An constant empirical pattern in the world is the cycle of intense warfare at regular intervals. The average interval of war between Great Powers is 34 years. The frequency of intense warfare is distributed according to a power law. In between intense wars are more frequent minor wars.

The cycles of warfare are strongly correlated to economic cycles.

This research gets controversial since it runs in the face of so-called “peace studies” which disregard this empirical data.

First, here are some of the empirical results showing the cycles. There is the Correlates of War dataset from 1820 to 1990. Jack Levy created a dataset that listed Great Powers and conflicts between them from 1492-1980. There’s also the Doran and Parsons study (1980). There’s been many, but no one can explain the causes of this pattern.

According to Levy’s research, the mean interval between Great Power Wars was 34 years. The maximum interval was 99 years.

The most intense and transformative wars occur roughly every 50 years. Each major war establishes a period of peace between the Great Powers, until the next transformative cycle, where wars escalate in frequency and intensity until they reach a major transformative war again.

Joshua Goldstein studied the same phenomenon and relied heavily on Levy’s dataset in Long Cycles: Prosperity and War in the Modern Age (He posted his whole book online).

“The great power war severity times series (fig . 11 .1) strongly suggests fifty-year cycles. In the middle years of the graph, around 1600-1800, are four regularly recurring war peaks (marked “WP”). Each is a sustained, high-fatality war that ends a series of wars of escalating severity. These peaks are spaced about fifty years apart and are followed by two more peaks (around 1870 and around 1915), also spaced about fifty years apart, though of shorter duration and no t preceded by a series of escalating wars. Only the final peak, World War II, does not fit the pattern, following too closely after the World War I peak. Going back to the years before 1600, three more peaks are visible, although much more weakly in the first two cases, which have similar though slightly shorter spacing.

Furthermore, there is a dramatic one-to-one correspondence between the recurring war peaks shown on the graph and the long wave peaks which are indicated on the figure by small arrows at the top. For nine successive long waves, until 1918, each war peak occurs near the end of an upswing phase period. From this, I date ten war cycles since 1495 (table 11 .3).”


So what patterns correlate to this pattern of warfare?

There are structuralism theories which try to explain why Great Power States rise, reach a peak in relative power, decline as competitors rise in power. How does this occur?

War begins because of political and economic transitions of power. The old international structure and international laws uphold a “status quo” which no longer exists, so the states rising in power have to overthrow the regimes. After the revolution, there is a punctuated equilibrium where the new regime reaches a balance of power.

I call it the Political Boom and Bust Cycle. The cycles represent a process of political evolution. The Long War was triggered by Industrialism and caused the transition from Monarchism to Liberal Democracy.

But this is still vague and merely descriptive. It also only marginally accounts for the power law frequency of warfare.

There are a rising number of International Relations theories which factor in the importance of long-term economic evolution as a causal factor of political change and warfare. War is about the political economy, and money funds the wars. As Cicero said, “The sinews of war are infinite money.”

Joshua Goldstein notes that wars economic upswings. In particular, Goldstein noticed that cycles of warfare correspond almost exactly with Kondratieff Waves – economic cycles of 50 years. The 50 year Great Power struggle occurs at the peak of the 50 year K-Wave.

K-waves describe long-term structural changes in the economy. This makes up part of Schumpeter’s theory of creative destruction.

The five recent cycles are:

“The Industrial Revolution–1771
The Age of Steam and Railways–1829
The Age of Steel, Electricity and Heavy Engineering–1875
The Age of Oil, the Automobile and Mass Production–1908
The Age of Information and Telecommunications–1971”

Joshua Goldstein created this image of the K-Wave and where wars fit in:

Here’s a longer explanation of the idea of K-Waves.

One of the limitations of this IR theory is the limitations of the economics theory. The K-waves hypothesis is not widely accepted much less proven. IR scholars await better data to understand the impact of economic changes on politics. For instance, we know there are economic cycles and changes, but we don’t know if they are structured or random.

We do understand some things. Money creates a dampening effect. Inflation or monetary shortages dampen the intensity and length of warfare. Wars occur less often during economic depressions because states lack tax revenue. This helps explain the recurring finding that wars occur after Depressions end – during the economic upswing. Some wars – such as World War I – are cut short because the states virtually went bankrupt before resolving the political cause of the war (this in turn makes a future war more necessary).

A radical re-structuring of the economy in something like a K-Wave would create problems in the international system. International Law is built to uphold the status quo of power-arrangements and cooperation. What happens when economic power re-aligns to favor a different state and new technologies make international laws and practices obsolete? The new state will rise and become frustrated with the status quo – this leads them to challenge the old regime.

Economics is probably not the only cause of transitions of power but it must be an important one.

This is linked to other findings. In particular, the power-law distribution of intensity of warfare, and the power-laws within wars. Wars might be a product of self-organized criticality, which I discussed in those posts. Simply, people act within a complex adaptive system. They respond to new signals and resources. Tensions accumulate into major events, such as war.

What is striking is that agent-based modeling re-creates cycles of violence through a handful of functions.
Epstein’s famous Civil Violence Models show the same pattern.

Agents grow dissatisfied with their government and turn against it faster than police can capture them. This leads to civil violence. Epstein found a surprising result. There are sudden spikes in civil violence with a mean frequency of 60 cycles. In between these brief violent outbursts is a punctuated equilibrium.


And the Frequency graph was just as stunning:

I am in no way suggesting that a simplified civil violence model holds the key to understanding war. I am just surprised that it recreates some of the patterns of conflicts.

I believe that new research into economic cycles and complex adaptive systems will lead to a greater understanding of violent political transitions of power in the international arena.