How Many Crimes Are There, and Why Does it Matter?

"Unreported Crimes" cannot be counted in any objective way that actually matters...

Some readers may know that before I became a criminal defense attorney, I was a librarian. I used to introduce students to basic academic research concepts by having them read my favorite academic journal article: How Many Birds are There, and Why Does it Matter? This article purported to measure the total number of birds on the planet by using a variety of different methods that all, strikingly, added up to a pretty similar number. It was a great way to show how something so seemingly uncountable could actually be reckoned pretty precisely.

(how many birds are in this picture? Also, is this one discrete object or lots of little ones?)

I fear I’m going to write the exact opposite thing today. I’m writing this article in response to a series by Scott Alexander trying to push back against arguments, often made by reactionaries, that crime is increasing. There is a well recognized decline in crime rates in many Western nations, which Scott Alexander argues is a real phenomenon and not some artifact of sampling. I kind of agree with him, with caveats below that will consume the whole article.

More broadly, I’m writing this because a lot of commentary on the criminal justice system and the immigration system in the United States centers around the vexing question of separating “real criminals” from “non-criminals.” This is very easy to do if you trust the results of the criminal justice system: we can speak of “convictions”, “prosecutions”, or “reports” of crime as some kind of objective measurement of “the crime rate” without thinking too hard about what “crime” is, because the whole point of the system is to sort that out.

It’s much harder to get an objective accounting of crime if we are skeptical of the system…as many are, especially the reactionaries Scott is addressing. These people don’t trust any of the three metrics above as useful measurements for the “objective” amount of crime in our society. In their view, these are not good measurements because a good deal of crime goes unpunished…even uncharged…even unreported in the first place.

And I don’t disagree! My point in writing this essay to demonstrate, perhaps ten different ways, that crime is not a discrete and objective phenomenon like “birds.” It’s essentially impossible to meaningfully discuss the NUMBER of crimes that go unreported, and any attempt to do so will yield millions upon millions upon millions of unreported crimes, more than anyone would care to prosecute…or even remark upon.

And I want to be clear at the outset that all of my examples are not legal sophistry: they are precisely what happens when one tries to apply fuzzy maxims of morality (“crimes should not go unpunished”) to the concrete work of law, which as I’ve really tried to emphasize repeatedly is a practical discipline, not a theoretical one. To illustrate this, I’m largely going to keep my examples confined to a few areas that I think are un-controversially crimes reactionaries care about: frauds, thefts, assaults, and drugs. I will cite to caselaw as often as I can.

I: BUT FIRST, A CIVICS LESSON

Just so we are clear on what a crime is. In the United States, crimes are, with very few exceptions, violations of the explicit text of a specific statute, passed by the legislature of a particular jurisdiction. The general proposition is that “no federal crime can exist except by force of statute.” Morissette v. United States, 342 U.S. 246 (1952). Every state has a similar rule. Thus when we are talking about “crime”, we are necessarily talking about a specific act which violates a specific statute of a specific jurisdiction. If an act meets all the “elements” of that offense, it IS a crime. If an act fails to meet all the elements, it is NOT. For example, one form of Second Degree Murder in Minnesota is committed by one who

1) “causes the death of a human being”
2) “while committing or attempting to commit a felony offense”
3) “other than criminal sexual conduct in the first or second degree with force or violence or a drive-by shooting”

Simple enough, but note that this definition of “second degree murder” is completely different than the one found in Arkansas and murder is, in general, a pretty straightforward crime. Later, we will be taking a walk through Minnesota’s theft statute and that thing is a beast.

I hate to start off at this level but it’s really really important. When we talk about whether a “crime” has occurred, we are necessarily engaging in a technical inquiry, asking whether the particular elements of a criminal statute in a jurisdiction were, or weren’t, met by a particular act. In the United States, that is what crime IS. There is no alternative definition.

I say this to preemptively deflect objections that my later examples are hypertechnical. If someone were to approach me and say “We should punish all criminals more harshly, especially people who smoke cigarettes” I think I would be well within my rights to respond “actually, cigarette smoking isn’t a crime.” What I mean by that is necessarily a very technical proposition, what I’m saying is that there is technically no statute…within a particular jurisdiction…the elements of which punish smoking cigarettes in the abstract.

With this civics lesson firmly in mind…

II: HOW MANY FRAUDS ARE THERE?

The federal wire fraud statute, one of the crimes that I deal with most often, punishes one who “having devised or intending to devise any scheme or artifice to defraud…transmits or causes to be transmitted by means of wire…ANY writings, signs, signals, pictures, or sounds for the purpose of executing such scheme or artifice…” Simply put, this statute is violated whenever a defendant transmits a wire communication in furtherance of a scheme or artifice to defraud. Rather importantly, the act of sending the communication is what is punished.

At the federal level, this is a very commonly-employed statute. You’ll see it on the Elizabeth Holmes indictment, on the NBA points shaving scheme indictment, and on the Sam Bankman Fried indictment. But hang on…if you look at those indictments you’ll see only a few counts of wire fraud are charged. And yet under the statute every single wire communication in furtherance of their scheme is a wholly separate wire fraud. Why were only a small number charged? The answer is that charging each and every wire fraud Sam Bankman Fried committed would occupy every judge and prosecutor in the nation until doomsday...and all for no consequential effect, since under the guidelines the total loss amount is all that matters for sentencing purposes.

Again, there’s no sophistry here: it’s a hard legal fact that each and every person defrauded by Sam Bankman Fried’s FTX suffered a discrete instance of fraud that could have been charged, but wasn’t. I have no doubt for each single instance it could have been added to the indictment, and a jury would have convicted. Each act is in every sense of the word a “real crime” that “could have been charged” with a “real victim” who suffered “real harm”…and FTX had more than a million customers. Most of these would not even enter the statistical universe of “reported crimes”, since each and every FTX customer didn’t ring up the FBI.

This single example dwarfs the total number of reported crimes (violent and nonviolent) reported in the city of los angeles in 2024…indeed it vastly exceeds even the total number of police service calls. And that’s just one guy. Think of how many scam emails you receive. Each of those is a wire fraud.

So: that’s a few million unreported, uncharged crimes we need to add to our total, right out the gate. I hope this one example serves to demonstrate that we are, at all times, adrift in a sea of crime that absolutely swallows the total number of crimes “anyone cares about” by several gazillion orders of magnitude…but in case I haven’t made my point I’ll go one step farther.

III: HOW MANY THEFTS ARE THERE?

I, like basically everyone reading this blog, was once the victim of a crime that was never charged…in fact I never reported it. Once, many years ago, someone stole my lawn furniture. I was disappointed, but reporting it would have been foolish: there was simply no way to prosecute such an offense, given that it happened one afternoon when I wasn’t paying attention… but make no mistake, this was a theft under Minnesota law, which punishes among other things one who “intentionally and without claim of right takes, uses, transfers, conceals or retains possession of movable property of another without the other's consent and with intent to deprive the owner permanently of possession of the property.”

So, perhaps the number of crimes that “truly occur” in the state of Minnesota should be increased by 1, to reflect the fact that this crime, while unreported, represented a single instance of theft against a truly sympathetic victim.

…but hang on a minute. This statute is subject to exactly the same subdivision as the fraud example above. After all, the thief made off with multiple discrete pieces of furniture. In each case, the thief “intentionally” and “without claim of right” took some movable property from me.

Lest you think this is ridiculous, rest assured there are numerous Minnesota cases addressing this exact issue. I’ve probably litigated it myself 3 or 4 times in my (relatively short) career as a Minnesota lawyer. In State v. Henline for example, a Minnesota lawyer took numerous unauthorized draws from a trust fund, the draws were for large amounts, and made over a period of years. Was this one theft or several? In Henline, the Minnesota Court of Appeals held that the entire course of conduct Henline committed could be charged as a single discrete theft…but a recent Minnesota Supreme Court case has thrown this case and others like it into doubt. That’s right, this issue has been litigated in Minnesota quite a few times.

And lest you think this is some quirk of white collar law, or Minnesota law, consider the U.S. Supreme Court case United States v. Wooden, high up on the list of my favorite SCOTUS cases ever. In Wooden, an enterprising criminal broke into ten different storage units on the same evening, smashing through the wall of one to break into the next. This was charged out as ten separate instances of theft (all of which Wooden was convicted of) and the Supreme Court had to unsnarl the nasty question of whether these ten crimes (it was uncontroversial that these were ten separate crimes) were committed on “occasions different from one another.” The Supreme Court said they were not, but notice, Wooden’s ten thefts, from the same place, in a single night, were charged as ten DIFFERENT crimes, simply because a prosecutor decided that made more sense.

This issue ends up being so messy that many jurisdictions, including Minnesota, expressly allow a prosecutor to aggregate lots of little thefts into one big one. Here: “the value of the money or property or services received by the defendant in violation of any one or more of the above provisions within any six-month period may be aggregated and the defendant charged accordingly in applying the provisions of this subdivision.” So, prosecutors are allowed to cut this knot by simply aggregating a bunch of little thefts, but that doesn’t change the fact that they could split up thefts almost infinitely, and charge a bunch of little ones, and if we are engaging in an analysis of how many thefts ACTUALLY occur, irrespective of which ones a prosecutor feels like charging, it feels like we have to actually examine each theft on its own terms, as a discrete act that could be charged in isolation.

Importantly, as well, this aggregation provision was added precisely to make it easier to address shopliftings and related small thefts. This problem is ubiquitous. Other jurisdictions have similar aggregation provisions, and for similar reasons.

IV: HOW MANY DRUG POSSESSIONS, DUIS, AND ASSAULTS ARE THERE?

Other examples are legion, but I think people get my general point, so I’ll make a few more briefly.

Drug Possession: In Minnesota, one commits fifth degree drug possession if they “possesses one or more mixtures containing a controlled substance classified in Schedule I.” I hope it goes without saying that this is a very commonly prosecuted crime. The rate at which this crime is COMMITTED is far higher…but it’s a little weird to talk about a “rate” for this crime. Notice that this crime is committed whenever someone “possesses” a “mixture or substance.” Let’s say they possess a small bag of cocaine and a small bag of MDMA, and they do so for six days, holding it for a friend who will later pick it up. Eventually, the pickup happens. Is this one crime? two crimes? (one for each bag or person?) Six crimes (one for each day?) twelve crimes? (one for each bag or person for each day?) twenty four crimes (one for each bad and for each person for each day?). We can go on. And yep, this isn’t sophistry. There’s a case on this question.

DUI: Driving while under the influence is probably the most frequently charged crime in Minnesota. It’s also a very harmful crime, it can do real damage. It’s a social ill. How does one meaningfully talk about the “number of unreported DUIs” though. Do we guess at the rate at which people drive drunk and are not caught by police? note that as in most jurisdictions the crime of Driving while impaired in Minnesota is rather specific: It doesn’t punish consuming ANY amount of alcohol and driving. It punishes one who has an “alcohol concentration at the time, or as measured within two hours of the time, of driving, operating, or being in physical control of the motor vehicle [of] 0.08 or more;” or one who “is under the influence of alcohol” when they operate a motor vehicle, meaning they are some degree intoxicated.

So: one criteria is a specific and highly scientific one made with reference to a specific testing device kept in a lab somewhere. The other is extraordinarily vague bordering on a general opinion. How in the world do we even talk about the rate at which this crime occurs for instances which are unreported? Am I under the influence of alcohol if I drive after having a drink 2 hours previously? 5 hours previously? This crime, like a lot of crimes, is difficult to analyze without reference to a jury…but an unreported crime definitionally will never see a jury.

Assault: “Unreported assaults” seems like exactly the kind of thing a reactionary would centrally be concerned about…but defining them is pretty complex: In Minnesota, one commits an assault if they “Intentionally inflict[] or attempts to inflict bodily harm upon another.” Another statute defines bodily harm as “physical pain or injury, illness, or any impairment of physical condition.” So…this raises a bunch of nasty questions about what actions would be assaults: thrown snowballs? unfortunate high-stickings? slaps to the face? flying tackles? Is there a consent defense? Some kind of exception for friends or sport? Litigating this seems very hard without knowing the precise facts of each particular assault.

Lesser Included Offenses and companion crimes: Many crimes are “lesser included offenses” of other crimes, meaning that one necessarily commits a lesser crime when they commit a greater one. As an example, murders are also generally assaults, robberies are also generally thefts, instances of first degree criminal sexual conduct are also instances of fifth degree criminal sexual conduct. Many other crimes also simply follow others, as companion crimes. Someone can, in the same instant, commit the crimes of murder, assault, discharging a firearm in city limits, possessing a gun as a felon, possessing ammunition as a felon, and kidnapping. Are these all separate crimes? It’s actually hard for me to imagine a criminal act which is not, in fact, multiple possible crimes that could all get charged at once.

It seems like however we slice it, we are dealing with a truly staggering amount of uncharged crime, so much that even slightly tweaking our definition of “crime” to more centrally involve the “kinds of crime people actually care about” would incorporate so many problematic inclusions as to totally drown out the total number of “actual” crimes reported or charged in a given year.

V: PRE-EMPTIVELY ADDRESSING SOLUTIONS TO THIS PROBLEM THAT WILL NOT WORK

We count “Crimes with Victims”, we don’t count “Crimes that have no real victim”: most of my examples above involve crimes with victims. Sam Bankman Fried probably victimized millions of people. So did Archer Daniels Midland and Enron. Did Phillip Morris? Did Dow Chemicals? These latter cases are examples of crimes that COULD have been prosecuted, but weren’t. They probably had hundreds of thousands or millions of victims. Also, importantly, many crimes we would probably like to include in the reckoning of “uncharged” crimes have no victim at all: Driving under the influence has no victim, neither does possessing or trading in drugs.

We count “Malum in Se” crimes, we don’t count “Malum Prohibitum” Crimes: Again, I’ve very carefully tried to pick examples of crimes that are very “real” crimes: defrauding an elderly person, which Sam Bankman Fried did hundreds of thousands of times, is a fraud, a species of theft. In Minnesota we punish fraud and theft under the same statute. Thus theft of the kind where you lift a candybar from a shop is punished using precisely the same statute, Minn. Stat. 609.52, which punishes theft of the kind where you cook your books and overstate certain sums. It’s the same crime. Defining “malum prohibitum” or “purely technical” crimes is pretty difficult too. Is DUI a malum prohibitum crime or a malum in se crime? What about possessing ecstasy? Possessing marijuana? Possessing a handgun as a felon? Again these are very commonly-charged crimes.

We count “violent” or “serious” crimes and not “minor” ones: I’d be more sympathetic if the discourse around uncharged crimes didn’t perpetually circle the issue of shoplifting, a relatively minor crime. Many, engaging in this discourse, would be outraged by the theft of a can of spray paint or a stick of gum. Most of my above examples involve far more “serious” crimes.

We just count unreported thefts: This is why so many of my examples involved theft. I’m a white collar lawyer, and the two crimes I bet I see most often in cases are federal wire fraud and minnesota’s theft statute. Both are capable of endless division, subdivision, and aggregation. Both could theoretically be used to charge everyone, or selectively saved for only “really bad offenders.” And that’s okay! That’s how law is supposed to work.

We count “crimes that would get charged if they were reported”: This requires a pretty acute understanding of what sorts of crimes “would actually get charged”, something that changes with each new administration that takes control of each jurisdiction. Sometimes, an executive might decide to crack down on shoplifting. Other times, they may decide it’s no big deal. Meaningfully talking about whether a crime is “the sort of thing that does get charged” requires understanding how a particular city, county, or state makes its charging decisions. It’s difficult to discuss in the abstract. Also, isn’t the whole point of this to talk about ACTUAL crime rates, irrespective of charging decisions or the whims of prosecutors and police?

We count “crimes a jury would actually convict on”: This is even more abstract. Now, for each crime, we have to consider a hypothetical jury and what they would do with hypothetical evidence presented at a hypothetical trial. This requires a pretty firm understanding of local sensibilities and trial law. Keep in mind that Juries have a well recognized tendency to acquit people even in the face of clear evidence if the crime seems very minor, or the defendant seems sympathetic. Conducting a mental “mini-trial” for every potential crime seems insane, especially since I expect Sam Bankman Fried, for example, WOULD be convicted for each and every one of those million victims if he stood trial for them.

VI: There are As Many Crimes as You Want, and Why That Matters

This matters because people get very angry about it. They claim the system is flawed or broken because huge amounts of crimes go unreported, or uncharged. They demand change.

…and my perhaps pedantic point is that we are always and forever not talking about some objective thing called “crime” that exists outside of abstract social convention, because ‘crime’ is not an objective fact.

A demand that “all crime be punished” would result in each and every one of us getting successfully prosecuted for dozens of crimes, most notably tax fraud.

So, when we demand an OBJECTIVE accounting of crimes, that they may be punished, we are in reality demanding an accounting of some very very small subdivision of all crime, “crimes I actually care about.”

But the problem is, that definition changes person to person. The entire sticking point of the debate about rising or falling crime rates is that a great many people DON’T care about some instances of shoplifting, or fraud, or assault, or drug possession. You don’t, and I don’t. All of us, every single one, looks at some crime, actual crime, and goes “you know what, it’s not so bad that this particular crime didn’t get reported or charged or punished.”

We just probably don’t agree on which crimes merit that treatment. For that, we have to look OUTSIDE the law, to policy…because at the end of the day, which crimes are worth caring about is a political question, not a legal one.