Crime Rates: A Good Moment for Political Instruction

An issue on which many are confused, and few productively cut through the noise

Law enforcement is one of the most tricky areas of government to discuss, partially because more people come to the conversation with knives unsheathed than usual (oh the irony).

This isn’t surprising. Public safety, public order, and justice in the courts and carceral system are important. They are at the foundation of what justifies government itself, and small mistakes in these areas can easily ruin lives.

This is all the more reason to get your facts straight, and to bend over backwards to ensure that you properly understand the policy landscape. You must be brave enough to disagree with your tribe when it is necessary, you must not crucify those who disagree with you, and you must build bridges to others when they are right. The staggering importance of this policy area demands nothing less, and much more still.

I’ll give you an example of how this works in practice, where we have:

• two “sides” that are unhelpfully entrenched in mutually exclusive positions, and

• a corrosive rhetorical environment that punishes the curious, well-intentioned, and the not-yet-sure.

The issue: crime rates, and what to think of them. Let’s examine a theoretical New York.

Crime rates

The sides:

• Team Data: Those who say that crime rates are down, and that anyone who complains about crime rising in cities is fear-mongering and ignorant of data.

• Team Lived-Experience: Those who say that they see more crime happening than usual, that they feel less safe, and that crime statistics are wrong.

If you are on either of these teams, pretend like you aren’t for the purposes of this post. I won’t tell anyone.

Whenever I see two sides line up across a rhetorical no-man’s land like this, my default suspicion is that both of them possess part of the truth, and that together they possess less than the full truth. Life and public policy rarely divide into an all-right and all-wrong side. This doesn’t mean that each side is equally correct or incorrect, just that each brings at least one piece of the reality puzzle.

In this case, both sides should ask themselves: is it possible that crime rates are falling, but the impact of crime (number of people affected) is rising?

This is definitely possible.

Scale

Urbanists often point to the positive benefits of urban agglomeration, one of which is positive externalities. If your neighbors plant flowers, you get to enjoy them. If they open a store next to your store, you likely reap a foot-traffic premium from the extra people they bring into the area, etc. Urban areas benefit from vast positive externality spill-over, and from economies of scale. This is (for urbanists) famously described in Geoffrey West’s book Scale.¹

But bad things scale just like good things, and that is the key to understanding how crime rates can fall, but crime impact can rise. Let’s look at three different examples of the effects of crime scaling.

A bomb

Imagine that someone sets off a bomb in an abandoned Queens warehouse at 3am. It explodes, but no one is hurt, and the people who were woken up probably assume something less dangerous happened (fireworks, probably).

Now take that bomb, and detonate it in a crowded public square at 1pm on a weekday in Manhattan. The effects will be devastatingly different.

But in both cases, the number of bombs that went off is just one.

The time and place of a crime affect its magnitude, and not all elements of magnitude are always captured in crime data. In the case of the bomb, all deaths would appear somewhere in the statistics—but the psychological spill-over into the population would not.

Let’s look at a less dramatic example, although it’s worth noting that NYC bombings used to be more common.

A burglary

Imagine that someone breaks into a house in rural New York and steals several valuable items while shouting at the home’s family to stay back. The burglar leaves, driving off into the still night, and no one is hurt.

Now take that same crime, and put it inside a New York City apartment building with paper-thin walls. The family in the burgled apartment feels the terror of the crime, but their neighbors probably do as well. They are unsure if they’re next, either that evening or one that follows.

In both cases, the number of burglaries is one. But the number of people who are justifiably afraid, and the number of people who have experienced crime, has gone up quite a lot.

A mugging

Imagine that someone is walking home from a club late at night along a sparsely populated street in Brooklyn. Suddenly, they feel a gun at their back and a stiff hand on their shoulder. A voice behind them demands their wallet and their phone, which they hastily hand over. When the gun is withdrawn, they turn to see a figure fleeing around the corner. No one else is around.

Now imagine that mugging happening in broad daylight along a busy Fifth Avenue. There is no question that this is different: possibly hundreds of people will witness it. They’ll not only be shocked by the nature of the crime, they will be shocked by the criminal’s boldness.

In both cases, the number of muggings is one. But how many people in each scenario can truthfully raise their hand and say they’ve witnessed a mugging in New York City? In the second scenario, hundreds could.

The lessons of scale

It is possible for the crime rate to go down, but for many more people to witness, and be affected, by crime. In addition to crime rates, you should also try to approximate criminal magnitude.

If you had two identical crime rates in two theoretical New Yorks, but the crimes in one were all during broad daylight and in crowded public places, that second New York would have many more people truthfully saying “I have witnessed crime. I have been psychologically affected by witnessing it. I know more people who have witnessed and been impacted by crime.”

So what does this mean for Team Data and Team Lived-Experience?

Well, they both have a piece of the puzzle. Team Data is correct to note both: (1) that crime rates are down, and this is good, and (2) that New York City is safe. But they are incorrect to fully dismiss Team Lived-Experience.

Team Lived-Experience, even if they don’t say it this way, is correctly noting criminal magnitude. Rising criminal magnitude is a danger to public order, public norms,² psychological health, and feelings of safety.

Brave individuals on both teams need to step forward and synthesize all of this information—they need to partially disagree with their team, and partially agree with their opposing team. In this way, they will get an accurate gauge of criminal reality, and be far more able to pursue correct policy remedies. They will also, ideally, make their respective teams realize that they were overly dismissive of the other side that was partially correct.

Questions of data

In the hypothetical scenarios above, I assumed that crime rate data was all correct. But crime data is never perfect, even in ideal conditions. And we are not in ideal conditions. Both Team Data and Team Lived-Experience should be extremely rigorous when they see crime rates go up or down, but especially Team Data.

What causes crime rates to go down

• People commit less crimes, and this is accurately recorded.

• Data collection errors, omissions, or methodology changes. Are you accurately counting the number of crimes? How many crimes aren’t being reported? Were you overcounting certain crimes, and now you’re correctly counting?

What causes crime rates to go up

• People commit more crimes, and this is accurately recorded.

• Data collection errors, omissions, or methodology changes. Are you accurately counting the number of crimes? How many crimes are incorrectly counted? Were you undercounting certain crimes, and now you’re correctly counting them?

The lessons of data collection

A lot can go sideways in crime rate data collection. A good default stance is to assume that your team is missing something important, or that there is another way to interpret the data you see. Do not be surprised if someone on the other team, operating in good faith, makes a great point.

OK, go back to your teams

I’m not saying you shouldn’t join a team. A lot of people do, and for good reasons. To begin with, it’s easier to advocate for an issue as part of a team. But if you want to understand the truth of crime in New York City, you must be willing to take a wider view than your team does.

You must ask yourself: are you interested in shouting at someone online or in person, or are you interested in understanding reality, and deploying public policy tailored to it?

If you are nervous to speak about law enforcement (or any issue) for fear of one of the teams in the field, I am on your personal team. If you find that you disagree with your team, and you are afraid of what they will do to you if you voice your disagreement, I am on your team.

Maximum New York is a civics school, and we believe that before you throw yourself into questions of what ought to be, you must first grasp what is. The pursuit of that ground truth is vital to a better—a Maximum—New York.

“Views of the same city from different perspectives.” From Midjourney.

1. See Scale, p. 274 and 276. Alternatively, you can view the study that these pages reference here, with particular reference to Table 1 and Figure 1. Although I think these data still hold, and are supported by many more recent studies, I would like to see these numbers run again with 2019 data (to avoid Covid complications).

2. People see, people do. If certain kinds of bad behavior are allowed in public, they will increase with no pushback. NYC is seeing this now with people not paying their MTA fares. Not paying for the MTA is socially allowed in a way that taking other things without paying is not. Some people think subways should be free (or: not paid for at the point of service, but via non-user income streams like taxes), so they don’t see it as stealing (taking the rightful property of another); but they’re probably unaware that they’re shifting norms in an undesirable direction with regard to stealing generally.