Ages of Discord: A Structural-Demographic Analysis of American History (AoD) was published on September 30 and is available for purchase through Amazon.com. AoD is a scholarly book – in fact, it’s a work of cliodynamics. This means that there are lots of equations, 25 data tables, and 74 (!) figures. I invested a lot of effort to make the text as clear as possible, given this heavily quantitative content. My editor, Simon Reynolds, made it even more readable. Nevertheless, and unlike Ultrasociety, it’s not a popular book. This is why I am not simultaneously publishing it on Kindle. When it comes to displaying math and graphics, e-books still trail behind good old tech of printing on paper.
Many conclusions in AoD will be new to most social scientists, and the implications of these conclusions are not always welcome. This is the reason why AoD is so dense with supporting data and models – I need to build a case that will be convincing to other scientists.
Many years ago I realized that scientific rigor and popular appeal cannot be combined in one publication, and since I wanted to do both, I simply needed to write two very different kinds of books. My plan is to do the same with AoD – translate its main message into a shorter, punchier, and more vivid publication (a brochure or a thin book). And I will not do it alone, but team up with someone who is good at writing for the public. So stay tuned.
Now a few words about my additional experiences with independent publishing, incurred during the production of AoD. As readers of this blog know, I started my own imprint, Beresta Books, a year and a half ago. I began by republishing Quantitative Analysis of Movement, my very first book published in 1998, whose rights reverted to me when its original publisher dropped it. Then came Ultrasociety, with which I went through the complete publication route: from writing it to making both paper and e-book versions available for purchase.
AoD offered some additional challenges, the main of which was developing an index. I didn’t bother with an index for Ultrasociety, because it’s a popular, rather than an academic book. Even more important is that it’s available in both print and electronic version (in fact, if you buy the paperback, for an nominal additional amount you can have the e-book). A searchable text, in my opinion, is much better than an index.
With AoD, for reasons I explained above, I am not publishing an e-book version, so including a good index was a priority. I first hired an indexer, but unfortunately, she did not come through. When the deadline passed, she wrote to me saying that her computer crashed and all the work she did on the index disappeared. Perhaps. Another explanation is that AoD is a complex book, crossing many scientific disciplines, so constructing a good index for it is quite a challenge.
I then found another indexer. This one came through, although he needed extra two days to complete the job. Still, this is understandable, given the scale of the job, and the quality of the resulting index (which was very good).
This was a good learning experience. I went through a similar process with visual designers during the production of Ultrasociety (what I did then was hire three of them and pay them for initial concepts for the front cover, and then going with the one I liked best). I now have a team of reliable professionals, with whom I have worked before, for all aspects of book production. This is worth a lot.
In principle, anybody these days can start their own imprint. But in practice, I understand that many of my colleagues will not want to invest money, and especially time, into building their own publishing company. For this reason, in 2017 I am planning to make Beresta available to some select colleagues—in other words, make transition from publishing just my own books to a small publishing house that produces books in areas of my interest and expertise. I have already had inquiries from two colleagues about the possibility of doing so. The truth is, dealing with traditional publishers (and here I talk about university presses, which are typically not as greedy as commercial publishers) is a lot of hassle. Often you have to deal with—and satisfy—reviewers who don’t understand what your book is about. You have no, or little control over cover design, and sometimes even over the book title. And you have to wait a year, sometimes years, for the book to come out after you are done writing it. In contrast, the whole production of AoD took months. Additionally, the publication was delayed, first, by my travels in SE Asia, when my Mac died on me in Java, and then by the first indexer failing to deliver. Thus, it is quite realistic to have a book published in months, not years. In my case, this was an important consideration, since I wanted AoD to be published before the elections—the current political situation being a perfect illustration of our own Age of Discord.
Congratulations on the new book. WRT the chart, it’s not clear to me the intended implication. Gilded Age I comes after the discord explosion. Gilded Age II precedes it.
It’s difficult to capture all the complexity of a real society’s history in one graph. The book explains what’s going on (a quick take: superimposition of two dynamics: structural-demographic waves and fathers-and-sons cycles).
Surprised to see that the well-being index rises steadily through the 1920s and 30s. despite the Depression….
I devote quite a lot of space to explaining this interesting dynamic.
That (the fact that your graphic posits the Great Depression with double digit unemployment and mass impoverishment and world turmoil as a time of rapidly escalating well being) has been my concern with this chart as well.
I guess the foundational question is what is your definition and measurement system used to establish this precise annual measurement with nearly perfect trends? I am pretty sure you don’t have annual well being survey data going back to the 18th C, so you must be composing the well being and stress indices based upon some other measurable factors. Since this is key to your book, would you care to elaborate?
In fact, I started building a site where I lay bare all data and empirical analyses. It’s under construction, but the part on well-being proxies is mostly done:
http://peterturchin.com/age-of-discord/
Thoughts on your well being index…. It took a bit of study, but clearly what you are trying to label “well being” is actually attempting to measure “relative labor scarcity”.
For example, you don’t measure absolute well being as indicated by median income adjusted for inflation and quality (which improves pretty much at a trend line of 2% a year for two hundred years), but relative median income to GDP. This results in the counterintuitive results that the Great Depression and the Great Recession of 2008 become the eras of unprecedented gains in “well being.” According to this measure, if half our chickens died, as long as the wealthy lost three fourths of their flock, we became relatively better off. This isn’t a measure of well being. If we reduced median income to $2 a day for everyone, it would register in your index as the apex of well being, not just for this measure but for several of the others as well (which use median wage in their formula)
On height and life expectancy in the 19th C, the consensus seems to be that Fogel’s data exaggerated the data up in the early era and down in the mid century due to adverse selection of his data on army volunteers. The minor mid century reductions in these factors are explained as most likely a factor of increasing urbanization just prior to the sanitation improvements which quickly followed industrialization. Again, the larger pattern is steady increase offset by an urbanization effect.
The immigration rate inverted component makes the labor scarcity angle clear
Long story short, there was no up and down curve of absolute well being. In every case the median well being at the lowest point on your curve was superior to the well being at the highest point in the earlier trend. The general trend for health, life expectancy, child mortality, height, median income, hours worked, leisure, education, freedom, equality of rights and so forth improved pretty consistently for two hundred years. You are plotting “relative labor scarcity” and labeling it “well being.”
On the elite side, you are to some extent measuring the inverse of labor scarcity by again using median wage in your denominator. When “elite” costs go up relative to median income, one would expect the ratio to mirror the data as it of course does.
I do still find the political polarization trends and how they track with relative labor scarcity interesting. Again, this has absolutely nothing to do with well being, but with relative well being of median workers. IOW, income inequality
looks at graph
Ah, nothing to worry about, I see!
© Marie-Antoinette
When will AoD be available on Amazon.ca?
Peter, was it your finding that things like murder rates and neighborhood demographic turnover had little impact on social soundness so long as wages were good and elites well behaved? Crime and riots apparently aren’t well captured by either the well being index or the PSI, as the 60’s and 70’s register as smooth sailing. I suppose there’s a good, perverse logic to that: elites did seem to manage that turbulence much better than they handle smaller crises today, and material standards of living were continuing to improve for the working classes. But I think people will be skeptical of indices that are insensitive to the deaths of cities or, as per other comments, depressions.
Roger & Pete:
I think that a distinction has to be made between politically/hatred/discord-inspired violence and suffering (which is what Turchin graphs) and “regular” violence and suffering (caused by non-political shocks). So for instance, there was a lot of suffering from the Great Depression, but (except for a small bump in riots and lynchings in the early 30’s, which may be due to the Great Depression), riots and lynchings mostly trended down from 1920 to WWII according to Turchin’s graph.
Likewise, while there was a relentless increase in “regular” violent crime from the mid-’60’s until around 1990 (which, IMO, was caused by leaded gasoline), after which violent crime started dropping as fast as it rose, after a major outburst of riots in the late ’60’s, the political discord-related violence (riots, lynchings, and terrorism) started dropping until the late ’70’s. And while regular violent crime has been dropping since 1990, the political discord-related violence (namely, terrorism), has been rising since then.
Richard, the appreciation of that distinction is the premise of my comment – what I’m curious about is what Peter found when he included things like illicit, apolitical violence and crime. It’s a natural question because it’s not immediately obvious that it’s valid to exclude apolitical violence given that the perceived inability or unwillingness of the state to protect its citizens would seem to be a potent delegitimizer – certainly crime played an important role in facilitating the political realignment associated with the Reagan trend reversal, and it’s a key rhetorical battleground as we speak. (Also because Turchin himself seems to include apolitical spree shootings in his model – though I’m not yet sure if that’s actually included in the PSI.)
Re: PSI, I see now that it’s just economic and demographic; likewise the WBI, which also includes biological measures like stature (will be interesting to see how the immigration/genetic confound was handled, there).
Anyway, if as I initially asked, it’s the case that crime makes little or no independent contribution to chaos – that it’s only as destabilizing as the structural-demographic fundamentals allow – that’s a pretty cool finding.
BTW, in a way, Americans have been really lucky in that (other than in the Civil War), internal politically-inspired violence just hasn’t caused that many deaths compared to regular run-of-the-mill crime violence. But that certainly hasn’t been true everywhere. In major countries like China, Russia, Germany, and Japan (and maybe India as well), I daresay politically-inspired violence has led to more deaths than regular crime violence over the past century. In all those countries besides Japan, a ton of those politically-inspired killings were due to internal strife.
Long way of saying that Americans really should pay attention to Turchin’s index even though more mundane crime violence has affected Americans more in living memory (though I have to say that I am pessimistic).
I’ve just finished your excellent although terrifying new book! Although you mentioned globalization a couple times as a complicating issue, and you capture one aspect of it through the trade deficit, I think it needs more attention in the modeling: adjustments that would affect the prescription for returning the US to an integrative phase. Let me begin with a personal digression …
During my first course on economics in college 40 years ago (I did not become an economist) I was taught Adam Smith’s conclusion that trade increases the wealth of nations, but it struck me then that there is a flaw: global markets in tradable goods would lead to unified global markets for the factors of production of those goods – thinking of wages in particular – where the pricing of those factors would ultimately equilibrate throughout the global market, just as pricing equilibrates throughout any unified free market. Thus labor income in the US together with, say, labor income in India (China had not yet opened up back then) would rise, but this would come from a sharp decline in the real incomes of US workers as the earnings of Indian workers rose even further in the aggregate.
When I asked the professor about this dire outcome, he assured me it was not in the cards because labor is not the only factor of production for tradable goods: there was our great advantage in capital per worker and our technological superiority in applying that capital, so the vastly higher productivity of labor in the US would allow the disparity in wages to continue. And as time passed, the productivity growth of the US, along with the rapid pace of technological advance which seemed to maintain a US lead, apparently supported my Professor’s view, so I set my worry aside.
But now I think it is clear that a global wage equilibration dynamic had been at work for a long time, thanks in part to the free flow of capital and technology to low wage countries. One result is the decline in labor demand that you call “remarkable” on p. 213. In the model, the imports associated with our trade deficit reduce the demand for US labor.
One would think the labor supply-demand balance could be improved and the US ‘’commoners’” wage/GDP made to rise by reducing labor supply: perhaps by slashing immigration. But my fear is that this would be of limited effect because the labor supply in all tradeable goods and services industries is now global, and the global wage for those industries is that of a labor pool that demands a lower wage than American “commoners” still receive. So in the modeling, the result of cutting US labor supply would just be a further cut in US labor hours demanded, leaving the balance and the resulting wage pretty much unchanged, at least in tradeable segments of the economy, which have grown over time thanks in part to technological advances that enhance tradability.
Yet as this has unfolded in recent decades I have taken some comfort in the fact that there is another globalization effect. In most tradable industries the American wage has declined toward the global wage, but in some it has risen as the result of globalization. Four specific examples are 1) globally famous entertainers, 2) renowned medical specialists at US teaching hospitals, 3) financial service professionals and managers at firms that have globalized, and 4) the owners of certain US technology companies, for example in social media. People like these now earn much more because their labor is in more demand thanks to their now being able to address global customers. My conclusion is that globalization is thus bifurcating wages and expanding income inequality in the US. Yet I thank goodness that there are US pop stars, oncologists, hedge fund managers and social media entrepreneurs, because without elites like these and their increased earnings, which partly offset the decline of the commoners’ wages, the US real GDP would be considerably weaker and perhaps in absolute decline!
The point here is that I believe globalization throws a wrench into the basic relationship between the aggregate incomes of commoners and elites in your two-class model. That model portrays a zero-sum game where US elites are competing for slices of the US economic pie with US commoners. But in this newly globalized world, US elites of the kinds listed above are taking slices of the global pie from the commoners (and even the elites) of the rest of the whole world, thus increasing the size of the US pie. If we crush or consolidate those kinds of US elites in order to reduce US income inequality, the pie that they no longer take from the world does not revert to US commoners, improving their standard of living: it reverts to commoners and elites elsewhere in the globe. US commoners’ wages then decline rather than rise because US commoners in tradable industries are still competing against the low global wage, while US commoners in non-tradable service industries lose the business of the large segment of US elites who had benefited from globalization: waiters are no longer getting crumbs from the slices of non-US pies that those US elites took.
So as great as your modeling is when applied to analyzing history, I believe adjustments need to be made to capture the present situation and to make suggestions for its rectification.
It’s far from clear that the crumbs from the American elites offset the immiseration of the lower half or so (if you are in the lower half).
If you look at Western Europe (certainly not a paragon of GDP growth), their elites have not risen as much, but their lower half is doing better than our lower half.
I’m not saying they do. I’m saying that the immiseration of the lower half in the US is much less the result of US elites taking the lower-half’s slices of a fixed US pie than it is the result of much of the US lower half increasingly competing in a now-global labor market with a low clearing wage. Eliminating the US elites is not going to alter that global wage competition (in tradable industries) and is not going to reduce the immiseration of all those US workers. Thus the model which says that reducing elite overproduction in the US will cure the commoners’ immiseration fails, or at least it does to a large extent. And the elites may be less overproduced in the US than we think if they are serving a global market, taking pie from the rest of the world.
Eric: Again, look at Western Europe. You take it as a given that there will be global wage competition that leads to immiseration, but why is the working class in Germany (for instance) doing better than the American working class? Are they not on the same globe?
You picked a special country! Within the EU, Germany is indeed the nation whose manufacturing workers are doing well (and there is even a labor shortage stimulating immigration into Germany) in significant part because they benefit from both the EU’s monetary union and single market. An astonishing 50% of German GDP is exported, much to other EU countries. If the EU hadn’t adopted a single currency, Germany’s extreme trade imbalance would cause the exchange rate of its currency to skyrocket, making German exports much less competitive and putting pressure on German wages. But Germany’s exchange rate in the EU is forever fixed at 1 Euro to 1 Euro. The flip side is Southern Europe — places like Spain — where unemployment runs 20% and youth unemployment 50%. There is similarly low income inequality in those countries, but terrible immiseration for young people without any prospect of getting a job. So my first point is that because of currency union you need to look at the EU as a whole and not cherry-pick Germany.
I admit that even if Germany had its own currency they still have a manufacturing know-how advantage that would persist, so their manufacturing workers are more protected against the low wages of global unskilled labor than are manufacturing workers in Spain and, I am sad to say, also in the US. Japan probably also has such a know-how advantage.
But to address your point, the lower half in Germany is very happy, and the lower half in Spain, Portugal and Greece are not in open rebellion, because European government social policy is highly redistributive. But the purpose of my original post was not to argue against highly redistributive policy, although I might try elsewhere. My purpose was to raise a couple of narrow points about the coupled differential equations in the book. One is that policies that reduce labor supply in a country may not improve real wages in that country, thanks to globalized competition between, in particular, that country’s lesser-skilled workers and the global pool of lesser-skilled labor. The other is the globalization of markets for certain elite skills, so that the success of elites in the US does not necessarily come at the expense of commoners in the US. But at the bottom of page 32 there is the statement “This formulation assumes the total economic output is divided among the elites and commoners …” and it seems that the same assumption of a domestic zero sum game is carried forward through the book, although I may have missed something since I have only read the book once.
My final point, which is not about modeling, is that elites whose skills are in demand globally are valuable assets for the US and should be cultivated for the sake of US GDP. The degree to which their incomes should be redistributed for the sake of commoners is another matter that I do not address.
I think my mistake was talking about crumbs to waiters, which made it sound like I was arguing for trickle-down economics. That was not my purpose here, and I should have left the comment out.
Thanks,
Eric
Right. I’m skeptical that stopping immigration will do much good economically, though it may be good culturally as everyone assimilates and adjusts to a new America (much like what happened after the 1920’s).
Redistributive policies and strengthening of unions, I believe, would do even more societal good (they started in the 1930’s), even if the total pie is shrunk.
Also, busting up of big national monopolies/companies and championing of regional/local interests/businesses (as was started by the trust-buster Teddy Roosevelt at the start of the 20th). That may lead to more sort-term inefficiency but more long-term societal health and vitality:
http://www.theatlantic.com/business/archive/2015/11/cities-economic-fates-diverge/417372/
All that led to the post-WWII golden age that the US experienced.
I’ve finished reading the book, and I think that it is excellent. Peter Turchin did an excellent job of covering American society with respect to the structural-demographic theory. I like some of the ingenious proxies that he chose, like choices of people to name counties after, and how many people visit national monuments. Also law and medicine as elite professions. I also like the perspective on the North-South split before the Civil War.
But about average human height and longevity, might we be running into physiological limits?
Here’s something interesting about just before the Civil War. Some politicians in states between New England and the Deep South had considered creating a Central Confederacy, neither strongly supporting slavery nor strongly opposing it. Mayor Fernando Wood of New York City proposed making that city a city-state, the Free City of Tri-Insula. Talk about national disintegration.
Looking more broadly, the French Revolution seems like it was led by elite aspirants who did not like being stuck in the Third Estate, the common people, behind the first two, the clergy and the nobility.
As PT notes, a way to avoid elite overproduction is to restrict who can join the elite, to close the patriciate. Like restricting how many people can get certain elite credentials, like MD degrees, or even what ethnic groups that joiners can be of, like certain elite universities trying to exclude non-WASP’s (White Anglo-Saxon Protestants). I confess that I find the latter solution very distasteful.
I must note that what’s happening in US universities seems almost like a microcosm of Gilded Age II. Elite overproduction — lots of administrators — combined with immiseration of many of the common people — poorly-paid adjunct professors.
Europeans have kept on increasing in height.
Any numbers on that? How does it break down by nation?
Also, has anyone tried breaking down physical height by social class?
Prof. Turchin,
do you plan to write a sequel to Ultrasociety at some point?
I enjoyed it a lot, but ends rather abruptly in the 1500s/1600s or so, it would be great to write another on the same subject but covering the modern era.
Will you publish a kindle version….I live in South Africa so the cost of postage is greater than the cost of the book
Do you the works of Emmanuel Todd, a french demograph, who use some sophisticated statistics tools and anthropology to study french and european societies?
He was famous by his prediction of the USSR’s fall in the 80’s.
Yep, Todd is indeed an interesting read.
The Economic Illusion was an interesting read about the possible reasons of the rise of neo liberalism and how mass education has sadly apprarently some socilogic limits.
His theories are basicly how family structures formed our different societies and even today influlence their thinking.
He sadly became a bit a personna non grata in France after the release of his book “Who is Charlie?”
https://www.theguardian.com/media/2015/aug/28/emmanuel-todd-the-french-thinker-who-wont-toe-the-charlie-hebdo-line