## Less than 1% of the city’s population accounted for more than half of its lethal incidents

Saturday, August 24th, 2019

Since Sept. 11, 2001, hundreds of Americans have died in terrorist attacks and mass shootings — while more than 100,000 have died from common street violence:

Urban violence accounts for most murders in the U.S., but politicians focus on everything except the violence itself, instead issuing sweeping calls to ban guns, legalize drugs or end poverty.

In a 2016 paper, my colleague Christopher Winship and I analyzed reviews of more than 1,400 studies on anti-violence programs around the world. We discovered that urban violence is sticky, meaning that it tends to cluster among a surprisingly small number of people and places. In New Orleans, for instance, a tiny network of less than 1% of the city’s population accounted for more than half of its lethal incidents between Jan. 1, 2010, and March 31, 2014. In Boston, more than 70% of all shootings between 1980 and 2008 were concentrated in less than 5% of the city’s geography. In almost every city, a few “hot people” and “hot spots” are responsible for the vast majority of deadly violence; the key to addressing the problem is to pay close attention to them.

I don’t think we’re supposed to look too carefully at these hot people and spots.

1. Aretae says:

The Pareto principle is multiplicative.

80/20
65/5
50/1

50% of insurance costs are 1% of the people
50% of customer service calls are 1% of the people
50% of the crime…

Etc.

2. Felix says:

Aretae, could you explain your arithmetic in simple terms? I don’t get the where and why of any multiplication.

Thanks.

3. Aretae says:

Felix,

Apologies for lack of clarity.

80% of 80% is 64 %. I rounded to 65. Sometimes to 2/3
20% of 20% is 4%. I rounded to 5.

80% of 80% of 80% is 51.2%. I rounded to 50.
20% of 20% of 20% is 8/1000. I rounded to 1%

If the pareto principle is:
20% of the inputs cause 80% of the outputs

Then pareto squared is that 4% of the inputs cause 2/3 of the outputs

And pareto cubed says that 1% of the inputs cause half the outputs.

4. Felix says:

Aretae, thanks. That’s very clear!

Hmmm. Gotta remember such a distribution may curve differently at the extremes than overall.