Separate Problems

The FBI has released its “Crime in the United States” stats for 2016. While there’s been significant discussion of whether and how the 22% rise in homicides since 2014 (and significant rise in violent crime) arose from a “Ferguson Effect” in cities with Black Lives Matter demonstrations following a police killing, another question is how closely the rise in violent crime has followed the opioid crisis and expansion of heroin use across the country. As Robert VerBruggen noted last year, there’s suggestive evidence tying increased cartel activity associated with heroin distribution to increased homicide rates by city, in recent years. It looks to me, however, like this does not translate over to an association with overdose rates:


In fact, the “Obamacare effect” appears to be associated with a lower increase in violent crime at a state level, although the difference is not statistically significant:



In particular,  Medicaid expansion states with large recent increases in overdoses have had lower increases in homicides than holdout states, or than Medicaid expansion states with smaller increases in overdose:


At the very least, this suggests that the Obamacare effect on overdoses is not just picking up a general social deterioration and increase in lawlessness but something specific to opioids and drug use as well as (perhaps) interaction with the medical system. More speculatively, I wonder if Obamacare has encouraged a segmentation of the market for illegal opioids, with the Medicaid expansion states facilitating the kinds of symbiosis between small time, low-violence Xalisco suppliers and pill-supplying medical providers that Quinones profiles in Dreamlandwhile the holdout states are more often scene to the higher violence, cartel-based distribution that Don Winslow wrote about in Esquire last year.


This may, however, be overthinking the issue. The simpler explanation for these patterns is still, probably, the Ferguson effect– the states with large recent increases in violent crime are almost exclusively those with a larger black populations; reduced policing in cities either in direct response to BLM-related demonstrations against police killings or preemptively due to worse police-community relations seems like the simplest and most plausible explanation for the recent rise in murder rates:


The New York Times earlier this year quoted one sociologist saying that pull-backs in policing in response to demonstrations were “actually a good thing,” but the marginal result at this point appears to be literally thousands of additional murders.

UPDATE: For what it’s worth, an OLS regression suggests that these are complementary rather than competing explanations: conditioning on % black and baseline murder rates, murder increases have been worse (larger) in states where overdoses were bad in 2013, but have been smaller in states where overdoses have increased from 2013 to 2015.

(1) (2) (3) (4) (5)
VARIABLES Change in Murder Rate 2014-2016 Change in Murder Rate 2014-2016 Change in Murder Rate 2014-2016 Change in Murder Rate 2014-2016 Change in Murder Rate 2014-2016
Murder Rate in 2014 0.00256*** 0.000840 0.0132*** 0.0123*** 0.0113***
(0.000950) (0.000766) (0.00251) (0.00238) (0.00216)
% Black Population 0.166*** 0.0990*** 0.107*** 0.118***
(0.0269) (0.0255) (0.0242) (0.0220)
Population 2016 -0.561*** -0.516*** -0.475***
(0.111) (0.106) (0.0957)
Overdose Rate in 2013 0.0995** 0.139***
(0.0377) (0.0359)
Change in Overdose Rate 2013-2015 -0.183***
Constant 4.175*** 2.859*** 3.722*** 2.034*** 1.861***
(0.416) (0.378) (0.350) (0.720) (0.649)
Observations 50 50 50 50 50
R-squared 0.131 0.520 0.692 0.733 0.789

Standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Here’s another way of looking at the last regression column graphically: when you control for baseline murder rates, baseline 2013 OD rates, population, and % black, the states with increases in ODs had lower increases in murder rates:


This is at least not contradictory to the idea that when cartels move in, they increase the price along with the violence, making overdoses more difficult, while a more libertarian approach to drug distribution increases overdoses while reducing violence. I may be omitting some obvious confound here though, of course.

2 thoughts on “Separate Problems

    1. Sometimes WordPress handles tables weirdly, and you might have to scroll around a bit to see the last few columns. I also added a graph showing the coefficient on the change in OD rates to explain what I meant.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s