Scott Sumner  

Drug overdose deaths: It's not just economics

Adam Smith on Why Patents Are ... Oil Prices are Complicated...

Let me start this highly unscientific post with a caveat, I think it's quote likely that economic distress plays some role in the recent rise in drug overdose deaths. Here I'll try to caution people that the role might be less that you would think, as casual empiricism suggests that many well performing states have high rates of drug overdoses, and vice versa. If you look at the data on state fatality rates from drug overdoses for 2015, you do see some of the expected pattern. Booming North Dakota had by far the lowest rate, and depressed West Virginia had by far the highest rate. But after that things seem less predictable.

There are several ways of thinking about economic distress. One is inequality. Utah is the most equal state, then Alaska, Wyoming and New Hampshire. In terms of median family income, Maryland is the richest state, then New Hampshire. So the Granite State is in a sweet spot, both very rich and relatively equal. But its drug deaths exceed the national average. The three states at the top of the gini table (i.e. most equal) also have above average median incomes. So they can be viewed as economic success stories. Utah even has a religion with wholesome values. And yet all three have above average drug overdose death rates, with Utah coming in a particularly bad 47th place (where 1st is lowest rate and 51st is highest.)

In contrast, the 5 states with the lowest rate of drug deaths are all in the northern plains area. (Note that the 4 most equal states and the 5 lowest drug deaths states are mostly white, but their drug death rates are vastly different.) Interestingly after the 5 plains states you have Virginia, followed by four very unequal states, with lots of poverty; Texas, New York, Mississippi and Georgia rounding out the top 10 for fewest drug deaths. Why do those 4 unequal states have relatively low drug death rates? (Mississippi is also quite poor.) I would have expected exactly the opposite; I would have expected their death rates to be far higher than the four best states for inequality, where economic distress is far less severe. If you prefer unemployment, New Hampshire has a 3.1% rate and Utah has a 3.5% rate.

Again, this is very unscientific, but I suspect that the correlation at the state level between economic distress and drug overdoses is much weaker than you might expect, even if at the individual level it is somewhat stronger. If America could have an income composition like New Hampshire or Utah, most people on the left, and many on the right would consider it a dream come true. Rich and equal. The bottom 20% in New Hampshire may have the highest living standard that the bottom 20% have experienced almost anywhere in the world, at any time in human history. And yet if we are somehow miraculously able to make the US just like New Hampshire, is this what awaits us?

New Hampshire saw at least 385 drug deaths in 2015, according to the latest tally from the Office of the Chief Medical Examiner -- but the actual total could be even higher, as some 45 cases are still pending toxicology.

Toward the end of last year, state officials began warning that the state could end up surpassing 400 drug deaths total for 2015. Even if the current figure remains the same, however, that would still make 2015 the deadliest in recent years when it comes to drug overdose fatalities.

Of the cases processed so far, nearly two-thirds of the drug deaths last year involved fentanyl, either alone or mixed with other drugs.

Again, at the individual level economic distress probably matters a lot. But I suspect that tackling those economic issues at the macro level isn't going to make all that much of a difference for the drug overdose problem. Even Bernie Sanders couldn't make America as equal as Utah, nor could Jeb! make us as rich as New Hampshire.

Comments and Sharing

COMMENTS (13 to date)
E. Harding writes:

Yup. The big issue here is that this massive drug problem is largely an American problem. Do Portugal, Italy, Greece, Cyprus, and Spain have it?

KLO writes:

There are lots of confounding factors here. One of them is lthe rate at which individuals in the populations most affected by drug overdoses are insured. West Virginia, Kentucky, and New Hampshire have low rates of uninsured. Insurance provides greater access to drugs, including many drugs of abuse, and thus can increase the number of overdoses.

Some states with large relative uninsured populations like Texas and Georgia, have low rates or moderate overdose.

Insured rate and income is only moderately correlated with income due to the effects of Medicaid and the varying eligibility rules states have. Thus, a low income state such as West Virginia can have a high insured rate. A middle income state like Texas, with tight eligibility requirements, has the lowest insured rate.

Scott Sumner writes:

KLO, Thanks, I had not thought of that point.

Robert VerBruggen writes:

I played around with the data on this a little last month. Here is a chart where I compared the CDC overdose data from 1999-2001 with that from 2012-2014. These are age-adjusted numbers and for non-Hispanic whites only. Basically Appalachia really stands out -- WV, OK, and OH were in the middle of the pack around 2000 but now are out in front.

I also threw together a quick model that included white poverty rates (from IPUMS if I'm remembering correctly). It suggested that even after controlling for the 1999-2001 rate, white poverty predicts the current rate. I *think* WV was something of an outlier here (high white poverty, unusually large increase), but unfortunately I seem to have lost the stupid spreadsheet.

Jason writes:

KLO, could a potential mechanism be:
lots of insured poor people -> large incentive to cheat to get valuable perscription drugs -> large supply

gamma writes:

A physician pointed out that many overdoses are accidental:

1. A patient builds up a tolerance to a medication, and then runs out. If there is a delay in renewing medication (travel, doctor unavailability, insurance snafu), the patient's body loses its tolerance for the high dosage to which he is accustomed. If he receives the medication again, and resumes the dosage to which he was accustomed, that dosage is now an overdose.

2. A patient takes her meds and falls asleep. Upon waking, the disoriented patient can't remember taking her meds, and takes them again, close enough to the first dose to ensure lethality.

If this sort of accidental overdose is as frequent as my physician friend indicated, it could certainly diffuse the correlation between unemployment and overdose.

E. Harding writes:

@KLO, I totally expected that. My thought when thinking of the White Death focused strongly on "access to the healthcare system", which Blacks and Hispanics have less of.

John Thacker writes:

Here's another page with uninsured rates before and after Obamacare. Note that Kentucky, Ohio, and West Virginia all had large effects from Medicaid expansion. It's possible that the increase in deaths is related to that.

Tom Jackson writes:

Scott links to a drug deaths site with a firewall, but this seems to have the same numbers:

Robert Evans writes:


Both of the reasons you cite are arguments for electronically (or internet-enabled) controlled dispensing.

How much would such devices cost? I could imagine price tags of $50-$100 each, and the possibility of insurance covering it.

Hook one up to a webcam and the device could also "verify" that the person took their medication, and at what time.

Floccina writes:

Great post!


Again, at the individual level economic distress probably matters a lot.

I doubt that, rather what leads to drug addiction and death also leads to economic distress. I would bet on it.

Matthew Lubic writes:

All the proposed "remedies" for drug overdosing will undoubetedly
save a few lives... hopefully those whose ovedose was accidental. But as regards all those others... the ones who CHOSE to become involved with drugs in the first place... I resent my money being spent trying to save them from their own self-chosen behavior, regardless of how it is they chose to make that decision.

Yes, some were made suceptable by bad parenting: wanting to be part of something, an "in" group at school... whatever. This is where intervention would accomplish the most good; BEFOREHAND, not afterward. Put the word out to parents and potential parents that their children are THEIR responsibility; that a child's first one to four years are THE most important; the time when they will be imprinted with personal and social responsibility. Later? Too late. The die is cast. (How many parents reading this even know what the phrase means... what the "die" in this case means? Ignorance reigns supreme in our increasingly Sociopathic society.)

Nation-wide, hundreds of millions of dollars are spent every year or after-the fact drug interentions, dollars that could go towards helping the truly needy, especially in states like West Virginia where poverty is almost a way of life; towards education of parents whose awareness of their own responsibility AS parents is highly questionable.

Let the "druggies" experience the consequences of their own self-chosen, self-inflicted stupiity, even death itself. Give it a couple of yerars for word to get out to those contemplating the use of Heroin, pain killers, Marijuana, and whatever, and I estimate within three years we'll see a significant decrease in drug use and drug overdose.

Sound harsh? It IS harsh. But so too is the burden placed on society. It should be evident by now that the on-going, after-the fact interventions aren't helping but rather are ABETTING AND ENCOURGING drug use. "Why should I worry about overdosing if there's at least a possibility somebody will come along to save me in time? To save me from myself?"

It's time reality was faced for what it is, not what we wish it was. Wishing otherwise is the road to ruin.

Comments for this entry have been closed
Return to top