Karachi, and the greater violence of New World cities

I like to ask some CWT guests (Charles Mann, Juan Pablo Villarino, and Alain Bertaud) why New World cities are so often so much more violent than Old World cities, including in Europe and Asia.  In the case of Asia, wartime episodes aside, it does seem that so many Asian cities are remarkably safe, especially for men but often for women too.

Recall one of the key principles of reasoning: look for the cross-sectional variation.

While Karachi is relatively safe now, before 2013 it had at least two decades of fairly extreme violence.

And what are some special features of Karachi history, relative to many other Old World and Asian cities?

The city had a very large “new” population, with Hindus (formerly the majority inhabitants) having left and migrants having come from many other parts of Pakistan, Afghanistan and Myanmar too.

The city was hit by a major wave of drug trafficking, heroin in particular.

The city was hit by a major wave of arms trafficking, run by thugs and mafias, often related to the wars in Afghanistan.

This is only one data point, but it supports hypotheses that higher levels of New World violence stem from relatively recent population shifts, drug trafficking, and arms trafficking.  When Old World cities have that blend, they too become quite violent.


Yes, but arms trafficking is not exogenous, unlike the other two. Arms trafficking usually results from drug trafficking or maybe a new population.

The hypotheses seem underdeveloped, which this is a blog post, so fair enough. However, it is difficult to look at nations that are rapidly urbanizing in Asia, like China but also others, and they still have low levels of violence. India has a low reported homicide rate, as does Vietnam, Malaysia, and Indonesia. My suspicion is that the population shifts are really a spurious correlation, if indeed there is a statistically significant correlation at all. The drugs and arms trafficking seem like the likely culprits, but of course you would still need to explain why you see drug trafficking in some areas but not others. In some cases I would suspect it just bad luck, more or less; the Andes are the best place in the world to grow coca, and the Central American Isthmus (including Mexico) is also a very logical place for logistics for distributing Cocaine to North America and other parts of the world.

Or Asians are more intelligent than the rest of us.

Is higher IQ associated with a higher future orientation? Murdering other people usually isn't a good idea in the long run (either you get caught, or if you live somewhere where the state doesn't catch murderers, you will eventually catch a bullet yourself).

Did someone say LA?

Mayor Garcetti plans to replace the entire population, he has invited the world's poor to come and get a free condo from LA city government, and vote twice. Quite a deal, and LA citizens are enthusiastic about being replaced.

Africa is Old World, no? Plenty of violent cities there and with raising populations, a likely shift in this proposed Old:New World urban violence model.

Africa is Old World, no?


I used to vacation in Mexico with my father frequently in the the 1960s-1980s, but when I went to previously pleasant Rosarito Beach south of Tijuana in 1996, I was struck by two things: how much richer the rich had suddenly become and how many security guards with automatic weapons were now standing around.

I presume this major change, which erupted into large scale violence in 2007, tracked to Mexico taking over the cocaine smuggling business from Colombia about the time of Pablo Escobar's death in 1993.

You're not supposed to provide data, that's not the MT way.

In LAPD novelist Joe Wambaugh's nonfiction book "Lines and Shadows" about Chicano cops in the Border Patrol in San Diego in the late 1970s, he portrays north of the border as more violent than south of the border. Mexico was a semi-dictatorship, and while most of its cops were peons, you could be sure that in Mexico there were a few cops who were smart and ruthless enough to keep the forces of disorder down.

Perhaps democratization undermined the old system? Or President Salinas got too greedy and unleashed chaos?

I suspect that in this century, however, if you speak Spanish and have family members in both the US and in Mexico and you aspire to a life of crime, you won't pursue your chosen career in America. Instead, there is more opportunity for bad men in Mexico.

So, it's now the opposite of the 1970s. But lots of influential people in the U.S. like George Soros are working hard to bring back 1970s style prison sentences, so maybe this flow will reverse itself once again.

Perhaps democratization undermined the old system? Or President Salinas got too greedy and unleashed chaos?

A thesis you cribbed from Mark Falcoff.

Mexico's homicide rate is quite unremarkable for Latin America, so unless you fancy every Latin American country was run by a PRI-like machine we couldn't see and they were all chucked out around the same time, it doesn't compute.

I’ve lived in a half dozen “first world” countries (7 if France counts). The least safe I’ve felt was any Nordic city on a Saturday night. Way less safe than my time in DC. Why? In the Nordic countries, when drinking in cities at midnight, on weekends, there are often multiple males who can punch you (or try) without — ceteris paribus — serious repercussions. In DC, they are potentially dead if they provoke the wrong person. Go figure.

The “knockout” game has claimed another life in the DC area. Violence is moving out of the drug trade and to just random violence as entertainment. https://www.unz.com/sbpdl/his-name-is-john-weed-59-year-old-white-man-murdered-by-two-black-teens-playing-knockout-game-at-fair-in-maryland/

Wow I wish that were verifiable, it plays to my prior that alcohol is very destructive, but I doubt that it's true that DC is safer, though I'd guess that Spain and Italy are safer than the Nordic cities because of less alcohol problems. I'd also guess alcohol is why Italy is way ahead of Denmark in life expectancy.

In DC it probably depends on the neighborhood. You're pretty safe in, say, Dupont Circle and other long gentrified areas.

Chris Blattman's papers, especially the recent Colombia focus - another country with large declines in violence - are quite useful in terms of analytical framework. Also Bates' book. Income and opportunity matter, breakdown of local government too. Then violence becomes an epidemic, like a strange zombie disease. Everyone around you is violent so you get drawn in, too. The epidemic eventually dies down if it can't find more hosts.

No recent population changes in the UK yet big spike in knife crime despite new blade restriction regulations. https://www.bbc.com/news/amp/uk-42749089

For the millionth time: UK =/= England and Wales. Knife crime in Scotland has dropped massively over the last few years.

Tirana-Durres is a major smuggling point for guns/drugs/cigarettes/etc. into the EU which has also seen major population shifts over the last decade but is extremely safe. The population shift has resulted in it being the least friendly city in Albania and given it an air of smug superiority but it's not lead to violence.

The level of organised crime is very high in Albania but there is very little spillover crime.

Organized crime is often negatively correlated with disorganized crime.

I suspect that strongly patriarchal societies, where the social and economic activities of women are constrained, tend to confine violence against women to within the home, leaving a surface impression to outsiders of relative safety in the cities. When women begin to achieve greater Independence and break down traditional restrictions, they face increased danger of urban violence (compare India, for example).

Not difficult answer to that, if one is honest. More inequality in the New World. Both the US and Latin America has white people at the top spot of their societies, and then an underclass of Blacks and Amerindians (the latter in Latin America, but you can replace this category for Latinos in the US). This creates a lot of resentment that ends up sometimes in crime and violence.

Funny, generally speaking, I resent criminals but not those richer than I am.

Also, I'd be interested in the whether the same things that make someone a criminal also make them part of an underclass. I don't doubt that being a criminal could make you a part of the underclass, but I do doubt whether being a part of the underclass makes you a criminal, especially one who victimizes members of that same underclass.

But I'd think that most likely, there's a third (or fourth, fifth, etc.) thing causing both.

I would suspect that you see higher rates of substance abuse, lower future orientation, lower intelligence, and higher scores in the “dark triad” personality traits among people who go on to commit crimes. For the most part, crime has a pretty low (negative) risk adjusted return, so people with enough intelligence and self control tend not to get involved in it. So a disproportionate number of criminals display the traits of “anti-social” personality disorder.

I would assume that you are suggesting that combination of high inequality within multi-racial societies and high inequality between races in those countries is what makes the new world unique. But then again you have Vietnam, Malaysia, and Indonesia, which all have quite low murder rates and highly racialized inequality. Granted, I think all of those countries have had anti-Chinese pogroms in living memory.

"This is only one data point, but it supports hypotheses that higher levels of New World violence stem from relatively recent population shifts, drug trafficking, and arms trafficking. When Old World cities have that blend, they too become quite violent."

Is this a reason to legalize all drugs without requiering a prescription? To me it points in that direction.
Remember Sam Peltzman's work on drugs.


The effect of requiring consumers to obtain prescriptions for pharmaceuticals on mortality is examined for a sample of middle‐income countries. In countries enforcing the requirement, infectious disease mortality is no lower and poisoning mortality is higher than in those not enforcing the requirement. A broader measure of government intervention—public expenditures on health relative to GDP—is shown to have moderately adverse effects on overall life expectancy.

What does the cross-sectional variation difference between Canada and the rest of the New World suggest? That plus the cross-sectional variation difference between Northern and Southern states lends support to chattel slavery as one difference. Nisbett and Cohen, in that general "cultural of violence" argument, also pointed to ranching as a contributor (linking that to Scottish Borders violence and other places), basically a "roving bandit vs stationary bandit" theory. One question is how long cultural differences can persist and be transmitted through generations, even to different immigrating populations and even if much of the original differences that spawned them are gone.

In other words, higher levels of New World violence can come from the original settlers, and be transmitted to immigrants. The USA outside Yankee states has had high violence for a long time.

Effective punishment is a real deterrent. Nobody steals where they might get their hands cut off. Because of political correctness and fear of being called a racist, most cities in US have decriminalized criminal behavior. The result is more violence. Socialism breeds violence. That's why Soviet police always carried sub-machine guns.

"higher levels of New World violence stem from relatively recent population shifts ..." which is why the US must allow unrestricted immigration. Have I got your logic right?

There's a large literature on this. I'm not an expert, but the big driver of large scale violence seems to be weakness or absence of the state coupled with struggle between multiple well funded groups for control. E.g. Brazilian police rarely go into favelas. Favelas were relatively safe until the rise of the drug trade provided funding for gangs. At this point, some gangs employ municipal managers for the neighborhoods they control, the PCC runs a small claims court system, and homicides drop precipitously when the PCC takes over a neighborhood. This only applies in areas where the PCC has uncontested control. In areas where it is competing with other gangs, homicides and other crime go up. There are several large gangs in Brazil and lots of competition, with lots of violence.

The Mexican state has historically been weak or absent in large parts of Mexico, both rural and urban. Crime was relatively low when drug gangs had co-opted the PRI and the PRI ruled the country. When PAN started winning major elections, gangs could no longer reliably protect themselves with bribery, and started building private armies. PAN is corrupt also, but transitions in political power disrupted established bribery systems. This led to violent competition with both the state and other gangs. Currently, areas where a single gang is dominant have low homicide rates, and areas where gangs are competing have high homicide rates. This applies in both urban and rural areas, eg Guerrero.

The CAR has rebel groups that compete with the government and try to control illegal mines for revenue. Where they succeed, they impose economically efficient taxes and economic activity in that area increases, But much of CAR is areas that no one group controls, and there's a lot of violence.

What these groups, including the governments, have in common is that they're not able to maintain long term control over an area. Violence in any area decreases as one group becomes dominant and then increases as the group losses power.

In SE Asia, states tend to be stable and powerful. There are no new political parties gaining control in Vietnam and China. States are strong enough to keep lid on inter-gang conflict. The Pakistani state is weak and violently competes with a well funded Taliban for power. Pakistani mafias are funded by drugs and arms sales.

European states have a long history of stability and power. The Austro-Hungarian Empire was already a police state in Beethoven's time. The really powerful gangs originate in places like Albania and southern Italy where the state is comparatively weak and there's a long history of organized groups competing through violence and funding themselves through smuggling and extortion.

Population shifts don't in and of themselves cause violence, but they can bring in new competition and upset an existing balance of power. Drugs are factor in that they provide funding. The Taliban funded its rise in Afghanistan by controlling the heroin trade. Arms trade provides both revenue and weapons. Competition between Asian gangs in Vancouver became more violent when they started smuggling in guns from the US.

But the big factors seem to be weakness of the state, well funded competition, and the ability of the state and its competition to disrupt each other and prevent any one actor from establishing long term control.

Comments for this post are closed