Explaining the crime wave of the 1980's.
The murder rate in the US suddenly got worse in 2020 and 2021, compared to 2019. That’s big news — things were getting better, from 1993 to 2014. Then there was a small uptick in 2015 and a bigger one in 2020:
I’ve written a few posts about the increase in 2020. A more interesting question is why things were so much worse in the 1980's.
Public perception tends to be confused. A majority of voters think that crime gets worse, every year. But the reality is that there are trends, it comes and goes for various reasons.
Let’s zoom out for the 100 year view. Wikipedia offers this graph:
America has seen 2 big crime waves in the past 100 years.
There’s some debate about crime rates in 1900. Most likely they weren’t that low. The FBI didn’t start counting until 1913, so the data before that is vague. Another source offers this picture of the beginning of the century:
Either way, there were two big crime waves in the 20th century.
Many people think that the 20’s had more crime because of alcohol prohibition, which started in 1920 and ended in 1933. People still drank when alcohol was illegal, but gangsters had to use violence to defend their business, and establish who could distribute it (think of Al Capone, shooting up his rivals with a tommy gun).
The crime wave from the 1960’s to the 1990's is more interesting because no one agrees what caused it — people on different sides of the political spectrum push wildly different theories for it.
Let’s go through some of the theories and find which ones hold up.
5 things that don’t explain the crime wave:
1. The Economy
Unemployment rates got worse from 1965 to 1980, and the murder rate also went up:
But unemployment got much better from 1982 to 1990, while the murder rate kept going up.
Unemployment also got just as high in 2009 as in 1982. There was no increase in murder during the 2008–2009 “great recession”.
Most studies agree that high unemployment is related to an increase in property crime, but there’s no consistent relationship between the economy and violent crime.
2. More Guns
People on the left blame gun violence on there being too many guns in America.
The crime wave of the 80’s was an epidemic of people shooting each other:
More specifically, it was an epidemic of people shooting each other with handguns:
Did the US have more guns during that time period?
The beginning of the crime wave lines up with a small rise in gun ownership. But gun ownership kept going up in the 90’s and after 2000, while crime declined:
This might be a misleading way to present the data. This is “firearms per person”. By about 2008, America had more guns than people. A person can own more than one gun, some people hoard ten or more.
Another way to look at it would be to ask how many households own a gun. That number has stayed about the same through the whole time period:
There is also no relationship between gun ownership rates and murder, across states in America. The most heavily armed states, like Wyoming and Montana, have some of the lowest murder rates.
There are a lot of guns in America. Just having the total number go up doesn’t seem to affect crime. We need to look for other explanations.
The crime wave was restricted to certain demographics:
There was a crime spike in 1980, which was most pronounced in young men, but it shows up in groups as old as 50. Then there was a second crime spike in the late 80’s and early 90’s, where men aged 14–24 did most of the shootings. There was little increase for older men, and none for men over 35.
There were also big racial disparities in who was being shot, with murder rates much higher for black men than white men:
Rates for Hispanics were somewhere in the middle:
So, the late 80’s and the early 90’s saw a sudden increase in black men, aged 14–24 shooting each other.
It’s possible that young black men were more likely to carry handguns during the 80's. But gun ownership rates for the whole country weren’t the problem. We need to look for explanations for why these men were shooting each other.
3. Gun Control
Conservatives sometimes like to argue that gun control causes crime. I’ve seen this image posted on reddit:
If I wanted to make the pro-gun control case, I could say the opposite:
The 1993 Brady Bill increased background checks for gun sales. Near the peak of crime in 1994, Bill Clinton passed an assault weapons ban. If I wanted to make the case for gun control, I’d credit those law, instead of “concealed carry permits”.
To be clear, neither of these graphs is accurate. Clinton’s assault weapons ban didn’t have much effect because most murder is done with handguns. The background checks from the Brady Bill probably didn’t have much effect on crime, either — prior to the bill, only 20% of criminals got their guns through legal means. Most of the violence of the 80’s was done by young black men who were concealing handguns, so the concealed carry laws that conservatives favor shouldn’t have much effect.
Is it possible that even more extreme gun control, like a complete ban on handguns, would stop crime?
Washington DC tried this, banning handguns in 1976:
The law was later struck down by the supreme court. In between those two dates, DC had a worse handgun shooting rate than almost anywhere else in the US.
The conservative meme also assumes that life was civilized pre-1900 and that the “wild west” was peaceful. More likely, this is just because we don’t have good statistics on murder from back then.
Other scholars have tried to estimate murder rates in the distant past, and concluded that things were less peaceful:
In the 1800’s, gang violence happened among many immigrant groups, like the Irish. Gangs of New York was set in 1862.
4. The Death Penalty
On the right, we occasionally see the argument that “murder is deterred by the death penalty”. We stopped executing prisoners in the 70’s and 80's.
I find this implausible. Murder is sometimes an impulsive act. Sometimes it’s a calculated one where the perpetrator hopes to get away. I don’t think anyone decides to kill based on the odds that they will be executed 20 years later, as opposed to being given life in prison. Young men, aged 14–24, are the demographic that would be least likely to reason about the consequences. Young gang members carry guns for protection, they’re far more worried about getting shot today than executed 20 years from now.
Also, there were plenty of executions in the 1930’s, along with high crime.
The one thing this idea has going for it is efficiency. Suppose executions had explained the entire crime wave of the 60’s through 90’s. In that case 100 executions per year prevented 10,000 murders.
Now suppose it had only a small deterrent effect, and explained only 2% of the crime increase. Then each execution would still save 2 lives. As we’ll see later, this is more efficient than other solutions.
I’m not going to spend much time considering this theory because I don’t see it used very often. Most conservatives argue that crime declined in the 90’s because of different reasons like changes in policing or jailing. Both are plausible theories that I will look into.
5. Abortion and Crime
On the left, we have the theory that abortion prevents murder by creating less unwanted children. Freakonomics author Steven Levitt proposed this idea.
Starting with Roe v Wade in 1973, more pregnancies ended in abortion, peaking at about 30% of all pregnancies being terminated in 1980.
Levitt suggests that unwanted children are more likely to become criminals and abortion gets rid of unwanted criminals. With more abortions, less criminals are created. 20 years after Roe V Wade, and 13 years after the peak in abortions, we saw a huge crime decrease starting in 1993.
When you dig into it, there are a lot of problems with this theory.
Levitt supports his argument with a scatter plot of states. Where abortions increased the most, murder decreased the most (from 1985 to 1997):
It’s not a strong correlation. Take out New York and the plot looks almost random.
The theory also doesn’t address why the crime rate spiked, to begin with, starting around 1965. Those children had been born in 1950, abortion was illegal at the time. Abortion was also illegal in 1940, and that cohort was less violent.
A lot of people have ethical issues with the idea. Abortion is about 3 times as common among black women as white women. So the idea suggests that crime goes down because more black children are aborted.
Abortion is an incredibly inefficient way to prevent murder. Levitt writes,
One additional abortion is associated with a reduction of 0.23 property crimes, 0.04 violent crimes, and 0.004 murders annually when a cohort is at its peak crime age.
In other words, you need to abort 250 fetuses to prevent one murder per year.
If the death penalty theory were true, then one execution would save multiple lives. If the abortion/crime theory was true, we’d need to abort 250 to prevent 1 murder.
Let’s put aside all the ethical issues. We still want to know: is this theory true or not?
The biggest problem with Levitt’s argument is the age of the shooters:
In the late 80’s and early 90’s, murder spiked among those aged 14–24.
At the peak of the violence, in 1993, those men had been born between 1969 and 1979. Roe vs Wade happened in 1973.
If this was all about unwanted children, then those teenage murder rates should have been high up until 1987 (14 years after Roe vs Wade) then started dropping and kept dropping until at least 2004 (24 years after the peak of abortions in 1980).
In fact, the opposite happened. Teenage murder rates were lower, then started to spike in 1984/1985, and kept going up until 1993.
If this was about unwanted children, then the rates should also start dropping in the 14–17 year old group, next in the 18–24 year old group, then in those 25+. Instead, there’s something that causes rates to spike for young men, all at the same time.
The crime wave of the late 80’s and early 90’s was especially notable in that younger teens, aged 14–17, started shooting each other:
And those teens were old enough that they were born after Roe vs Wade. It’s hard to salvage Levitt’s argument against this data. There has to be something bigger going on to explain the wave of young black men shooting each other, starting in 1984.
5 Things that do explain the crime wave
1. Crack Cocaine
From Wikipedia:
Crack first saw widespread use as a recreational drug in primarily impoverished neighborhoods in New York City, Philadelphia, Baltimore, Washington, D.C., Los Angeles, San Francisco, and Miami in late 1984 and 1985; this rapid increase in use and availability was named the “crack epidemic”, which began to wane in the 1990s.
This is, far and away, the best explanation for violence in the mid-80’s to late 90’s.
Steven Levitt wrote a later paper analyzing the crack cocaine epidemic. This never became as famous as his work on abortion and crime.
Crack use lines up with the start of the murder spike. The data shows why murder spiked much more for black and hispanic men:
Levitt still believes that abortion was a factor in crime declines, but admits that the crack epidemic was also a factor.
I would agree that abortion could have some factor in the decline, but it’s not the least bit obvious from the data. I would have to crunch the numbers carefully and control for other factors to find it. For the period from 1984 to 1999, crack cocaine causes a clear spike in murder, while any effect of abortion is subtle.
The crack cocaine theory still leaves a few things unresolved. It doesn’t explain why crime declined in the 90's.
Levitt’s data shows that cocaine deaths continued past 1994, even as crime declined.
In the mid-80’s, crack usage was also associated with lots of low birth weight black babies, but this went down by 1999:
Levitt writes:
The link between crack and adverse social outcomes weakens over the course of the sample. Even though crack use does not disappear, the adverse social consequences largely do. Thus, by the year 2000, we observe little impact of crack, which accounts for much of the recovery in homicide rates and child outcomes for Blacks over the period. We hypothesize that the decoupling of crack and violence may be associated with the establishment of property rights and the declining profitability of crack distribution. The fading of adverse child outcomes may be attributable to the concentration of crack usage among a small, aging group of hardcore addicts.
The other part he doesn’t mention is that we put a lot of people in prison. As we’ll see, this had an effect on crime.
The crack cocaine epidemic also doesn’t explain the earlier rise in crime, from 1964 to 1980, only what happened from 1984 to 2000.
2. The Baby Boom
We’ve addressed the peak of crime starting in 1985.
We still haven’t figured out why it started rising in the 1960’s to begin with.
Part of this was simply demographics:
Most violent crimes are committed by young men. When there are more young men in the population, crime is higher per capita.
Birth rates were high in 1910.
Birth rates dropped in the 20’s and hit a low during the great depression of the mid-30’s. Crime was low 20 years later, in the mid-50's.
The baby boom started in 1945 when world war 2 ended. It peaked in 1960.
Crime started going up 20 years later in 1965 and hit a peak in 1980. Crime should have peaked in 1980 and declined, but the crack epidemic followed afterwards and made things worse.
1960 was the year birth control pills were approved. So, this is a bit like the abortion and crime theory — making less children makes society older, on average. Crime rates go down, per capita.
Canada also had a baby boom at the same time:
And also had a crime wave starting in 1965:
Demographics don’t explain the entire US increase in murder from 1965 to 1980. Demographics predict murder should go up, but not by 100%. Even on an age adjusted basis, homicide rates also went up. They doubled for both black and white men:
Why did people become more murderous?
I’m not sure. Maybe there was already violence related to drug turf wars starting in the 60’s and 70's, well before crack cocaine in the 80's.
Maybe it was migration of blacks from the south to northern cities (and the ghettos that formed there, especially after manufacturing jobs dried up).
Maybe it was an increase in single motherhood, that started in the 1960's:
It’s hard for me to pin down the reasons.
Why did crime fall again, after 1994?
3. Reduced lethality of murder
In some sense, it didn’t fall as much as we think it did.
A paper published in 2002 points out that shootings have become less lethal, because emergency medicine has improved.
If we graph murders and aggravated assaults, relative to 1960, we find that assaults have gone up at a faster rate than murders have:
We can compute a lethality rate of each assault, defined as:
murders divided by (murders + aggravated assaults)
The rate goes down every year from 1960, presumably because trauma care got better. There’s a big drop after 1974, which maybe came from trauma care lessons learned in the Vietnam war.
Lethality flatlines starting around 1990, there’s a small uptick in 2020 (maybe hospital care got worse because of covid?)
One alternate explanation is that the definition of assault has changed, and we now include less lethal assaults, maybe things like domestic violence. That would mean the changes in lethality aren’t real.
One reason to think that lethality has gone down is that the lethality of car crashes fell at almost the same rate:
Including the data on lethality, we might conclude that life in 2010 wasn’t actually as peaceful as in 1960. Things are still better than the height of the crack epidemic in 1990, but otherwise assaults are still about as bad as 1980.
It’s still not quite that simple. Robberies are back down to 1960’s levels:
Rape numbers are down from the peak but still near 1980’s levels:
4. Mass incarceration:
We also started locking up huge numbers of people, starting around 1980:
Most studies think this accounts for about 10–20 percent of the drop in crime.
These arrests weren’t all for violent crime, most of them were actually for drugs:
But, criminal behavior has some correlation. If you lock up people for robbery and assault, you’ll prevent a few murders. If you lock up lots of drug users, you’ll prevent a few murders, as well. Just like abortion, the ratios are bad. You need to jail many people to save one life.
We also started giving longer sentences for each type of crime:
5. More Policing
The US hired about 14% more police officers in the 90’s. It’s estimated that this cut crime by about 5 or 6%.
The effect is subtle in most places, but New York City shows an extreme example. NYC cut its police force from 1970 to 1980, then started to increase it again:
Murders in NYC hit a peak in 1980, then started dropping. There was a second surge in the late 80’s with the crack epidemic, then they continued dropping.
NYC has continued to fare much better than it did in 1990, while some other big cities have ticked back up to those levels of violence:
It would be worth investigating what NYPD has done differently than Chicago or Philadelphia, if we want to solve 2020’s problems.
Putting it all Together
Crime started rising in the 60’s simply because of the size of the young generation. That should have caused a 50% increase in murder, but violent crime actually went up 200%.
Some other factors (drug wars, ghettos, single mothers?) caused a larger social breakdown.
Things should have peaked in 1980, but they got worse because of crack.
After 1994, things got better due to some combination of:
•reduced turf wars for crack
•increased policing
•mass incarceration
Abortion wasn’t the main factor in the decline, but it may have some smaller influence after you correct for these things. The economy, high gun ownership, and gun control laws didn’t have much impact.
Steven Levitt made an estimate of how much effect each factor had:
Even in Levitt’s model, abortion is not the biggest factor — mass incarceration is.
As I showed earlier, Levitt’s model doesn’t match the ages of the offending groups, the criminals started getting younger when his abortion theory would have predicted the opposite. I think that’s because Levitt is overestimating the effect of abortion and underestimating the effect of crack cocaine.
I think he may also be missing the advances made in trauma medicine. Property crime has continued getting better since 1994. Shootings today are still about as bad in 2020 as they were in 1980. Emergency room medicine has gotten better, so we’ve reduced the likelihood that each shooting victim will die.
Whatever social breakdown started in the 60’s hasn’t gone away. Whatever social factors create more criminals than 1960 still exist. But we’ve decided to keep large numbers of people locked up.
And then things suddenly got worse in 2020. Based on prior trends, we should wonder if 2020 saw any changes in drugs, incarceration, or policing.
