Crime and Covid spiked after protests

The news says we just had a month of mostly peaceful protests, but my twitter feed is constantly filling up with violent videos. How bad was June? Let’s look at some statistics.

The riots started in Minneapolis. Here’s the weekly rate of burglary, in Minneapolis, for the last 5 years:

There was also a historically large spike in assault:

Shootings in Minneapolis tripled in June, from 24 last year to 76 this year.

Most of the victims survived. Murder overall didn’t go up much:

I saw some buildings burn down. Was that notably bad?

Yes and no. There was definitely an arson spike, with the protests. Prior weeks have been similarly bad, but usually the fire department can respond. We probably had lot more buildings burning to the ground.

Let’s move on to New York City. The first trend I see is that many crimes went down, when covid hit:

The number of rapes went down, during covid lockdowns:

Rape seems to be down about 30%, nationwide, in most cities I’ve looked at. Presumably, everyone’s staying inside and social distancing, so less sexual assaults happen. Also, rape went down last year for Christmas. I’m not sure if that’s just random noise. The same patterns show up in “other sex crimes”:

Assault also went down during the pandemic, but recently spiked:

Grand theft auto increased during the protests, to levels higher than last year:

Burglaries saw a huge surge during the riots:

Shooting victims doubled since last year. Let’s show this on a longer time scale, since it’s such a big spike. Shootings in NYC, in June 2020, were higher than any June in the last 20 years:

Murders, however, have not gone up remarkably. Most of the shooting victims survived.

Most big cities don’t offer a good interface for graphing crime, so I’ll just give some one month summaries.

In Chicago, murders went up 83% compared to last year, shootings are up 96%, burglary 77%.

During the last month in Philadelphia, burglary went up 1000%, shootings and homicides increased.

The videos I saw on twitter kind of represented the Philadelphia Burglary issue. Thieves were blowing up ATMs to get the cash.

In Atlanta, murders and shootings also doubled compared to last June:

I picked those 5 cities because I kept seeing videos of violence in each place.

This crime wave didn’t affect every city in the US. Things seemed roughly normal in LA, with little excess violence and only a 20% increase in Burglary:

The protests themselves have only directly killed 26 people, but the associated crime spike is bigger.

We have 30 more murders than usual in Chicago, 10 in Philadelphia, 6 in Atlanta, maybe 5 in New York. Nationwide, I would take a rough guess that 100–200 more people than usual were murdered in June, and more than 1000 were shot. It’s hard to extrapolate the results to smaller cities. Maybe the FBI will aggregate all the data, sometime next year.

In short, June was bad. Protests killed some people, others died amidst the increased violence.

Expand the time window a bit, and you’ll notice that we live in a historically peaceful time. Here are murders in New York City in an average June, going back to 1993:

Average shootings in June:

Burglaries for a typical June:

And Grand Theft Auto rates:

In 1993, an average month in New York City was deadlier than a riot in 2020.

In a sense, there are bigger questions to think about:
“why has crime gone down so much since the 90's?”
“if we reduce policing or change sentencing, will crime go back up?”

Rudy Giuliani, famously tough on crime, was elected in 1994. Even Democrats were worried about crime in the early 90’s. Here’s Joe Biden speaking in 1993:

Another big question is whether we will end up with a president Biden that thinks the way he did in 1993, or one that wants to defund the police.

The last big question is what the protests did to covid. I offer two headlines, posted one hour apart:

So, as we can see, it’s simple. Republicans spread covid but liberals do not.

And that’s just science!

Okay, the second headline is easier to analyze. Here’s Tulsa’s case rate:

Trump’s rally happened on June 20th. Tulsa’s caseload started increasing two weeks prior. There is a spike of cases on June 24th (two days after the headline was written). I’m not sure if that’s related — usually the virus takes more like 1–2 weeks to incubate.

I‘m sure that Trump’s rally did spread covid, with 7,000 people gathered and cheering, mostly without masks. But we probably haven’t seen the effect in the data, yet. The headline has no relationship to reality.

Let’s take a look at some other cities. Here are 6 from the west coast:

Cases start spiking around June 14th, across most of California and in Seattle. Portland starts rising a bit earlier. LA has been on an upwards trend for a while, but spikes around June 14th. And that is about 1–2 weeks after the big protest events.

The case spikes don’t seem well timed by the reopening of bars. Seattle didn’t reopen bars and restaurants until June 10th. Oregon on June 19th. California opened restaurants, for outdoor dining only, on June 19th and never reopened bars at all.

The spikes could be related to memorial day, but the time lag seems too long (3 weeks afterwards).

While held outdoors, often with masks, protest events were very large. Seattle had one March estimated at 60,000 people, Washington DC supposedly exceeded 100,000. People met in smaller groups before and after. Some of the protests were sustained. Seattle had an outdoor autonomous protest zone + homeless camp where people were in close contact with each other for weeks.

In Texas:

Big spikes in Houston, San Antonio, and El Paso, started around June 14th. Austin starts trending up on June 8th. Dallas was maybe already trending up in early June. Nothing in Amarillo. Amarillo did have a protest, but the crowd size was estimated at 500.

Idaho is having a spike. Basically all the cases look to be coming from Boise, which had a protest:

In Wisconsin, there’s a big spike in Madison (liberal college town that protested), which generally exceeds the rest of the state.

All that said, the national picture is complex, and there are surely many ways that covid is spreading. Arizona and Florida were both trending up fast before the protests (there looks like some acceleration around June 14th, especially in Phoenix):

And there are places that don’t fit the pattern.

Minneapolis and Chicago both went through earlier surges and now they’re both picking up again, starting on June 14th.

NYC has no visible spike in June. NYC already has a high level of herd immunity. 20–30% of NYC residents tested positive for covid antibodies in April, after a horrible first wave. Cuomo is still celebrating the achievement:

Most suspiciously, Washington DC seems to have no spike in cases, despite having had the biggest rally. They did have a first wave. And maybe most people at the rally were visitors from out of town, who then took it home?

Or maybe none of the protests were superspreader events. All case growth could just be effects of reopening the economy. The timing of the spikes could all be a coincidence and maybe big crowds are safe. In that case, the only thing we have left to fear is Trump rallies.