shipwrecks.jpg

      There are no significant enough improvements in shipping technology to explain the significant decline in Mediterranean shipwrecks. The best explanation is a decline in economic activity.

      An increase in overland trading would also not explain the decline, since overland transportation before railroads was orders of magnitude more expensive than sea transportation. According to Diocletian's price edict, travel by river was 10 times more expensive than travel by sea and travel by road 50-60 times higher.

      There is also a dramatic decline in lead production.

      http://en.wikipedia.org/wiki/File:World_Lead_Production.jpg

      Just one quick note on the title and labeling of the chart: the data for lead production comes from Greenland ice cores. The chart says "world lead production," but the Greenland ice caps only have particulate residue from Anatolia to the Mississippi basin region and for mega-particulate events such as massive forest fires.

      Anyways, the decline in lead production could have been caused by the exhaustion of silver mines. That doesn't, however, contradict the notion that the Roman Empire was in continuous economic decline. Geoffrey Ingham referred to the Mediterranean political economy as a "military-coinage" complex. Conquest was funded with coinage, the land from conquest was distributed among the soldiers and slaves were used to work in mines to produce more coinage. An exhaustion of the mines would throw a wrench into this whole complex.

      Environmental factors play a role as well, such as the blighting of the soil that occurs over centuries in societies using pre-modern agricultural methods (with the exception of areas where there is frequent flooding, such as the monsoon regions of China and India as well as along the Nile).


      Delete
    3. The Han Dynasty had earthen floors; the Roman Empire had plumbing and aqueducts.

      Also, look at miles of roads built and number of shipwrecks found. No comparison.

      Rome had sea trade, and the Han had not yet built the Grand Canal. That's what Morris thinks made the big difference.

      Han agriculture was more productive than Roman agriculture, except in certain places like Egypt where the land was really good. But Rome had more arable land total. Remember, the Han grew mostly wheat; China hadn't invented really good rice cultivation techniques yet.

      Delete
    4. The Han had the blast furnace. The bloomery would not be completely replaced by the blast furnace in Europe until the Industrial Revolution. In warfare, Chinese trebuchets could send heavier stones than the torsion catapults of Rome--the gravity powered trebuchet would eventually replace the Greek torsion powered catapault in medieval siege warfare.

      The total amount of canals in Chinadwarfed that of the Roman Empire. The longest canal in the Roman Empire was the canal of pharaohs at 35 miles long. Four Chinese canals were over 60 miles long, and one was over 200 miles. The total length of canals in Egypt, where most of the canals of the Roman Empire were concentrated, was 100 miles. Chinese canals could easily handle 100-200 tons compared to a few dozen tons for Roman canals. Of course it would only be fair to note that the Romans didn't need to build canals because they had the Mediterranean.

      The Romans built twice as much roads, that were paved, but the Chinese had a smaller administrative area and their roads were on average twice as wide as Roman Roads.

      The Han were very sophisticated in hydraulic engineering. Even before the Han of three great canals built during the Qin dynasty, 2 of them are still used today. Canals were also built for irrigation purposes,water distribution was handled by a complicated network of troughs and trenches, and mechanical devices such as water wheels and siphons were used to raise and lower water levels. Through access channels in their irrigation systems water could be sped up to allow it to cool in hot environments, or slowed down to allow it to gather heat in warm environments.

      The Han mostly grew millet and rice. By the Warring States period, an agricultural laborer could support 7 people with 4.5 acres, whereas a Roman farmer could only support 3.6 people on the same amount of land. The Han produced around 870 kg of millet per hectare, whereas the Romans produced about 410 kg per hectare.

      In terms of urban population, if you believe the high count hypothesis, the Roman Empire had the largest city of the two empires, but Han cities were on average larger, with their largest city at 500,000, fifth largest at 270,000, and tenth largest city about 240,000 people, whereas the Roman Empire's largest city had 1 million, 5th largest 125,000, and 10th largest 80,000. However,the low-count hypothesis of Rome, which is supported by archaeology and cliodynamics, puts the population of Rome at about 300,000-500,000.

      I just don't think GDP comparisons of past societies are all that relevant, regardless if its Maddison's method or Morris's method. It's just too simplistic.

      If we were to compare standards of living, technological development, state power, or economic models, what would we find? Well, the standards of living were about the same between the two powers. In both empires the vast majority of the population lived under a Malthusian regime. An average person in Rome and China lived at about 1.5-2 times subsistence level during normal times. Osteological analysis, as reported by Scheidel, supports the Malthusian interpretation of the Roman Economy--average heights tended to fall under Roman rule, and did not start rising until the 5th century CE. Technological development was similar in magnitude but different in kind--perhaps you could give the edge to Rome, but both experienced tremendous innovation, albeit in a limited way. Both were severely limited by the preindustrial constraints of time and space. At the end of 1,000 years the Romans were no closer to an industrial revolution, to annihilating time and space, than they were at the beginning. As for state power, neither could project very far.

      Perhaps we could say that the Romans had a better economic model. I don't think that's relevant either, just because I believe that our modern economic institutions did not evolve from Greco-Roman society, but from the Middle Ages--specifically the high middle ages, where I do believe there was a break.

      Delete
  17. Anonymous5:24 AM

    I don't follow the Maddison criticism. Why the difference in his GDP per capita figures if agricultural productivity was assumed to be uniform? Wouldn't the distinctions indicate he took some manner of economy-wide productivity differences into account--given that Chinese and Roman farmers must have achieved similar yields?

    Sadly, he is no longer around to ask.

    ReplyDelete
    Replies
    1. Here is what Noah wrote

      " 1. Assume that pretty much everyone prior to the 1700s was a farmer.

      2. Very carefully estimate agricultural productivity and population density in ancient times.

      3. Multiple population by agricultural productivity to get GDP."

      The assumption is not that agricultural productivity was uniform, but that most everyone before 1700 was a farmer(or to put it in another way, the composition of the labor force was uniform). By assuming that the population was predominantly employed in subsistence agriculture everywhere, he can make GDP per capita comparisons just by looking at differences in agricultural productivity.

      Delete
  18. "look at Chart 3; if "other" = Rome, why isn't 1 AD the "golden age of the West"?"

    Because by that time most of the Roman Empire was outside Western Europe. (And don't you mean Chart 1?)

    ReplyDelete
  19. Anonymous9:32 PM

    A lot of ancient China traditions, culture, and ancient methods being used to their every life are still alive today. China had really still survive that up to now.
    study in china

    ReplyDelete
  20. I don't think anyone would consider your criticisms of the Maddison data to be presumptuous. I'm pretty sure that the consensus is that while the Maddison data is a good place to start, it is definitely not the best data available.

    ReplyDelete
  21. Anonymous4:59 PM

    Do not we have a good economic data on ancient Mesopotamia from all those cuneiform tablets? I think there is more then a handful of studies on many aspects of Mesopotamia's economy including international trade. Which was quite big from prehistoric times (how can anyone neglect it, anyway?)... it is said that the only commodity produced in Mesopotamia was wheat, everything else they had to import (a bit of exaggeration, but not too big).

    ReplyDelete
  22. Does 'other' include the South American/Central American populations? Weren't they relatively productive as well in ancient times? Or are they beside the point for this discussion? I suppose so.

    ReplyDelete
  23. Therefore, China's rapid growth is just a case of reversion to the mean; we should expect China to go back to being the same percent of world GDP that it always was before. buy from china

    ReplyDelete