The danger of the linear-thinking fallacy and the diminishing marginal utility of alcohol consumption

Unlike most elites, Charles Wheelan freely admits that he was “born on third base.” Although he probably inherited less than celebrities like Donald Trump or Bill Gates, Wheelan was born into a stable family that could afford to send him to an elite high school and then an ivy league college. When you are born in privilege like these people, there aren’t many things that can knock your life out of privilege, but illness is the most common reason. The illness that should be the easiest to prevent is the one that killed Donald Trump’s older brother: substance abuse. As Wheelan said,

…alcohol abuse is one of the things that can seriously derail the lives of otherwise privileged people… at Dartmouth and most other college campuses, binge drinking is the most serious social issue. 

The Harvard Study of Adult Development has followed Harvard sophomores for 70 years starting in 1937. It found that alcohol abuse was the most common way for this privileged group to lose socioeconomic status.

This is because most youths (especially wealthier youths) are not very rational in their use of alcohol.  People in their late teens and early 20s have the highest hazard rate for alcoholism, so it is a problem, but there are ways to moderate the risk.  First, people should only use the minimum effective dose for any drug.  Many college students still haven’t figured out that more is not always better.  Binge drinking shortens life expectancy for all ages and increases risk of alcoholism.   Alcohol companies love drinking games that result in binge drinking because they rely on alcoholics for most of their revenues.  Everything eventually hits diminishing marginal utility including alcohol, but many youths get too drunk to notice when drinking more is producing more harm than benefit.  Lifehacker recommends making a plan while sober to limit consumption to two drinks because that reduces risks of harm and it is easier to avoid accidental overconsumption if you make a plan while sober than if you wait to decide until after your judgement has already been compromised by a couple drinks.  This is part of the “mindful drinking” movement.

Youths are the most prone to mindless drinking of any age group. They have more problems with binge drinking and its problems than people over age 25 partly because of the linear thinking fallacy.  This is the overwhelming tendency for people to extrapolate from an experience and think that because one drink feels good and two drinks feels better then ten drinks must feel way better than two drinks.  That fallacy leads to vomiting, blackouts, injuries, violence, and six Americans dying of alcohol-poisoning every day. 

This is the mental model many youths unthinkingly have in their heads:

This is because the easiest way to think is linear when in reality, all consumption, including technology, eventually exhibits diminishing marginal utility as consumption increases. The real relationship between drinking and fun looks something like this:

This is the standard medical model for describing the effect of any drug, known as the dose–response or exposure–response relationship. Contrary to the bunk of homopathy, as any drug dose approaches zero, it’s clinical effect also approaches zero. For example, when someone drinks a quarter of a beer, his brain won’t feel much effect of the alcohol (although there can be a significant placebo effect). Then, there is increasing marginal effect as the dosage increases and the next half beer can have, perhaps, ten times more psychoactive effect.

But as with any drug, diminishing marginal benefit kicks in as dosage continues growing and the positive clinical effect eventually plateaus. Unfortunately, the negative side effects generally continue to rise after positive effect stops growing which causes the net benefit (labeled “fun” in this case) to decline with excessive dosage.

The scaling will differ for different people, but for most people a third drink gives little immediate benefit over two drinks and should be avoided because it causes long-term health problems that probably exceed whatever short-term marginal benefit there may be. And by the fifth or sixth drink, many youths can expect to regret a hangover, vomiting, elevated accident risk, and social regrets. A few more drinks beyond that can lead to immediate death from alcohol poisoning, so the marginal utility of alcohol clearly becomes negative at some point when people drink too much. An average of six Americans dies from alcohol poisoning every day.

Most youths are too inexperienced with alcohol (and other drugs) to realize that the benefits stop rising with increased consumption whereas the costs of drinking continue to rise exponentially. And once people get inebriated, they are no longer good at making rational decisions and analyzing concepts like the diminishing marginal utility of alcohol.

Addiction theory advises against routinely using any addictive drug and against bingeing on large does because these behaviors increase the body’s desire to consume more partly because they shift the curve to the right and make small doses less effective while increasing the desire for larger doses.  That is addiction. Days of fasting between doses (and smaller doses) help reset the body’s defenses and reduce tolerance so that small doses recover their effectiveness.  

My main worry with marijuana legalization is that it will make marijuana use routine enough to cause a large increase in addiction.  Right now marijuana rarely causes addiction, but that might be due to the fact that very few people ever use Marijuana enough times to pass the threshold where it alters their brain structures enough to produce addiction. In the past, very few Americans have ever used it 100 times in their lifetimes and that was partly because it was illegal, and so most people never found enough space in their lives where it was socially acceptable to use it routinely enough to become addicted.  Of the people who have used marijuana more than 100 times, it has higher self-reported rates of problems than for people who have used alcohol more than 100 times, so as casual usage patterns change, we might discover that marijuana is much more addictive than alcohol. The research just isn’t clear yet.

Plus, the hazard rate for addiction is much higher for youths than for older adults.

“The younger you are, the more vulnerable your brain is to developing these problems,” Dr. Levy said. Youths are also more likely to become addicted when they start using marijuana before the age of 18, according to the Substance Abuse and Mental Health Services Administration. Furthermore, there is growing evidence that cannabis can alter the brain during adolescence, a period when it is already undergoing structural changes.

It is clear that higher doses of marijuana are more likely to produce unpleasant side effects.

Furthermore, the potency of marijuana has been skyrocketing. In 1972, the average THC content of marijuana was less than 1%. Now it is MUCH higher:

A study by ElSohly et al., published in Biological Psychiatry in April 2016, looked at changes in cannabis potency from 1995 to 2014… The team found that marijuana had an average potency of 4% THC in 1995…. [They] found a noticeable increase in the amount of sinsemilla found in the cannabis they tested in the early 2000s. By 2014, the average THC content had risen to 12%… Today, it is normal to find marijuana with a THC content of over 20%. There are samples with a recorded THC level of over 30%, which is insanely high… Apparently, the average THC in Colorado weed is 18.7%! Expert Jonathan Page from the University of British Columbia says he doesn’t pay much attention to strains with an apparent THC of over 28%. In his opinion, these results are either faulty or based on the fact that breeders add extra cannabinoids to the strain to make it stronger.

In case you didn’t understand what it means to add extra cannabinoids, that means that chemists are cheaply synthesizing novel compounds that are cousins of THC and breeders are adding them to marijuana in order to get more bang for the buck. There are thousands of types of artificial cannabinoids and nobody knows what effect they will have. Plus, illegal drug dealers are doping their dope with (hopefully) trace amounts of fentanyl to cheaply increase it’s kick. Some marijuana products advertise THC levels exceeding 95%.

Furthermore, CBD levels are dropping at the same time. According to one marijuana aficionado, the best ratio of THC:CBD is 1:1, but that is rare today:

the THC:CBD ratio went from 14:1 to 80:1 within 20 years. This has major repercussions when it comes to analyzing the potency of marijuana.

The higher the THC:CBD ratio, the higher their addictiveness. Today’s ‘marijuana’ products, “are as close to the cannabis plant as strawberries are to frosted strawberry pop tarts.”

Nobody in the public health establishment feels that alcohol consumption has any possible net health benefit for college-age people, but for people over age 65, the costs and benefits are different and there is some disagreement.  Alcoholism is going to have a much smaller harm after retirement. One the predictions of the theory of rational addiction is that old people should get addicted at high rates because their life expectancy is so short that it doesn’t matter as much if alcoholism eventually messes up the rest of their lives.

Small amounts of alcohol can even protect against cardiovascular disease which is the main killer of old Americans.   This is the kind of research that alcohol companies love to fund, so you might take it with a grain of salt.  But the heart-protective benefits of alcohol only work for people with a significant risk of heart disease.  People younger than 30 don’t have significant heart disease risk, and even moderate alcohol consumption raises the incidence of lots of other illnesses including cancer, so there is zero net health benefit for young people.  Accidents are the main cause of death among college students and alcohol is the biggest contributor.   Although alcohol use is down considerably among American youths, older people are abusing more.  German Lopez reports that alcohol deaths are a large and growing problem:

According to new data from the Centers for Disease Control and Prevention, after controlling for age, the alcohol-induced death rate reached 8.5 per 100,000 people in 2014, up from 7.1 in 1999 and 7 in 2006.
As a result, nearly 31,000 people died by alcohol in 2014, up from 22,000 in 2006. That means more people died to alcohol in 2014 than the nearly 29,000 who died from opioid — including heroin — overdoses, but fewer than the nearly 34,000 who died to gun violence or car crashes that same year.
Still, the 2014 data likely under-counts alcohol-related deaths, since it only includes deaths induced directly by alcohol, like liver cirrhosis. It doesn’t include deaths from drunk driving, other accidents, and homicides committed under the influence (alcohol is linked to 40 percent of violent crimes). Counting those deaths, alcohol’s death toll in the US reached 88,000, according to the CDC — and that’s before accounting for the recent rise in alcohol-induced deaths shown in the chart above. All together, this puts alcohol behind only tobacco, which is by far the deadliest drug in the US, in terms of total drug deaths.
...Why are alcohol-induced deaths on the rise?
…For one, Americans are drinking more. According to the latest National Survey on Drug Use and Health, the number of Americans who reportedly drank in the previous month slightly increased as alcohol-induced deaths did: from 51 percent of all persons 12 and older in 2006, when deaths began to climb, to 52.7 percent in 2014.
…But why are people drinking more? …alcohol is more affordable than ever: According to a 2013 study…as a result of rising incomes and falling prices, alcohol is more affordable than it has been in 60 years. And …most US alcohol taxes aren’t indexed for inflation.
…[Secondly, the] opioid painkiller and heroin epidemic could have made alcohol deadlier. When taken together, alcohol and opioids interact to intensify each other and make each other more dangerous. So it’s possible that alcohol deaths rose as more people used — and died from — opioids throughout the 2000s, and many of those deaths [may have been] counted as alcohol-induced deaths in the CDC data.
Another factor could be greater use of the anti-anxiety medication, benzodiazepine. …these drugs were prescribed more and more throughout the 2000s. Since benzos can intensify alcohol’s effects, it’s possible they [caused] more alcohol-induced deaths in the same way opioids might have.
…Many, many studies, for example, have found benefits from a much higher alcohol tax. A recent review of the research from David Roodman, senior adviser for the Open Philanthropy Project, made a case for a higher alcohol tax:
“[H]igher prices… cause less drinking. A rough rule of thumb is that each 1 percent increase in alcohol price reduces drinking by 0.5 percent. …I estimate …a 10 percent price increase would cut the [cirrhosis] death rate 9-25 percent. For the US in 2010, this represents 2,000-6,000 averted deaths/year.”
…So for the US, boosting alcohol prices 10 percent could save as many as 6,000 lives each year. To put that in context, paying about 50 cents more for a six-pack of Bud Light could save thousands of lives. And this is a conservative estimate, since it only counts alcohol-related liver cirrhosis deaths — the number of lives saved would be [almost three times] higher if it accounted for deaths due to alcohol-related violence and car crashes.

This estimate of the social cost of alcohol is still too conservative because it doesn’t count alcoholism-related crime, accidents, lost productivity, and family breakdown.  Alcohol also causes a lot of domestic abuse that is left out of the numbers above because it falls short of murder.

A 10% increase in alcohol prices causes about a 5% decrease in alcohol consumption.  That is because alcohol is addictive and so its demand is inelastic.  Even so, a mere 5% reduction in alcohol use reduces the number of alcohol deaths by a large amount (up to 20%) because of the diminishing marginal utility of alcohol.  The marginal social utility of excessive drinking is negative, so a small reduction has a big benefit.  If everyone only drank one drink of alcohol per day, alcohol wouldn’t be a significant problem and it might even increase average life expectancy because of the reduction in heart-attack risk for people over 30.  Unfortunately, alcohol is addictive and it is hard for people to have one drink everyday without wanting more and more over time.  Excessive drinking causes most of the deaths and it only takes a little decrease in consumption to save most of the lives.

An increase in alcohol price can help people avoid excessive drinking because it has a bigger impact on the problem drinkers than on more responsible drinkers.  Someone who only drinks one beer per day is going to be impacted 1/10th as much as an alcoholic who drinks ten beers per day.  Because the budgetary impact on problem drinkers is ten times bigger than on responsible drinkers, the problem drinkers cut back the most in response to a price rise and and that is where a bit less drinking is most likely to save lives (primarily their own).

A big objection to alcohol taxes is fairness. They are very regressive because low-income people (and people at the median) pay a much higher percentage of their income as alcohol taxes than rich people. But alcohol taxes arguably benefit non-rich people more because they give us a greater incentive to avoid excessive drinking.   Rich drunks don’t care as much about a 50-cent increase in the price of a six pack and will keep drinking and dying despite the price increase.