Opioids and the Labor Market

Do not believe those who tell you the only labor market problems have been demand side!:

This paper studies the relationship between local opioid prescription rates and labor market outcomes. We improve the joint measurement of labor market outcomes and prescription rates in the rural areas where nearly 30 percent of the US population lives. We find that increasing the local prescription rate by 10 percent decreases the prime-age employment rate by 0.50 percentage points for men and 0.17 percentage points for women. This effect is larger for white men with less than a BA (0.70 percentage points) and largest for minority men with less than a BA (1.01 percentage points). Geography is an obstacle to giving a causal interpretation to these results, especially since they were estimated in the midst of a large recession and recovery that generated considerable cross-sectional variation in local economic performance. We show that our results are not sensitive to most approaches to controlling for places experiencing either contemporaneous labor market shocks or persistently weak labor market conditions. We also present evidence on reverse causality, finding that a short-term unemployment shock did not increase the share of people abusing prescription opioids. Our estimates imply that prescription opioids can account for 44 percent of the realized national decrease in men’s labor force participation between 2001 and 2015.

The fact that the demand side blade of the scissors can be powerful does not imply the supply side blade does not matter, no matter how many snide tweets you may read to the contrary.

The paper is by Dionissi Aliprantis, Kyle Fee, and Mark E. Schweitzer at the Cleveland Fed.

Via Ilya Novak.


'Do not believe those who tell you the only labor market problems have been demand side!'


Sounds like good research.

I'm confused by this sentence: "Geography is an obstacle to giving a causal interpretation to these results, especially since they were estimated in the midst of a large recession and recovery that generated considerable cross-sectional variation in local economic performance."

Isn't the identification of causal factors usually helped by having different events happening in the various localities?

I'd say you are normally correct--but only if the events are random or at least not serial correlated. However, specific to this paper, they note Appalachia as a specific challenge for measuring the effect of the recession. For example, it is a rural region with generally low labor force participation and high opioid use. It's going to be relatively insensitive to the recession and recovery. Other regions might be highly sensitive.

In theory you could control for insensitive labor markets by finding an associated/correlated variable, but I think it is an identification challenge to find and understand which regional variation is persistent versus temporary.


"Even as seizures of the drug known as speed, ice and “ya ba” in its various forms reached a record high last year, street prices have dropped, indicating increased availability, said a report released by the United Nations Office on Drugs and Crime.

The agency said methamphetamine has become the main drug of concern in 12 out of 13 East and Southeast Asian countries, up from five a decade ago. The only exception was Vietnam, where heroin is considered the major problem."


Egads! A sudden worldwide problem. Not just the work force, the cartel money seeps into politics, governments go foul. This all impacts trade, enough to flip a minor downturn into a severe recession.

The reference to "prescription rates" clouds the issue. The drugs of choice are not available by prescription. Also the deaths from opioids are for the most part caused by fentanyl which is added to the illegal drugs so that the illegal drugs can be diluted for more profit and still give the addict a high. unfortunately fentanyl is so powerful that very tiny amounts can be fatal doses and the street dope seller is just not able to make accurate safe combinations of these drugs. For the most part prescription drugs are a second and desperate choice of addicts who for some reason cannot get their drug of choice. Prescription pain killers are not the porblem

"prescription opioids can account for 44 percent of the realized national decrease in men’s labor force participation" or....
decrease in men's labor force increase the use of opioids to get away of a reality whitout prospects.

I recall reading a piece a while ago about the increase in the percentage of the population receiving SSI disability. Small town doctors who were interviewed said they helped workers meet disability eligibility because they knew the workers had no other employment in small factory or coal mining towns where the industry had shut-down. Something similar could be going on here.

Negative marginal product men are incapable of good labor anyway. So mission accomplished! America will be a nation of winners when all of its losers die out!

This guy gets it.

But then who will we sell the opioids too? The collapse of the highly-addictive drug industry is a national crisis for the pharmaceutical industries, we need to find some drug to market to poor people so that they can get back on their feet, immigrant communities in the Southwest should do the trick this time!

Why opioids are prescribed that easily in first place?

The most interesting part of the article is that white adult men are the ones that get more frequently an opioids prescription for pain.

Legally, any doctor can prescribe any medication. They are licensed to it. The unwritten rule is that a cardiologist does not prescribes medication for a long term skin problem. The most recognized example is that "only" psychiatrists can prescribe psychoactive drugs, even though all licensed doctors can.

Half or more of the opioids are prescribed by primary care doctors. If they don't prescribe them, they may get sued. Perhaps pain should be a another specialization, where only fully trained people can prescribe them. That ensure the doctor knows what is going on and it would also relief lots of doctors of the legal risk and the pressure to prescribe.

Opioid prescribing is coming down at a pretty good clip the past year or two. CDC issued prescribing guidelines, which are helping. Also, some insurers and a lot of states have imposed new limits (e.g. no more than a 7-day supply to start) on opioid prescriptions. https://www.cdc.gov/drugoverdose/prescribing/guideline.html

It is generally true that any doc can prescribe any drug--with two major exceptions. Docs need special training and permission from the DEA to prescribe buprenorphine for addiction. That and methadone, which has its own special requirements, are by far the most effective treatments for addiction but a lot of docs don't have time or desire to get the training. The funny part is, these same docs can prescribe any drug for pain without any extra hassle.

"Why opioids are prescribed that easily in first place? "
Because the Sackler family is flock of vultures that loves blood money.

1) They work, and life without them is hell for people with chronic pain

2) The 'opioid epidemic' narrative is built on lies: https://reason.com/archives/2016/05/18/opioid-epidemic-myths

The link you share contains the following:

"The risk of addiction also has been exaggerated. According to NSDUH,
those 259 million painkiller prescriptions in 2012 resulted in about 2 million cases of "dependence or abuse," or one for every 130 prescriptions."

2 million cases of dependence or abuse per year, nothing to see here......sure? Of course the percentage is low, but the consumer population is large, thus the absolute number of deaths is deemed unacceptable.

I think only 2 kinds of people would point at the "one for every 130" instead of the 2 million: opioid manufacturers/sellers and technocrats (the derisive meaning of the word).

They seem to have listed both for good context. Also, it is abuse or dependance, not deaths. I think it's valuable to note for each abuse there is 129 people who use it as intended. Especially with all the press about an epidemic.

If you look at 130 random people doing ANYTHING — from snowboarding to smoking — you’ll find that somewhere between 10-30 are doing it excessively.

@Cytotoxic: very valuable context: 5-6% of cocaine users become addicts after 2 years of use. https://www.nature.com/articles/1300681

It can be said that opioids are safe, they're only 5 times less addictive than cocaine =)

Finally, this quote from the reason.com link. I wonder if the author really wanted to advocate for the safety of opioid painkillers:

-We lose sight of the fact that the prescription opioids are just as addictive as heroin," says CDC Director Thomas Frieden.-

Reason has previously written about how the dangers of illegal drugs have all been exaggerated.

"Half or more of the opioids are prescribed by primary care doctors. " And what percentage of SSRI drugs are prescribed by primary care doctors?

The real change in opiod prescribing is that these were once prescribed only for acute pain, or for the terminally ill (who, after all, will never need to de-tox) but now they are prescribed for chronic pain.

My gut reaction is that it's better that a hundred junkie's OD than that one patient is denied adequate pain relief.

Nonetheless, the bottom line is probably the extent to which medicine remains a profession and not just a business.

Sending everyone to a pain specialist for meds sounds like a great way to jack up medical spending even higher.

The opioid epidemic is simply a fact. One cannot deny a more than quadrupling of overdose deaths -- 70,000 in 2017 -- is an epidemic.

Are there a lot of people who take painkillers who don't get addicted and die from it? Of course, but that has always been the case.

That's silly logic. Quadrupling =/= epidemic.

Even if it is, it's not like opioids are contagious. There's no problem here.

These are massive numbers.

"Our estimates imply that prescription opioids can account for 44 percent of the realized national decrease in men’s labor force participation between 2001 and 2015."

Yes. And it should mean something for policy.

Revisit "welfare vs freedom" as you will.

Yes let’s make OxyContin controlled and hard to obtain. Then let’s send addicts to prison to increase their welfare. And make it easy to only obtain fentanyl, which will kill tens of thousands of them. But that’s better, because our intentions are more important than the hundreds of thousands of lives we ruin. After all, if someone is doing something we disagree with they should risk an overdose death.

Certainly they’ll be better off doing heroin in prison while suffering regular daily violence than taking an Oxy at home.

There are idiots who will become drug/alcohol/gambling addicts. There are mentally ill people who will kill themselves with guns. Warping the legal system to prioritize the 0.01-1% of people makes for terrible law.

What is the cause of the demand side negative outcome: is it the opioid abuse or is it the long-term unemployment that drives labor to abuse opioids (or alcohol)? Okay, blaming opioids for the abusers weakness (i.e., inability to resist opioids or alcohol) just provides an excuse for the abuser's behavior. But people, especially men, are ingrained with the notion that one's value (and, hence, self-esteem) is tied to one's job. In areas with high unemployment, spouse abuse and divorce rates go up and marriage rates go down. I recall the 1983 film Mister Mom in which Michael Keaton played an unemployed husband and father and is treated as a running gag. Why can't unemployed factory workers be like Michael Keaton?

Whoops, supply side negative outcome.

Random slightly-related anecdote. I was visiting a small Texas town on business. My business connection there also owned the local Taco Bell. He had just built it but he had to delay opening of it for 3 months while he searched for enough staff, as the locals were in great percentages failing the drug tests. And this in a town with unemployment off the chart. Such a vicious circle.

I realize this might seem like a dumb question, but why doesn't he hire people who fail the drug test? If they turn out to be poor workers I presume he can sack them.

'but why doesn't he hire people who fail the drug test'

Theft is probably the most straightforward answer. Addicts need money, and don't care how they get it - and a minimum wage fast food job will not provide the necessary funds. However, walking out with the cash from the register might do for a couple of days, or selling all the store's beef out the back.

There are definitely other reasons, but theft is straightforward, and a reason not to hire addicts in the first place.

How do you hold both this position and your first comment on this post regarding ZMP?

Because ZMP is a concept that Prof. Cowen would prefer you to believe in than snide tweets? https://marginalrevolution.com/marginalrevolution/2010/07/zero-marginal-product-workers.html

Lol Peak Prior, calling out someone for being snide.

-5 points to Gryffin_Clock for lack of self awareness

-10 points to Gryffin_Clock for not understanding how labor markets work

To be fair:

+8 Gryffin_Clock for the low skilled addicts will steal from you answer.

However, he could have gotten an extra +8 for also pointing out the liability issues involved with hiring an employee who has failed a drug test.

That doesn't seem like normal addict behavior. In my experience most addicts seem able to hold down a job. My current workplace is pretty unique in that no one I know is an addict. (But my addiction is pretty obvious. They tease me with it.)

And those are pretty effective tests if they can detect addition rather than just the presence of drugs.

'That doesn't seem like normal addict behavior.'

Depends on what you are addicted to - the U.S. is a big marketplace for cocaine, and that is not cheap. Basically, legal addictions are fairly cheap, but once you go on the black market, the prices go up - as does the variety of products available for sale.

I think if the people who want to work at Taco Palace are addicted to cocaine then then American lower class is doing better than I thought it was.

Could be a matter of corporate policy. He's a franchisee...he's gotta run it the way Yum Brands or whatever tells him to run it.

West Texas? Anyone who's not smoking marijuana for a couple weeks to pass a drug test is doing it to get an oil job, not to work at freaking Taco Bell.

"We also present evidence on reverse causality, finding that a short-term unemployment shock did not increase the share of people abusing prescription opioids"

What about long term unemployment?

And no mention of the Xbox? Tsk tsk.

Is it possible that what's making at least a chunk of these people unemployable is pain/injury that drove them to opioids in the first place?

Also, what idiot actually believes that the supply side doesn't matter? Who are these imbeciles?

Also, this just demonstrates why we need mass immigration like nothing else. Time for some able workers.

"Also, what idiot actually believes that the supply side doesn't matter? Who are these imbeciles?"

This is probably the first time that even the hint of demand side problems has been mentioned by the author of this blog.

Legalize it at the federal level. Opiates are one of the few drugs that actually works for pain when you actually need a drug for pain (tylenol is a joke). Cox2 inhibitors also work but they have also been demonized but they'll kill you less often than opoids--or tylenol. If people want to waste their lives high on drugs so be it. Many fewer would die if they could get their drugs legally at market rates (which would be dirt cheap).

One, tiny problem is that opiates are known to cause central pain sensitization. Prescribing them in great quantities not only will make it easier to kill people through respiratory depression, but that they will actually stop working for pain relief entirely. Having seen many patients who are addicts go through surgery and find that their body has so downregulate mu opioid receptors that the endogenous opioids no longer function.

As far as fewer people dying. Oh please. That was not what the statistics showed for alcohol. Nor what the preliminary data shows for marijuana. And frankly I cannot think of a single addicting drug with high rates of lethality that has killed fewer people with price reductions.

I mean I know the argument is that you will get cleaner, easier to dose substances with legalization ... it just seems to be more than completely offset by increased use by the marginal users who consume the vast bulk of any vice.

No one has ever died from a marijuana overdose.


No but plenty have died from MVCs while people were intoxicated on marijuana.


Extrapolating results from the early legalization on the rate of MVCs to the rest of the country suggests somewhere in the ballpark of 2000 people will die from marijuana intoxication during MVCs each year with full legalization. Around 500-700 of them will not be the ones intoxicated (extrapolated from all intoxication MVCs). Maybe 300-400 will be dead children.

So no, thanks, but rather than buy crap peddled by journalistic hacks from 2014 about a farcical claim (which the medical literature now disputes with case report level claims); I will go with one of the better research outfits on motor vehicle deaths that has a 60 year history devoted solely the most salient issue.

Certainly my personal experience involves plenty of death certificates involving people with reaction times shot to hell by marijuana who somehow decided to drive home and lost control of the car at speeds that proved lethal for someone (pedestrians, passengers, and drivers). But I am sure the tox panels the forensic pathologist ran were all just lies.

Wouldn't the effect on prices be an obvious way to show the difference? The minimum wage lower bound is a complicating factor, but did we see wage increases in nontradable sectors in opioid affected areas? I doubt it ...

That's fine. The jobs recovery, which appears at long last to have run its course, suggests that maybe a million or so people never came back to the labor force, a MUCH lower number than most economists were thinking five years ago.

Doesn't seem like it passes simple math tests.

Just pulling out my pencil and envelope, I think they are claiming that about 1/2 of all males that are both:

1. of working age and should be in labor force, and;
2. on long term opioid prescriptions,

fell out of the workforce permanently during this period. that doesn't seem reasonable at all.

Here are some commonly accepted facts:
66 prescriptions per 100 people in 2016
50,000 people per million on long term opioid prescriptions in USA (5% of population)
LFPR dropped around 4.5% between the years they studied

You can find these numbers anywhere, widely accepted to be true

Just fiddle with these numbers to account for olds who are out of workforce and any other material adjustments to the percentage of people who should be in the labor force. You'll end up with somewhere between 4-6% of people of working age are on long term opioid prescriptions.

It ends up where this study thinks that a gigantic amount of the people on long term pain meds at any point during 2000-2015 dropped out of the labor force and never went back.

This doesn't seem even remotely reasonable, and less so if you've actually worked with and know people on long term pain meds, like we all do.

Always first step in things like this, check to see if what you find makes sense. This does't seem to make sense.

You didn't finish the math. They are saying 44% of the decline in the LFPR is explained by opioids. By your numbers that is 2.25% of LFPR. You say long term opioid Rx is 5% of the total population. Where is the mismatch? I'd suspect if anything, those with long term opioid Rxs are predominantly male and working age (and rural). Seems like it would be pretty easy to get to some plausible numbers if you ask me.

I left it as an exercise for the reader and I don’t think you did your work very well.

These numbers mean somewhere around 1 in 10 people who have ever been on a long term script dropped out of the labor force and never returned. No way is that true.

This math says a truly giant percentage of the people on long term opioids dropped of the labor force permanently if the authors are to be believed.

They didn’t drop out for a bit - they exited and have never come back.

Labor force size 160 million in 2018. At LFPR of 62.5 that’s a cohort of 256m There is no way 2.2*2.56m= 5.6million people dropped completely out of labor force due to opioids.

It’s a giant number. Way, way too large.

This would be so obvious to everyone in the USA. We would each - all of us - have multiple stories along the lines of “well aunt Jeanie got hooked on pills and she stopped working at the age of 46.” It’s 1 in 50 people - so if youve ever heard stories about heart attacks, well you would have heard more about people leaving their jobs due to pills and never working again.

It would be such a significant risk of losing your job if you go on the pills - so obvious to everyone and certainly every doctor over that time frame. “I make this prescription for cronic pain and a year later 20% of them never work again”.

That’s what these authors numbers say and of course we don’t see anything like this at all.

The top 2% of drinkers, that I see get wheeled in on daily basis, consume, by their own admission, over 12 beers a day.

How many personal stories do you have of people who drink that much?

2-3% of Americans are problem gamblers. These are people, per the formal criteria, who have lost jobs, relationships, homes, and even their food money to gambling. How many people do you know who have been evicted for gambling? Per our statistician, about 1% of our catchment area has about 1% of people who are problem gamblers who have lost a spouse, home or job to their gambling disorder.

How many close friends or family meet that description for you?

The truth is that we self-segregate. Most likely you are educated to at least a college level. Most likely the vast majority of your friends are similarly educated. Most likely you make above average income. Most likely all your personal contacts make the same.

The truth is the bottom 1% of the population survives on something like $3 a day. Unless you work with homeless folks with poorly treated psychiatric issues, you most likely know none of them. And you most likely do not the worst 1% of alcoholics, the worst 1% of gamblers, or the worst 1% of violent people. All of these distributions follow a power law and their tails lead lives utterly unlike those of anyone who has regular internet access.

I could completely buy that 5 million people dropped out of the labor force to opiates. After all it is not just getting high, it is also the fact that you get fun diseases like Hep A, Hep C, endocarditis, TB, and the like once you make the transition from pills (expensive) to injectables (cheap). Your body doesn't much care if you have endocarditis because you got addicted 5 years ago and shared needles; heart failure will still make you unable to do manual labor. Likewise, if you sell sex for drug money, syphilis does not care if you stop using, the literal holes in your brain are still going to be with you. Likewise, if you are intoxicated and crash your car, get into a bar fight, or let an infection fester because you are intoxicated and cannot feel pain ... those with also stay with you.

Beyond that, being high is part of what makes being "disabled" or "unemloyed" tolerable. Getting hooked on anything is going to alleviate the tedium and social opprobium of being jobless. Changing social mores will allow snowball effects where people who are not drug users do not feel so bad about sitting home and playing video games.

So, no, I will be shocked if you can demonstrate personal connection to the action bottom 1% of any social pathology. Regardless you are confusing a stock and a flow; disability can easily continue after drug use stops. And lastly, it is not like things happen in a vacuum, the effects on the non-using friends and family members of users will amplify any changes.

In what alternate reality are rx pain pills, which are generally covered by insurance, including public programs, pricier than, say, heroin? After foot surgery I had an rx for Percoset (60 pills- I threw most of them out eventually) and the total cost to me was $10. And yes, millions of people are functional addicts who manage to hold a job even with a booze problem or pill habit.

The one any of the addicts I treated for ODing this week live in?

Do you think people jab needles into their veins rather than take pills for funzies? Street prices for liquid morphine are typically half the price of pills around here (literally the same drug and dose); shockingly the dealers charge more for a product that is more convenient even when ceteris parabis holds.

But that is atypical. Far more common is selling some admixture of heroine (50% more morphine-equivalents per weight than your oxycodone) and fentanyl (roughly 6700% more morphine equivalents per weight than your oxycodone). Fentanyl makes opiates cheap (and dangerous). Carfentanyl, used by some of the really cost effective dealers is roughly 670,000% more potent than your oxycodone.

This makes injection opiates MUCH cheaper to the end user than your oxycodone. Heroin is just diacetyl morphine, it hits the same exact mu opioid receptors as oxycodone, morphine, fentanyl, and carfentanyl. The homeless prostitutes I treat are not going to splurge on fancy pills, they are going to buy the cheapest opioid they can afford ... which is some heroine, cut with filler, and laced with fentanyl/carfentanyl.

I suspect you got those pills back in the heyday of "pain as a vital sign". Dosing like you describe is literally illegal in my state under current law unless the prescribing physician jumps through many, many hoops (of which I know of zero physicians who would as it would cost somewhere north of $1000 of physician time to clear; they would write for less divertable fentanyl instead).

I got those pills after surgery in late 2016- not exactly the Old Stoned Age. And I have a hard time believing that heroin is cheaper than a 10$ copay. Don't black markets generally drive up prices? What I will agree on is that heroin packs more of a punch than prescription pills (grantrd, I've only had the latter). Pills are beer to heroin's 100 proof liquor. So yes, I can see an addict's preference for it, and cost be damned.

opiate/oid addiction
disproportionately harms people
who don't graduate from college
hey paul Krugman
we actually knew that already

It should hardly be a shock that people who become zombies on drugs will be fired and have trouble finding new work. I saw this with a family member, who had a legit problem requiring treatment and probably would have ended up on disability anyway. Thankfully she's past that now, with the help of medical marijuana.

Now take this premise and apply it across other domains and factors that affect the supply side of employment.

Allow me to summarize. We tried very hard to control multiple confounding variables using imprecise, sometimes observational, data sets to conclude the following while ignoring the skyrocketing death rate from illicit opioid use. "Taken at
face value, our results suggest that solving the opioid crisis would substantially improve economic conditions in counties that have had high levels of opioid prescriptions by boosting the prime age male participating rate by more than 4 percentage points." This seems to suggest that making opioid abusers switch to questionable purity illegal alternatives will improve economic conditions for their community. Did I get that right?

"Our estimates imply that prescription opioids can account for 44 percent of the realized national decrease in men’s labor force participation between 2001 and 2015."

Is this in the same way that reduction in nautical piracy can "account for" some high percent of global warming?

Comments for this post are closed