Structural-Demographic Cycles of History

Cliodynamics is doing quantitative history as a science, complete with testing theories about historical trends. This includes theories of cycles of history, a long-contentious kind of theory. In fact, many historians are very skeptical of overall theories of history, thinking that the only safe course is to content oneself with description and maybe some low-level theories. But a theory of cycles of history, the structural-demographic theory (SDT), has come out of cliodynamics research, and it seems to be well-supported. Biologist and historian Peter Turchin has popularized it in several articles.

Preindustrial Societies
Many large-scale preindustrial agrarian societies follow cycles lasting roughly 300 - 400 years, cycles that are readily understood with a structural-demographic analysis. There are two main phases, sometimes called in the professional literature "A" and "B", but here called integrative and disintegrative, or if one wants to be very short, upward and downward.


 * Integrative - centralized, unified elites, strong state, order, stability -- may gain territory
 * Expansion (Growth) - population increases
 * Stagflation (Compression) - population levels off, elites increase
 * Disintegrative - decentralized, divided elites, weak state, disorder, instability -- may lose territory
 * Crisis (State Breakdown) - population declines, elites continue, lots of strife
 * Depression - population stays low, civil wars, elites get pruned (killed, exiled, demoted)
 * Intercycle - if it takes time to form a strong state

Disintegrative phases usually do not have continuous violence. Instead, bursts of violence occur about every 50 +- 10 years, with fragile peace in between. This cycle covers roughly two human generations, and is sometimes called "fathers and sons" cycles. One generation fights, and its successor generation does not want to relive that experience. Thus it does not fight. But the successor of that successor has much less memory of that fighting, and it is much less inhibited. So it fights.

List of Cycles
Years are: beginning of integrative, integrative to disintegrative, end of disintegrative.

(from Turchin and Nefedov, Secular Cycles, Turchin, Ages of Discord)
 * Ancient Rome
 * 650 - 500 - 350 BCE - Regal (early kings)
 * 350 - 130 - 30 BCE - Republic
 * 30 BCE - 165 CE - 285 CE - Principiate (the familiar Roman Empire)
 * 285 - 540 - 700 - Dominate (eastern empire)
 * Early Medieval Germanic
 * 480 - 640 - 700 - Merovingian
 * 700 - 820 - 920 - Carolingian
 * 920 - 1050 - 1150 - Ottonian-Salian
 * Medieval and Early Modern England
 * 1150 - 1315 - 1480 - Plantagenet
 * 1485 - 1640 - 1730 - Tudor-Stuart
 * Medieval and Early Modern France
 * 1150 - 1315 - 1450 - Capetian
 * 1450 - 1570 - 1660 - Valois
 * 1660 - 1780 - 1870 - Bourbon
 * Medieval and Early Modern Russia
 * 1460 - 1565 - 1620 - Muscovy
 * 1620 - 1905 - 1922 - Romanov
 * Imperial China (when unified)
 * 200 BCE - 10 CE - 40 CE - Western Han
 * 40 - 180 - 220 - Eastern Han
 * 550 - 610 - 630 - Sui
 * 670 - 750 - 770 - Tang
 * 960 - 1120 - 1160 - Northern Sung
 * 1250 - 1350 - 1410 - Yuan
 * 1410 - 1620 - 1650 - Ming
 * 1650 - 1850 - 1880 - Qing

The United States
Industrialism has made such radical changes in society that one might expect them to behave very differently from preindustrial, agrarian societies. But surprisingly, a structural-demographic analysis is valid for at least some industrialized societies. In particular, the United States, a nation that is a close analog of the larger preindustrial nations of the past. The United States is large, it is one of the first nations to industrialize, and it has had relatively little outside vulnerability for most of its history. So it makes a good case study for cliodynamic research.

Peter Turchin and others have done so, and they find that the US does fit the structural-demographic cycle model rather well, though the cycle is faster than for preindustrial societies.

They have looked at some 1590 incidents of political violence between 1790 and 2010, incidents like riots, lynchings, labor strife, and terrorism. Looking over the decades, the US had relatively little political violence in its early decades. But it went up and peaked in 1900, and it then declined to a trough in 1950, and it is now climbing again. Superimposed on this long-term cycle is three spikes in violence, around 1870, 1920, and 1970, though not around 1820. Thus, the US seems to have a long-term cycle of violence with a 50-year two-generation cycle of spikes added to it.

They have continued with looking at several social indicators and seeing how they correlate. They have had to remove trends and normalize the variation for easier comparison. Here is what they used and the direction of correlation relative to "good" times. Many of these variables are found with proxies, which are listed here.


 * Political Violence: ---
 * Ordinary-People Well-Being
 * Employment prospects (foreign-born fraction of workers) ---
 * Relative wage (median wage / GDP per capita) +++
 * Health (life expectancy and adult physical height) +++
 * Family (age of first marriage, a measure of social optimism) ---
 * Elite (over)production
 * Top Wealth (largest fortune / median wage) ---
 * Education Cost (tuition of Yale, an elite university / median wage) ---
 * Elite Fragmentation (political polarization in Congress) ---

The ordinary-people measures correlate well with each other, as do the elite measures, and ordinary-people well-being correlates in the opposite direction with elite overproduction. So:
 * Ordinary people do well, not many elites, not much violence
 * Ordinary people do poorly, many elites, much violence

The US thus had an integrative phase from its founding to the 1820's and the aptly-named Era of Good Feelings. It then had a disintegrative phase lasting until the 1890's and the Gilded Age, named because upper-class opulence concealed a lot of misery and squalor and social problems. This was followed by another integrative phase, starting with the aptly-named Progressive Era and ending in the 1950's and early 1960's, with Dwight Eisenhower's and John Fitzgerald Kennedy's presidencies. We are now in another disintegrative phase, one that started with the Sixties-radicalism era of the late 1960's and early 1970's, and that has continued into a second Gilded Age.

The first disintegrative phase included the Civil War, and civil wars are a common part of disintegrative phases of preindustrial societies. This suggests a novel hypothesis for that war: too many "elite aspirants" competing for a limited number of positions.

Also in that disintegrative phase was lots of immigrants. This provoked an anti-immigrant movement that called itself the Native American Party. Its members would habitually say "I don't know" about their organization, something that got them named the Know-Nothings. It was not very successful, and business leaders loved the flood of immigrants that the US got back then, because it was ideal for forcing down wages.

But the immigrants themselves started getting restless, and the US Congress sharply reduced immigration in 1925. Also back then was some major labor strife, including the Blair Mountain Mine War of 1921. It seemed as if the US would go the way of Russia, and that apparently scared the US elites into compromising with the working classes over the early to mid twentieth century, helping to create the integrative trend over that time.

Some early-twentieth-century methods of reducing elite overproduction were not very pretty, it must be conceded. The American Medical Association controlled -- and still controls -- how many people could get MD degrees and become doctors, and elite universities like Harvard tried to keep out Jews and keep themselves White-Anglo-Saxon-Protestant only.

If historical trends continue, the current disintegrative trend should end in the 2020's, and there should be a spike in political violence around then. So it is not going to be a very pleasant time. We may already be seeing some of it, with neo-Nazis on the march and a president who finds it difficult to find fault with them.

Some policy recommendations follow from this history:
 * Increase workers' pull in workforces: labor unions should not just be for business leaders and for elite professions like doctors.
 * Have a more egalitarian distribution of wealth. People may still get rich, but they would not be allowed to force down everybody else's wages for whatever reason that they might claim.
 * Restrict immigration. This policy recommendation may be hard for many liberals to take, it must be conceded. It also does not help that many opponents of immigration are bigoted xenophobes.