Sunday, January 18, 2015

Blood for Oil

Here's an interesting one, oil causes crime in North Dakota. Seriously. Well sorta, but we'll get there.

First, check out this graph:

 You'll notice from the above graph that aggravated assaults seem to increase along with oil production in North Dakota.  This is a particularly interesting graph given that aggravated assault rates in the US have been steadily declining for the last 2 decades.

So what gives?

Well, here's the story as I have it in my head so far:

North Dakota has benefitted economically from the shale oil boom. The discovery and mining of the Bakken Shale Shelf from 2000 to 2012 saw an explosive growth in the number of barrels per day produced by North Dakota.

In near lock-step with this growth has been a huge growth in the aggravated assault rate within the state. The fracking industry runs on young males who do not have extensive family needs; the hours are long and the work is tough. Many companies provide corporate housing set on large campuses away from urban areas predominantly inhabited by these young workers. Additionally, the boom means that the jobs pay well and these young men find themselves with ample extra income to spend. It is likely that these elements create a sort of powder keg. Small towns with nothing to do, filled with young men who all have disposable incomes.

While not explicitly supported by the data, I conjecture that many of these aggravated assaults are bar fights. For further exploration would be alcoholism and drug use rates in these areas. A more granular look at specific areas in North Dakota where crime has increased would also help with analysis.

Aggravated assaults tend to change one year prior to changes in production of Oil. This lead-follow pattern indicates that Oil companies hire close to a year prior to reported changes in production. Note the inset: I offset the Daily Oil Per Well line by one year to better illustrate the pattern.

Kinda cool huh?

Of course, I am well aware that the graph is far from rigorous and all the data problems that go into finding a causal link. The graph poses an interesting question over how to plan for future economic booms.

Anyways, I got my data from the following places:

FBI Unified Crime Statistics - http://www.fbi.gov/stats-services/crimestats
EIA -  http://www.eia.gov/dnav/pet/pet_crd_crpdn_adc_mbblpd_a.htm
North Dakota Monthly Bakken Oil Production Statistics -  https://www.dmr.nd.gov/oilgas/stats/historicalbakkenoilstats.pdf