Immigration and Crime, Eight Popular Claims Checked From Both Sides

Every few months a single horrible crime makes the news, the suspect turns out to be in the country unlawfully, and the argument begins. One side says it proves the whole system is a menace. The other side says it proves nothing, because immigrants commit fewer crimes than everyone else. Each side then spends a week talking past the other.

Both are doing something worse than being wrong. They are refusing to ask what kind of claim they are making. A claim about an average and a claim about a case are different animals, and the whole fight turns on keeping them apart. So let us do the thing that is rare on cable: test eight popular claims, from both directions, and give a verdict on each.

The ground rules

Before the claims, three definitions and a warning.

An average claim says something about a group as a whole: that the rate of conviction among one population is higher or lower than among another. It is tested with data over many people.

A case claim says something about one event or one failure: that this man should not have been on the street, that this agency should have acted. It is tested with the facts of that case.

A policy claim says what government ought to do. It rests on values as well as facts.

Here is the warning. The data are imperfect. Few jurisdictions record immigration status when they record an arrest or a conviction, the status of the person is often not known at arrest, and the number of people in the country unlawfully is itself an estimate. Any honest analysis starts by saying so. I will point to the kinds of records available, and I will tell you where each is limited, and I will leave out figures I cannot stand behind.

Claim one: "Immigrants commit more crime than native-born Americans."

Where it comes from. It is the instinctive claim of people frightened by a particular case, and it appears in speeches with the confidence of a fact.

The test. The best-known source is state conviction data, and the most-cited example is Texas. The Texas Department of Public Safety records the immigration status of people it has arrested and convicted, checking them against federal records. Researchers who have analyzed those records have generally found that the conviction rate per capita among people in the country illegally is lower than among native-born Americans. Other analysts, including researchers at the Center for Immigration Studies, reply that the state data are incomplete, that people convicted of crimes can be missed or counted differently, that the estimate of the unlawful population in the denominator is uncertain, and that the data count convictions, not arrests or victims.

Verdict. As a claim about the average, it is not supported by the best state records we have. The honest conservative position is to concede the point about averages and move to the claims that survive, below. A movement that wants to win arguments should not stake its credibility on a claim its own best data contradict.

Claim two: "Immigrants commit less crime, so there is nothing to see."

Where it comes from. The mirror image. It is offered to close a conversation.

The test. Suppose the average is lower. What follows? Very little. An average tells you nothing about the particular man who has been removed five times and is in the country again, or about the victims of a crime that a prior removal would have prevented. The rate at which some group commits crime is not the same as the number of crimes that were preventable. If a single preventable murder happens because a government did not do what the law told it to do, the fact that the average rate is low is no comfort to the family. A country's duty to its citizens does not shrink because the odds are good.

There is a second weakness. The people most likely to be counted as "immigrants" in a conviction dataset include a large population of lawful, settled residents and naturalized citizens, who have every incentive to follow the law. Lumping them in with people who entered unlawfully obscures the question the policy debate is about.

Verdict. True as an average, if the data hold, and irrelevant to the policy question. A low average is a reason to avoid collective suspicion of any group. It is not a reason to stop enforcing the law against individuals who have committed serious crimes.

Claim three: "Crossing the border illegally is only a civil matter."

Where it comes from. A common talking point in debates about enforcement.

The test. The statute is where to look. Under Title 8 of the United States Code, section 1325, entering the country at a place other than a designated port of entry, or eluding inspection, or entering by false statements is a federal crime. A first offense is a misdemeanor, punishable by up to six months. Section 1326 covers re-entering after being removed, which is a felony and carries a longer sentence, with penalties that rise if the person had prior felony convictions.

The nuance is the one the Supreme Court set out in Arizona v. United States in 2012, where Justice Kennedy's opinion observed that as a general rule, it is not a crime for a removable alien to remain present in the United States. In other words, the entry can be a crime; the continued presence without status is generally a civil violation, handled in removal proceedings.

Verdict. The claim is half right and half wrong. Illegal entry is a federal crime, and re-entry after removal is a felony. Remaining without status is, as a general rule, a civil matter. A speaker who says "it's just paperwork" and a speaker who says "every person here unlawfully is a criminal" are each blurring a line the statute draws carefully. The reason to enforce the entry statute is the same as the reason to enforce any other: a law with no consequences is a suggestion. See The Employer Who Hires Off the Books Is Breaking the Law, Not Helping for the same point on the employer's side.

Claim four: "Most of the people prosecuted federally are violent criminals."

Where it comes from. A reading of federal statistics that treats every case as a crime of violence.

The test. The U.S. Sentencing Commission publishes data on federal sentences each year. One consistent finding is that immigration offenses, chiefly illegal entry and illegal re-entry after removal, make up a large share of federal criminal cases. That is a fact about what federal prosecutors are charging. It is a fact about immigration offenses, though, not about violence. A person sentenced under section 1326 for re-entering after removal has committed the offense of re-entry; that does not mean he committed a violent offense. Counting those cases as evidence of a crime wave misreads the statistic.

The same point cuts in the other direction. The share of noncitizens in the federal prison population is inflated by exactly these immigration offenses, which can only be committed by noncitizens. Using it to show that noncitizens commit more crime generally is circular.

Verdict. Misleading from both sides. The figures are real but they measure the enforcement of immigration law, and they cannot carry the weight either camp puts on them.

Claim five: "Sanctuary policies make communities safer."

Where it comes from. The standard defense of local rules that limit cooperation with federal immigration authorities.

The test. The argument is that people who fear deportation will not report crimes or testify, so refusing to cooperate with immigration enforcement builds trust and reduces crime. It is a plausible argument and it has defenders among police chiefs. The counterargument is that when a local jail declines to honor a federal detainer, a person already in custody for a crime can be released instead of transferred, and that some of those released will offend again. A federal detainer is a request, not a warrant, and the legal fight about what jails may or must do is a real one; I walk through the mechanics in What an ICE Detainer Actually Asks a Local Jail to Do and Why It Matters.

What the evidence shows depends heavily on which study you read and how it defines a sanctuary jurisdiction, and the studies point in different directions. That should humble everyone.

Verdict. Unproven in both directions. A claim that sanctuary policies make communities "safer" has not been established, and neither has the claim that they cause a surge in crime. The strongest conservative argument is not statistical. It is about law and accountability: a person who has been convicted of a serious crime and is subject to removal should be transferred, not released, and the local officials who decline should be answerable to voters for the result.

Claim six: "Cases of released offenders are rare flukes."

Where it comes from. The defense of the system when a case makes headlines.

The test. Here the case claim matters, and I want to be careful, because I will not pretend that one case defines a population. Consider the 2015 shooting on a San Francisco pier. The man charged had been removed from the country several times before and had been released from a county jail despite a federal detainer. A jury in 2017 acquitted him of murder and manslaughter and convicted him only of being a felon in possession of a firearm, a result that surprised many and that shows why the trial record, not the headline, should govern what we say about a case.

The Laken Riley Act, signed in January 2025 after the killing of a student in Georgia, requires federal detention of unlawfully present people who are charged with certain offenses such as theft. Whatever you think of its sweep, it was Congress's response to a pattern in which people with criminal charges were released and then went on to offend.

What do such cases prove? Not that immigrants are dangerous. They prove that a government that knows a person is a danger, and has the legal power to hold or remove him, can fail to do so, and that the failure has victims. That is a claim about competence and accountability, and it is true whatever the averages say.

Verdict. Both sides are half right. The cases are not a statistical argument. They are a governance argument, and they are strong ones.

Claim seven: "Smuggling is a victimless service."

Where it comes from. The notion that a person who pays for passage and the person who sells it are engaged in a voluntary exchange.

The test. The federal smuggling statute, 8 U.S.C. 1324, criminalizes bringing in, transporting and harboring people who are in the country unlawfully, and prosecutions under it are a regular part of border-state federal dockets. The record of what happens on the smuggling routes includes extortion, abandonment, deaths in the desert and in trucks, and the exploitation of those who cannot pay in cash. A smuggler's customer arrives owing a debt to people who have every reason to collect it. I tell the story of the economics in How a Smuggling Network Prices a Border Crossing and Who Gets Paid.

Verdict. False. The people who suffer most from an unsecured border are often the migrants themselves, which is a moral argument for enforcement that tends to be missing from the debate. A policy that lets the cartels set the price and the terms of entry is no kindness to anyone.

Claim eight: "Crime fell while immigration rose, so immigration lowers crime."

Where it comes from. A tidy-looking chart.

The test. National crime rates did fall sharply from the early 1990s into the following decades, and the unlawful population grew over part of that period. Two lines moving in opposite directions look like cause and effect. They are not proof of it. Many other things were changing at the same time: policing methods, incarceration rates, the age structure of the population, the economy, and more. To claim that immigration caused the decline you would need a design that holds those other factors constant, and a time-series chart does not do that.

The same discipline applies in reverse. A conservative who points to a rise in a particular crime in a particular year and blames immigration has made the same mistake in the other direction.

Verdict. Correlation, not causation. Distrust any argument that can be summed up by a chart with two lines.

What the eight verdicts add up to

Line them up and a pattern appears.

1. The claim that immigrants as a group commit more crime does not hold up against the best records we have.
2. The claim that this settles the policy question does not hold up either, because the policy question is about individuals, rules and accountability.
3. Illegal entry is a federal crime, and re-entry after removal is a felony, while mere presence is generally civil. Precision serves the argument.
4. The cases of released offenders are real, and they are failures of government. They deserve answers and legal fixes, not an averages argument.
5. The smuggling economy is a criminal enterprise that preys on migrants as well as on border communities, and it is the strongest humanitarian argument for a secure border.

A conservative should be able to hold all five at once. We can say that most immigrants are not criminals, because the evidence suggests it, and we can say that a government has a duty to remove those who are, and not to release people it has the power to hold. Both statements are true, and putting them together is the discipline that separates an argument from a slogan. For a different angle on how Americans have written this distinction into law, see The 1986 Law's Real Failure Was the Fake Document Industry It Created.

A note on the Founders

Why does the law treat the border as a legal line at all? Because the first duty of government, as the Declaration of Independence puts it, is to secure the rights of its people: "to secure these rights, Governments are instituted among Men." The Preamble to the Constitution lists "insure domestic Tranquility" and "provide for the common defence" among the purposes of the Union. A government that does not control who enters, and cannot remove those who have forfeited a claim to stay, is failing at the first of its jobs.

That is a principle, not an average. It does not depend on whether the people crossing are more or less likely to commit crimes than the rest of us. It depends on whether a nation has the capacity to decide.

The case in brief

1. The best available state records, including Texas conviction data, do not support the claim that immigrants as a group commit more crime than native-born Americans, though the data have limits.
2. A low average does not answer the policy question, because the duty to remove those who have committed serious crimes is about individuals and accountability.
3. Illegal entry (8 U.S.C. 1325) is a federal crime and illegal re-entry after removal (1326) is a felony; remaining without status is generally a civil matter, as Arizona v. United States observed.
4. Cases of released offenders are failures of enforcement and local cooperation, and they justify legal fixes whatever the averages say.
5. Smuggling networks prey on migrants and border communities alike, which makes a secure border a humane policy as well as a lawful one.

What you can do: look up your state's published criminal-justice data and ask your sheriff or county jail, in a public meeting, how it handles federal detainers, then compare the answer with what your state legislature has set in law.