, 54 tweets, 16 min read Read on Twitter
Latin America accounts for 8% of 🗺️population, but 36% of homicides (2012)

Why do so many people kill each other?

Awesome new book by Professor Deborah Yashar @Princeton.
What explains temporal, regional, & subnational variation in Latin American homicides?

- Expanding illicit economies & high-stakes incentives for control of trade
- Weak/complicit states
- High competition between rival organisations

As someone who spends all their time theorising social change, I absolutely 😍this kind of comparative analysis.

Asking, does your theory of X explain:

- Change over time,
- Differences between countries, &
- Subnational variation?

If not, can you really explain X?
So, illicit organisations identify places with where state security forces are ineffective or complicit.

This changes over time.

And the drug trade obvs compounds corruption.
Importantly, the geography of the drugs route does not entail violence therein.

Homicides only occur when rival organisations try to control that lucrative drug route.
Why have homicides surged in Guatemala?

There were crackdowns in the Caribbean & Mexico.

Guatemala = weak/ complicit police & judiciary

Seeing these permissive conditions, state & non-state actors built up a drugs trade in G. Homicides peak in subnat places where they compete
So what happens when rival illicit organisations forge a truce?

El Salvador is one of the most violent places in the world.
But in spring 2012 its 2 main gangs made a truce, & the gov created 'peace zones'.
Cue immediate (though short-lived) drop in violence.
Nicaragua is one of the poorest countries in Central America:

Ranked last for pc income, youth urban unemployment, illiteracy, social spending.

And conveniently located for the US drugs trade.

So, would you expect a relatively low or high homicide rate?
Curiously, Nicaragua's homicide rate is much lower than its neighbours.

Why might this be?

Capable, dedicated, professional police, with high homicide clearance rates, & low corruption. Working with the community.
But how on earth is the Nicaraguan police so professional, while Guatemala and El Salvador's police services are widely noted for their inefficiency, incompetence, and corruption?
Departing from previous repression, the Sandinistas developed a revolutionary police service, committed to public service.

Or is that just lefty romanticism?
This revolutionary police cadre stayed on after the Sandinistas loss in 1990.
But got new leaders.
The police became more independent, less partisan.
Committed to law and order, & pro-actively working with the community

[NB. this is **RELATIVE** to Guatemala & El Salvador]
So unlike elsewhere in Central America, the Nicaraguan police tried to contain gang activity through prevention & pro-active community policing

Given this *relatively* stronger police presence, it appears less favourable to drug traffickers. So avoids the negative feedback loop?
30% of Nicaraguans think the police are involved in organised crime - so it's clearly not some fantastical utopia.

But this perception of corruption is lower in Nicaragua than elsewhere (save Chile).
SEGWAY: Many people are helpfully responding to this thread by highlighting that state violence has recently increased in Nicaragua. Yes. Absolutely. See:

But as social scientists, we can still ask, why was violence relatively low before this period?
Also, worth distinguishing between homicides in general and state-sponsored killings as a subset of this. Yashar's book is on the former.

I don't know how much the recent spike in state-sponsored killings affect the national average homicides? I defer to regional experts.
It's presumably quite tricky to count gang membership. Not exactly something you tick on the census.. But here are police chiefs estimates for the 2000s.

Again, why is gang membership relatively low in Nicaragua?

[Side note: this data for Brazil looks inaccurate to me??]
There were some nationally representative polls in Nicaragua in 2010:

85% expressed a positive opinion of the chief of police, Granera. She was the most popular public official.

50% expressed a positive opinion of the police. It had wide(ish) public support.
In comparison to Mexico, El Salvador, Guatemala, & Honduras,

the Nicaraguan police (before 2014) embodied an "esprit de corps (one where individuals join for a collective cause and are sanctioned when they violate it)".

Young police officers emulated public-minded superiors
2014: the Nicaraguan legislature approved a law moving the police so that it was no longer independent, but under direct orders of the president (for the first time since 1979).

2018: President Ortega used the police to crush demonstrators.
So the Sandinista revolution spawned a public-spirited police, which gained independence after 1990.

Positive feedback loop: more competent, less corrupt.

So the drug trade did not flourish - relative to M, El S, & G.
And the police did not get so corrupted by drug money.
But as we saw in 2018, institutional path dependence and positive feedback loops can be undercut by changes at the top.
So we see two pathways:

(1) Countries with corrupt, weak police, impunity. Permissive operating environment for drug gangs, who fight over turf, & bribe police. Negative feedback loop. Chicken and egg problem - H, El S, G, M.

(2) Stronger police, positive fb loop - Chile, CR, N
So, how to switch from one pathway to the other?
I 😍comparative politics: examining several cases together, we see what's important, in generating differences & similarities


Hungry for another way to understand social change!

So what about web-based social science & computational models of networking?

Does behaviour spread just like information & infectious disease contagion?
Via weak ties?

No, says Centola

People may resist innovations that threaten etb practices & hierarchies
So key isn't to increase info or lower the costs
But change the network
Once I have a cold or share news, my contacts get a cold & hear that news

But new behaviour is different

1) New adoptees must want to accept it
2) Acceptance depends on their beliefs about what others do
3) This requires exposure to multiple sources, seeing many others changing
Adopting a new behaviour is costly
Others may disapprove, & reprimand you
So, rather than unilateral deviation, individuals heed their peers' behaviour.
Behavioural change reqs:

- Strategic complementarity: Twitter becomes more fun if our chums are users👯‍♀️

- Credibility: verification through multiple confirming sources 🤓

- Legitimacy: we check to see if our pals approve 🤗🤗

- Emotional contagion: EXCITEMENT SPURS ACTION! 🤩
Most work on social norms focuses on "beliefs": i.e. (2) and (3) above

So, i 😍 Damon Centola's focus on EXCITEMENT!

[Tangent: if you're waaay into this groove, also check out Dr Roni Porat's work on emotions & social norms scholar.google.com/citations?user…]
Refs on social norms and EXCITEMENT! 🤩

[this may surprise precisely no one but i am pretty convinced by the contagious power of enthusiasm]
Right on.
To test these ideas, Centola ran web-based computational experiments:

Comparing random & clustered networks (ie. weak & strong ties)

Dark lines show informational diffusion.

Info diffuses rapidly in both networks.

But behaviour changes more rapidly in the clustered networks
Participants were 67% more likely (p < 0.001) to join the forum after receiving a second invitation than after receiving only one invitation

32% more likely after receiving a third invitation

- Credibility &
- Strategic complementarity.
That's a neat experiment. But what about RL?

In authoritarian regimes, radical dissent is shared through trusted networks [strong ties]

Dissidents don't mouth off to people they don't know [weak ties] - given risks of repression

This curbs change - as Hannah Arendt argued
If people are scared about repression, they only express dissent within their small, trusted networks [strong ties].

So other groups [with whom they have weak ties] don't hear or know about their dissent

Cue pluralistic ignorance [my favourite phenomenon!].
OK! Suppose I want to launch my evil empire... 😈

How do I use these insights to encourage behavioural change, & overcome resistance?


[Recognise that individuals are not solitary islands, but influenced by their peers].
Centola runs a bunch of computational experiments,

We see that behavioural change endures iff there's network clustering:

Enabling social innovators to reinforce each others' early adoption of new behaviour.
Centola experiments on his students: gym classes & online platforms

1) Classes & 💰indiv rewards
2) Classes, rewards, & comparison to anonymous others 🏋️‍♀️🏋️‍♂️

3) Classes, online chat, & team rewards
4) Classes, online chat, comparison to other teams, & rewards

Which does best?
The students were more likely to go to the gym if they saw other teams' scores.

Surprisingly, the team score did NOT foster free-riding

Curiously, 'social support' produced the WORST outcome!!

Non-gym goers create inertia, legitimising non-attendance, others stop too!!
Curiously, I think this echoes Yashar's point about the police in Central America.

If your colleagues are corrupt, frustrated, despondent, lacking commitment to public service, this curbs investment & innovation.

So how to build espirit de corps, when no one has it, or sees it?
Here I'd flag Liberian & Gambian civil servants' observations of PDIA:

Identifying problems, collectively addressing them, securing small wins, builds collective efficacy: they see they can do it, together


In 1925, 1 in 23 Americans was a member of the Klu Klux Klan.

What are the parallels between the rapid rise of white supremacy then & now?

My third and final book of the weekend.
1920s: White, Protestant small-scale producers felt threatened:

- Waves of immigration in early 1900s
- African Americans -> Northern cities
- More women entered the labour force, & gained the vote
- Low-skilled labour & mass production
- Small-scale producers couldn't compete
In 1926, Du Bois wrote:
Small-scale producers couldnt compete w/ the rise of MACHINE PRODUCTION! 🤖

1920-1922, unemployment rose from 5 to 12%

Many Klansmen were small-business producers, or low-level white collar workers

[But economics isn't everything; some were managers/ professionals]
Besides the rise in mass production, there was also a collapse in foreign demand for US agricultural produce.

When Historian Nancy MacLean looked at the financial records of Klansmen, she found > 50% had suffered 💰 losses between 1918-1927
White supremacy in the USA - 1920s and 2010s..

States ripe for the KKK:

- States with a sudden spike in mass production, industrialisation, displacing small-scale producers
- Agricultural economies producing farm goods, no longer sought by European markets
This is new to me, though will surprise precisely zero gender & politics scholars

Trump did well in counties where relatively few women worked.
White supremacy rises in response to perceived threats to economic, social, & political entitlements - McVeigh & Estep.

[they don't go down the binary 'is it economic or cultural?' route]
Echoes this fab article in @SF_Journal

When white Americans think that minorities are thriving (politically/ economically/ numerically), they become more opposed to welfare benefiting minorities

Wetts & Willer (2018).
Likewise in India, poor people (born into historically privileged Brahman castes) are more likely to vote for caste-identities [rather than class-based redistribution] when they feel that the caste is threatened.

Pavithra Suryanarayan (2018) in CPS.
So why did the KKK's appeal rescind?

Some say they lost moral authority after the leading Klansman's rape & murder of a young woman.

But @joshuarothman suggest a structural explanation: life got rosier, & anxieties subsided.

Missing some Tweet in this thread?
You can try to force a refresh.

Like this thread? Get email updates or save it to PDF!

Subscribe to Alice Evans
Profile picture

Get real-time email alerts when new unrolls are available from this author!

This content may be removed anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!