In my last blog, I gave a summary of the elite
overproduction theory of political instability. Overall, I think it could be an
important theoretical perspective on the cause of some of history’s worst
conflagrations. But how much confidence can we have in it? What limitations
does it have? Can it contribute to a political theory of a more harmonious and
fair society?
It’s difficult to say how strongly the “structural
demographic” theory is supported by the evidence. So far I’ve located only a
scarce few papers that flesh out a potential mathematical model for analysing
instability using demographics and analysing historical situations using the
model. One of them was written by Turchin himself in 2013 and models the
Antebellum US and the US in the 21st century. Another was written
about Ukrainian politics around the Euro Maidan, a situation which Turchin also
draws upon in his book, End Times.
Turchin’s paper “Modeling Social Pressures Toward Political
Instability” is dense with formulae that aim to model the chances of political
instability for a society based on data that he and other social scientists
have gathered into the SESHAT project, a database of demographic information on
historical societies. The math looks forbidding, but it is based on a few main
assumptions. These are that the potential for instability is based on the
mobilization potential of both elites and commoners; i.e. the number of
politically and economically frustrated people exist in a society. This takes
into account the number of young people (the most likely revolutionary foot
soldiers), how many live in cities, and the relative wage, a measure of how
wages levels track economic growth. Turchin assumes that when wages are
depressed relative to growth, the elite are taking the surplus, and this contributes
to an increase in their number. However, if this goes on too long, the elites
start to feel pressure – there is neither enough surplus to go around, nor jobs
in government capable of maintaining their ambitions for power. Thus, their
mobilization potential goes up, and instability becomes more likely.
This is all very impressive, but despite some of its
subtlety, it feels a bit like looking into the code for a strategy video game
like Victoria 3, where the player has to provide the various castes of their
chosen 19th century nation with enough of their needs to prevent
rebellion while preparing for war and conquest with their rivals. That is to
say, it actually feels really quite blunt. A lot more blunt than I was
expecting. In End Times, Turchin defends the need for the models to be as
simple as they can be without leaving anything important out, but this
obviously leaves us with something that feels, not coincidentally, like a video
game. However, I think Turchin would not mind this objection too much. It
feels, especially in the heat of political action, that the human will is wild
and free, and it is a bit disappointing to find out such expressions of freedom
can be reduced to what looks like a game of Risk. However, all insight into
human behaviour generates new possibilities, so perhaps oversimplification
might focus attention on what is important in the political arena.
However, until there are more papers applying this model to
different societies and crises, it’s difficult to say how strong the evidence
is right now. Until then, a critique of it which involves criticizing its lack
of depth or consideration of the hundreds of potential mitigating factors
involved in political crises is more or less hot air. Of course, it misses out
the rich detail of the factors involved in this or that war – it is supposed to
focus on what makes up most of the variance; and in doing so, it will reduce
the apparent importance of such factors as ideology or specific grievances.
The Sri Lankan Civil War
In an effort to see if it does make some sense, I decided to
learn about the political instability in a country I previously had very little
knowledge of: Sri Lanka. What little I did know about it is that it endured an
almost thirty-year civil war fought on ethnic grounds, between the northern and
eastern Tamils, and the majority Sinhalese. I remember the war ending in 2009,
but beyond that, I had very little background in it. I seemed to me a good test
case for the theory, because if it showed that ethnic tension was the root of
the violence, this would act as a falsification of the importance of elite
overproduction. Thankfully, I found a detailed, sensitive retelling of Sri
Lanka’s history on YouTube, by travel journalist Alex Rothman here: https://youtube.com/playlist?list=PLhRLWBHRg9F2DlPmWNgeYR4pMlAOWi_dt&feature=shared.
Watching it, it really was tragic to think that such a beautiful place had
been subjected to a generation of bloody violence; and I think this is one
reason why research like Turchin’s is compelling – perhaps truly understanding
the causes of such violence is a key to preventing it.
Without going into too much detail, the Sri Lankan civil war
officially began in 1983 when a group of militants known as the Tamil Tigers
(the LTTE) attacked a government patrol, and the majority party retaliated by
instigating a massacre of Tamil civilians. The war then continued off and on
until 2009, with the LTTE making regular attacks on government forces, even
carrying out suicide bombings, in the hopes of creating a breakaway Tamil state.
Other nations were also involved, most notably India, resulting in the
assassination of a former Indian prime minister, Rajiv Gandhi, by the LTTE.
Eventually, the Chinese intervened and provided the government with weapons it
used to eliminate the Tigers and end the war. The UN estimated that 100,000
people were killed in the war.
It's easy to assume that this conflict was motivated purely
by “ethnic tensions”. Many Sri Lankan Tamils resented their minority status in
Sri Lanka, and were fearful of Sinhala persecution, with plenty of
justification – there had been anti-Tamil pogroms on several occasions after
independence had been gained from Britain in 1948. However, I find this
explanation just too easy, even lazy – it feels like a “western” stereotype of
global south conflicts such as Rwanda, or even European ones, such as Yugoslavia.
However, there are plenty of examples of minority groups in ethnically
segregated areas living peacefully next to majority neighbours in the same
country. There are even cases where the right to self-determination of that
region was denied, and this still did not lead to long term violence. Think of
Scotland and Wales, or the recent breakaway attempt of Catalonia from Spain.
Though the anti-Tamil pogroms certainly fueled the militant imperative for
Tamils to demand their own state, we must also ask what sparked these “ethnic
tensions” in the first place?
More than ethnicity
As a small, post-colonial nation, Sri Lanka unsurprisingly
struggled to develop economically during the 20th century. A
hangover from British administration was that there were a disproportionate
number of Tamils in the government service, and this rankled the Sinhalese
majority who wanted the state to act in their interest first. Their first
strike was to institute the language act of 1956, replacing English with
Sinhala as the language of official documents – making many Tamils in government
who were unable to speak Sinhala obsolete. As Rothman notes, to get elected,
Sinhalese politicians only had to appeal to the majority sentiment on the basis
of shared identity, and this gradually increased Sinhalese power vis a vis
Tamils in mid-century Sri Lanka, which was also going through a huge population
boom, doubling between 1946 and 1971.
Already in the Sri Lanka story, by the 1960s, we have some
of the factors in Turchin’s model of instability – a country struggling
financially, and experiencing a youth bulge in the form of a giant population
boom that was to mature in the 1980s. Add to this the fall from elite status
experienced by Tamils in the government service, and the increasing insecurity
of Tamils in the face of a movement of Sinhala supremacy, and the causes of the
civil war are starting to line up according to the model that Turchin proposes.
On the way to the civil war, though, was another event that hints further at
the power of the structural demographic theory, one that I wasn’t really
expecting to find – the JVT rebellions.
A key piece of evidence supporting Turchin’s theory in
relation to Sri Lankan political instability is that the population boom, and
increasing higher education created a large population of educated young Sri
Lankans without graduate job opportunities. Some of these educated youth turned
to the JVP, a Maoist party that aimed to start a rural rebellion against what
it saw as a corrupt government and an inconsequential left-wing. In his video,
Rothman argues that it was anti-elite, indeed, that its leader, Rohana
Wijeweera, was not a member of the elite. But he was certainly what Turchin
would consider a counter-elite: he was a medical student, whose father had been
a member of the Sri Lankan Communist Party, and who had studied in Moscow.
Students formed the vanguard of the party, and in 1971, they instigated a
rebellion. Ultimately, the rebellion was crushed and Wijeweera was imprisoned,
but not after 5000 people had been killed by either the JVP or the government. Crucially,
the insurrection had little to do with ethnic tensions – the JVP was made up
mostly of Sinhalese and went on to be reborn as an ultra-nationalist party that
offered a return to Buddhist traditionalism when it rebelled again during the
1980s.
The whole episode demonstrates that Sri Lanka during the
1970s and 1980s was just such a politically unstable tinder box as Turchin
describes – large numbers of educated youth, embittered Tamil elites shut out
of politics, a growing population of economically immiserated workers and
peasants, and a struggling state, all contributed to a generation of political
violence. The LTTE Tigers were formed in 1972, a year after the JVP
insurrection, by Velupillai Prabhakaran, the youngest son of a government
official from a wealthy family in the Hindu religious elite. He’d become
radicalized when the government had made it more difficult for Tamil students
to enter university – further raising barriers to government service and elite
status. The Tamil Tigers were thus also raised by a counter-elite, not one
which wanted to replace the government, like Wijeweera, but to gain
independence from it.
Is this evidence?
This appears to me plenty of evidence that elite competition
for economic and political control of Sri Lanka was a major, if not the major
driver of instability leading to the civil war and insurrections. It appears
that there was an overabundance of Sinhalese elite aspirants who took aim at
Tamils in order to get a seat in government. Indeed, there were apparently so
many Sinhalese aspirants that some of them took aim at other Sinhalese through
the JVP. This ultimately led to Tamil elites organizing a brutal fightback and undertaking
a thirty-year civil war.
I’d expected to find some confounds to Turchin while
bingeing the history of Sri Lanka, but I actually found plenty of evidence to
show the strength of this mode of analysis – it would really be interesting to
use Turchin’s formulae to model post-colonial Sri Lanka and see if this period
of instability would be predicted even without reference to ethnicity.
Unfortunately, I think that might be beyond my current scope – given that it
would be difficult to get the data to examine the demographics of post-colonial
Sri Lanka in that kind of detail. If so, it would show that Turchin might be
right that ideology seems to come after demographics – a justification of the
often-brutal things counter-elites feel they must do to further their
interests.
Still, it is hard to deny that identity was a huge factor in
deciding who were the elites, and who were the counter-elites; but Turchin
wouldn’t disagree. Instead, it could be argued that when there is demographic
pressure, the fractures will form most readily along ethnic lines, given that
identity is a key motivator for the collective action necessary to govern or to
rebel. It somewhat blurs the elite/commoner distinction, though, when the
interests of both elite and commoner align in this way. And this is the
question that Turchin’s theory seems to pose for politics more broadly: how
well do the interests of elites and commoners align? Hopefully I will write
about the relationship between Turchin and political science in another blog.
Whereas in scientific terms, the story of Sri Lanka is
nothing like conclusive evidence for Turchin’s theory, learning about it
certainly persuaded me that a structural demographic theory of history is pointing
in the right direction – and worthy of more exploration.
Is Peter Turchin right? – a Sri Lankan case study
In my last blog, I gave a summary of the elite overproduction theory of political instability. Overall, I think it could be an important theoretical perspective on the cause of some of history’s worst conflagrations. But how much confidence can we have in it? What limitations does it have? Can it contribute to a political theory of a more harmonious and fair society?
It’s difficult to say how strongly the “structural demographic” theory is supported by the evidence. So far I’ve located only a scarce few papers that flesh out a potential mathematical model for analysing instability using demographics and analysing historical situations using the model. One of them was written by Turchin himself in 2013 and models the Antebellum US and the US in the 21st century. Another was written about Ukrainian politics around the Euro Maidan, a situation which Turchin also draws upon in his book, End Times.
Turchin’s paper “Modeling Social Pressures Toward Political Instability” is dense with formulae that aim to model the chances of political instability for a society based on data that he and other social scientists have gathered into the SESHAT project, a database of demographic information on historical societies. The math looks forbidding, but it is based on a few main assumptions. These are that the potential for instability is based on the mobilization potential of both elites and commoners; i.e. the number of politically and economically frustrated people exist in a society. This takes into account the number of young people (the most likely revolutionary foot soldiers), how many live in cities, and the relative wage, a measure of how wages levels track economic growth. Turchin assumes that when wages are depressed relative to growth, the elite are taking the surplus, and this contributes to an increase in their number. However, if this goes on too long, the elites start to feel pressure – there is neither enough surplus to go around, nor jobs in government capable of maintaining their ambitions for power. Thus, their mobilization potential goes up, and instability becomes more likely.
This is all very impressive, but despite some of its subtlety, it feels a bit like looking into the code for a strategy video game like Victoria 3, where the player has to provide the various castes of their chosen 19th century nation with enough of their needs to prevent rebellion while preparing for war and conquest with their rivals. That is to say, it actually feels really quite blunt. A lot more blunt than I was expecting. In End Times, Turchin defends the need for the models to be as simple as they can be without leaving anything important out, but this obviously leaves us with something that feels, not coincidentally, like a video game. However, I think Turchin would not mind this objection too much. It feels, especially in the heat of political action, that the human will is wild and free, and it is a bit disappointing to find out such expressions of freedom can be reduced to what looks like a game of Risk. However, all insight into human behaviour generates new possibilities, so perhaps oversimplification might focus attention on what is important in the political arena.
However, until there are more papers applying this model to different societies and crises, it’s difficult to say how strong the evidence is right now. Until then, a critique of it which involves criticizing its lack of depth or consideration of the hundreds of potential mitigating factors involved in political crises is more or less hot air. Of course, it misses out the rich detail of the factors involved in this or that war – it is supposed to focus on what makes up most of the variance; and in doing so, it will reduce the apparent importance of such factors as ideology or specific grievances.
The Sri Lankan Civil War
In an effort to see if it does make some sense, I decided to learn about the political instability in a country I previously had very little knowledge of: Sri Lanka. What little I did know about it is that it endured an almost thirty-year civil war fought on ethnic grounds, between the northern and eastern Tamils, and the majority Sinhalese. I remember the war ending in 2009, but beyond that, I had very little background in it. I seemed to me a good test case for the theory, because if it showed that ethnic tension was the root of the violence, this would act as a falsification of the importance of elite overproduction. Thankfully, I found a detailed, sensitive retelling of Sri Lanka’s history on YouTube, by travel journalist Alex Rothman here: https://youtube.com/playlist?list=PLhRLWBHRg9F2DlPmWNgeYR4pMlAOWi_dt&feature=shared.
Watching it, it really was tragic to think that such a beautiful place had been subjected to a generation of bloody violence; and I think this is one reason why research like Turchin’s is compelling – perhaps truly understanding the causes of such violence is a key to preventing it.
Without going into too much detail, the Sri Lankan civil war officially began in 1983 when a group of militants known as the Tamil Tigers (the LTTE) attacked a government patrol, and the majority party retaliated by instigating a massacre of Tamil civilians. The war then continued off and on until 2009, with the LTTE making regular attacks on government forces, even carrying out suicide bombings, in the hopes of creating a breakaway Tamil state. Other nations were also involved, most notably India, resulting in the assassination of a former Indian prime minister, Rajiv Gandhi, by the LTTE. Eventually, the Chinese intervened and provided the government with weapons it used to eliminate the Tigers and end the war. The UN estimated that 100,000 people were killed in the war.
It's easy to assume that this conflict was motivated purely by “ethnic tensions”. Many Sri Lankan Tamils resented their minority status in Sri Lanka, and were fearful of Sinhala persecution, with plenty of justification – there had been anti-Tamil pogroms on several occasions after independence had been gained from Britain in 1948. However, I find this explanation just too easy, even lazy – it feels like a “western” stereotype of global south conflicts such as Rwanda, or even European ones, such as Yugoslavia. However, there are plenty of examples of minority groups in ethnically segregated areas living peacefully next to majority neighbours in the same country. There are even cases where the right to self-determination of that region was denied, and this still did not lead to long term violence. Think of Scotland and Wales, or the recent breakaway attempt of Catalonia from Spain. Though the anti-Tamil pogroms certainly fueled the militant imperative for Tamils to demand their own state, we must also ask what sparked these “ethnic tensions” in the first place?
More than ethnicity
As a small, post-colonial nation, Sri Lanka unsurprisingly struggled to develop economically during the 20th century. A hangover from British administration was that there were a disproportionate number of Tamils in the government service, and this rankled the Sinhalese majority who wanted the state to act in their interest first. Their first strike was to institute the language act of 1956, replacing English with Sinhala as the language of official documents – making many Tamils in government who were unable to speak Sinhala obsolete. As Rothman notes, to get elected, Sinhalese politicians only had to appeal to the majority sentiment on the basis of shared identity, and this gradually increased Sinhalese power vis a vis Tamils in mid-century Sri Lanka, which was also going through a huge population boom, doubling between 1946 and 1971.
Already in the Sri Lanka story, by the 1960s, we have some of the factors in Turchin’s model of instability – a country struggling financially, and experiencing a youth bulge in the form of a giant population boom that was to mature in the 1980s. Add to this the fall from elite status experienced by Tamils in the government service, and the increasing insecurity of Tamils in the face of a movement of Sinhala supremacy, and the causes of the civil war are starting to line up according to the model that Turchin proposes. On the way to the civil war, though, was another event that hints further at the power of the structural demographic theory, one that I wasn’t really expecting to find – the JVT rebellions.
A key piece of evidence supporting Turchin’s theory in relation to Sri Lankan political instability is that the population boom, and increasing higher education created a large population of educated young Sri Lankans without graduate job opportunities. Some of these educated youth turned to the JVP, a Maoist party that aimed to start a rural rebellion against what it saw as a corrupt government and an inconsequential left-wing. In his video, Rothman argues that it was anti-elite, indeed, that its leader, Rohana Wijeweera, was not a member of the elite. But he was certainly what Turchin would consider a counter-elite: he was a medical student, whose father had been a member of the Sri Lankan Communist Party, and who had studied in Moscow. Students formed the vanguard of the party, and in 1971, they instigated a rebellion. Ultimately, the rebellion was crushed and Wijeweera was imprisoned, but not after 5000 people had been killed by either the JVP or the government. Crucially, the insurrection had little to do with ethnic tensions – the JVP was made up mostly of Sinhalese and went on to be reborn as an ultra-nationalist party that offered a return to Buddhist traditionalism when it rebelled again during the 1980s.
The whole episode demonstrates that Sri Lanka during the 1970s and 1980s was just such a politically unstable tinder box as Turchin describes – large numbers of educated youth, embittered Tamil elites shut out of politics, a growing population of economically immiserated workers and peasants, and a struggling state, all contributed to a generation of political violence. The LTTE Tigers were formed in 1972, a year after the JVP insurrection, by Velupillai Prabhakaran, the youngest son of a government official from a wealthy family in the Hindu religious elite. He’d become radicalized when the government had made it more difficult for Tamil students to enter university – further raising barriers to government service and elite status. The Tamil Tigers were thus also raised by a counter-elite, not one which wanted to replace the government, like Wijeweera, but to gain independence from it.
Is this evidence?
This appears to me plenty of evidence that elite competition for economic and political control of Sri Lanka was a major, if not the major driver of instability leading to the civil war and insurrections. It appears that there was an overabundance of Sinhalese elite aspirants who took aim at Tamils in order to get a seat in government. Indeed, there were apparently so many Sinhalese aspirants that some of them took aim at other Sinhalese through the JVP. This ultimately led to Tamil elites organizing a brutal fightback and undertaking a thirty-year civil war.
I’d expected to find some confounds to Turchin while bingeing the history of Sri Lanka, but I actually found plenty of evidence to show the strength of this mode of analysis – it would really be interesting to use Turchin’s formulae to model post-colonial Sri Lanka and see if this period of instability would be predicted even without reference to ethnicity. Unfortunately, I think that might be beyond my current scope – given that it would be difficult to get the data to examine the demographics of post-colonial Sri Lanka in that kind of detail. If so, it would show that Turchin might be right that ideology seems to come after demographics – a justification of the often-brutal things counter-elites feel they must do to further their interests.
Still, it is hard to deny that identity was a huge factor in deciding who were the elites, and who were the counter-elites; but Turchin wouldn’t disagree. Instead, it could be argued that when there is demographic pressure, the fractures will form most readily along ethnic lines, given that identity is a key motivator for the collective action necessary to govern or to rebel. It somewhat blurs the elite/commoner distinction, though, when the interests of both elite and commoner align in this way. And this is the question that Turchin’s theory seems to pose for politics more broadly: how well do the interests of elites and commoners align? Hopefully I will write about the relationship between Turchin and political science in another blog.
Whereas in scientific terms, the story of Sri Lanka is nothing like conclusive evidence for Turchin’s theory, learning about it certainly persuaded me that a structural demographic theory of history is pointing in the right direction – and worthy of more exploration.