Extraordinary cash payments helped fuel a historic increase in drug and alcohol deaths.
Extraordinary cash payments helped fuel a historic increase in drug and alcohol deaths.
T he pandemic produced an abnormal number of deaths among people who did not have the virus. Through the first 15 months of the pandemic, U.S. deaths from drugs and alcohol were almost 50,000 above the corresponding months in 2018 and 2019 and 40,000 above the prior trend. Drug and alcohol deaths have never increased so much over such a time frame and may be the single greatest driver of recent declines in American life expectancy.
The people who died from substance abuse are a lot younger than people dying from Covid. One measure is life years, which counts each death according to the remaining years of life expectancy for people that age. More than 7 million life years were lost from drugs and alcohol during this time. For men, their part of the 7 million is more than the life years they lost from Covid.
For every drug or alcohol death, there are more than 100 people with an ongoing substance-use disorder, which is modern parlance for drug or alcohol addiction. For alcohol, narcotics, or meth, I estimate 47 million Americans have substance-use disorder, 8 million above the trend from before 2020. This change will contribute to worsening of several metrics of economic and social activity on an ongoing basis. Employment and labor-force participation are likely among them.
The supply-demand toolkit readily explains this. Our federal government was sending people massive amounts of cash, on average $30 per household per day. Normally incomes go down during a recession, but during the pandemic, personal incomes increased by record amounts.
We know from prior research that liquidity is quickly followed by elevated mortality from alcohol and drugs. This effect is especially strong when the liquidity is not coming with any incentive to invest in job-related human capital. Indeed, the pandemic payments were disincentives to work.
To quantify the income or liquidity effect as well as price effects, I consider alcohol, narcotics, and psychotropics separately. Narcotics deaths almost all involve opioids of one type or another. Deaths involving psychotropic drugs primarily involve crystal meth.
Alcohol has its own price changes, which I measured in a previous post. For illegal drugs I focus on the full price, which is the sum of a money price and an opportunity cost. Consuming dangerous drugs uses up time to acquire them, administer them, and recover from them. The full price therefore involves the value of time, which is where unemployment-benefit supplements come in.
Unemployment benefits have always been around, but the large federal supplements to unemployment benefits over the past two years were novel. These supplements were policy-induced shocks to the value of time, in addition to changes to the value of time coming from other sources.
The bulk of drug deaths involve opioids and methamphetamines, both of which I expect to respond to the pandemic liquidity and unemployment supplements. As I found in work conducted before the pandemic, opioid markets are more responsive because the expensive organic products that use the poppy plant, such as heroin or prescriptions, coexist with the synthetic products, such as fentanyl and its analogs.
Because of the substitution effect, opioid supply slopes downward in practice. An increase in opioid demand induces some organic consumers, or perhaps their dealers, to switch to cheaper fentanyl. It’s similar to the shift we saw during the pandemic from drinking alcohol at bars and restaurants to drinking at home.
The chart below shows the predictions of the economic model as a red line. It is the sum of the predictions from an alcohol model, a narcotics model and a crystal-meth model, each of which reflects prior trends and seasonal fluctuations, income effects, and price effects such as the unemployment supplements. The actual mortality is shown as red circles.
The model and data follow the same basic pattern. Both show deaths peaking twice at about 180,000 annualized deaths. Both predict one peak early in the pandemic and another in early 2021, because that is when the unemployment supplements were introduced.
According to the model, drug deaths between April 2020 and June 2021 were about 11,000, corresponding to more than 400,000 life years lost, above trend due to the substitution effects of unemployment bonuses. Substitution to home alcohol consumption explains another 7,300 deaths corresponding to more than 200,000 life years. Even though less than 1 percent of money from stimulus checks, the rent moratorium, and unemployment supplements went to spending on opioids or meth, the income effects still explain another 20,000 alcohol and drug deaths or about 750,000 life years.
Meanwhile, suicide rates fell slightly. These findings do not contradict observations that the pandemic elevated feelings of depression and anxiety in the population, but they do bring into question the thesis that alcohol and especially drug mortality during the pandemic were primarily driven by new feelings of depression or loneliness, since groups with high alcohol and drug mortality also had high suicide rates in the past.
Perhaps it’s all a coincidence that actual mortality resembles what we would expect from historical price and income effects. Other explanations for these patterns are welcome, but studying the incentives, which distinguishes economic understanding from public-health analysis, deserves serious consideration when it comes to causes of drug and alcohol mortality.