Interesting that Ca and Tx data nearly mirrors this simple analysis of racial makeup vs racial statistics on homocide.

    per_capita_data = {'white': 3, 'asian': 1, 'black': 21, 'hispanic': 6}
    
    cali_cencus = {'white': 37.7, 'asian': 14.8, 'black': 6.5, 'hispanic': 37.7}
    texas_cencus = {'white': 42.4, 'asian': 4.8, 'black': 12.6, 'hispanic': 39.1}
    
    def calculate(per_capita, state_stats):
        rate = 0
        for race in per_capita:
            rate += per_capita[race] * (state_stats[race] / 100)
        return rate
    
    print calculate(per_capita_data, texas_cencus)
    print calculate(per_capita_data, cali_percentages)

NYC had one of the lowest murdered rates in 2017, St Louis city nearly had as many homicides as NYC.

http://amp.kansascity.com/news/local/crime/article191211704....

Population of NYC: 8,537,673

Population of St Louis: 312,000

NY has restrictive gun-controls, is quite urban/dense and has the 2nd highest gdp/capita by state. They're all correlated with lower murder rates, that helps explain things a bit on NY.

The one thing NY does have is high inequality measured in gini, the highest of all states. That usually drives crime/murder rates. But it's probably because of rich outliers (extreme upper capital class), rather than a big gap between lower-uppermiddle.

Canada has far more restrictive gun control than NY, yet we are already on track to beat NY state in shootings this year with some cities having a 94% increase in violent gang offenses involving illegal handguns according to StatsCan. The difference must be policing and whatever gang strategy NY state has developed to get kids not to join them.

Could you reference me? Canada's murder rate is about 1/2 of the NY rate, and about 1/3rd of the US rate, and is about half of its peak 40 years ago. Canada has roughly the same amount of murders as NYS but the latter has 54% (let's say half) the population.

It's true that the rates are rising for a few years in a row, but this is nothing new. If anything, Canada's typical pattern is to see a rise a few years in a row, followed by a stronger drop. [0]

It's also true that gang related offences were part of the recent rise. But let's also note that they entail about 100 of the 600 murders in Canada, 15.5% in the last year for which I can find data. (2016). The vast majority of murders are done not by strangers but by friends or family, not related to gang violence.

[0] http://www.statcan.gc.ca/daily-quotidien/171122/cg-b001-eng....

Doesn't Venezuela have strict gun laws as well. I don't think private gun ownership is even allowed.

So how effective is that alone? It seems per capita income would be the prevalent factor.

About 95% of the world (not counting the US) has a lower per capita income than the poorest US cities.

Arguably the poorest US state, West Virginia, has a typical murder rate of around 3 to 4 per year per 100k.

Vietnam has been extraordinarily poor for the last half century, only recently beginning to climb economically. Its murder rate is typically 1 to 1.5 or so.

Per capita income probably only has a correlation in regards to the resources you have available to deal with crime problems if such presently exist, rather than being the defining characteristic of whether eg murder will be prevalent in a nation.

> About 95% of the world (not counting the US) has a lower per capita income than the poorest US cities.

Sure if the average salary in Vietnam is $150/month then yes per capita income is going to to be lower than the poorest cities in US. $150/month is a different story in NYC or SF. It would have to be adjusted for purchasing power or maybe just looking at poverty rates.

> So how effective is that alone? It seems per capita income would be the prevalent factor.

It isn't, alone, in a cherry-picked example. But ceteris paribus it's a explanatory factor.

I mean, one could also use your argument the other way around and question whether income is effective as a factor by itself. Cherry pick some country with similar levels of median wealth and a wildly different murder rate and gun laws. Like say the Netherlands and the US. Doesn't really prove a point, it'd be silly to now claim income isn't an important factor.