The Categorical Approach and Comparing a Statute to a Definition

An immigration ground that depends on a conviction asks a narrow question: was this person convicted of an offense of a particular kind. Answering it does not involve reading a police report or hearing what happened. It involves putting two texts side by side, the statute the person was convicted under and the federal definition, and asking whether one fits inside the other.

The two texts

On one side is the statute of conviction, stripped down to its elements: the things a prosecutor had to prove, or that the person had to admit, to be convicted at all. On the other is the generic federal definition of the immigration category, which may be spelled out in the statute or may have to be constructed from the ordinary meaning of the term.

The question is whether every conviction under the state statute necessarily satisfies the federal definition. If it does, the conviction is a categorical match and the ground applies. If the state statute reaches any conduct the definition does not, there is no match, and no conviction under that statute qualifies.

Note what that last sentence means. It is not a case-by-case inquiry. Where the statute is overbroad, everyone convicted under it falls outside the federal category, including people whose actual conduct sat comfortably within it.

The least of the acts

The comparison is made at the bottom of the statute rather than in the middle. The Supreme Court has directed that the conviction be presumed to have rested on nothing more than the least of the acts criminalized by the statute.

So the working method is to find the least serious conduct a person could be convicted of under the statute, using its text and the state's own case law interpreting it, and to ask whether that conduct meets the federal definition. A theft statute that reaches temporary takings, a drug statute that lists a substance not on the federal schedules, or an assault statute that reaches reckless conduct can each fail the comparison on that basis.

The government's standard response is that a merely theoretical reading is not enough and that the person must show a realistic probability of the statute actually being applied to the broader conduct, usually by producing a case in which it was. Courts have not treated that requirement identically, and several have declined to apply it where the statutory text plainly reaches further on its face.

Where the method applies

The method governs wherever an immigration consequence turns on a conviction of a described offense. That covers the aggravated felony list, the controlled substance grounds, the firearms ground, the domestic violence ground and the crimes involving moral turpitude grounds, and it applies both to grounds of removability and to the bars on relief that use the same categories.

It also applies to definitions built with numerical thresholds, though those work differently. Where the statute attaches a sentence length or a loss amount to the category, the threshold is treated as a circumstance-specific fact rather than as an element, and it is proved differently, a distinction taken up in theft, fraud and the loss amount threshold.

The controlled substance grounds carry their own version of the problem, because a state schedule that lists a substance not federally controlled makes the state statute broader than the federal definition. That analysis is worked through in drug convictions and the one narrow exception.

When the statute lists alternatives

Many criminal statutes are written with lists. A statute may prohibit entering a building, a vehicle or a boat; or may prohibit distributing any of a long list of substances. When that happens the first question is whether the list sets out separate elements, creating several distinct offenses, or merely different means of committing one offense.

The distinction decides whether any document from the criminal case may be consulted at all. Where the list is elements, a limited set of records may be examined to determine which offense the person was convicted of. Where the list is means, the statute is indivisible and the record stays closed, however clearly it describes what happened. That mechanism, and the documents that may be read, are covered in divisible statutes and the documents a judge may read.

Using the method in a real case

The work order is consistent. Obtain a certified copy of the conviction record and identify the exact subsection of the exact statute, in the version in force at the time of the offense. Break that subsection into its elements. Identify the federal definition being asserted. Compare them, at the least conduct the state statute reaches. Then check whether the state's own courts have construed the statute more broadly or more narrowly than its text suggests.

Two practical cautions attach to that sequence. The first is that the version of the statute matters: legislatures amend criminal codes constantly, and comparing the current text against an older conviction produces the wrong answer. The second is that a conviction record naming only a chapter and section number is not enough, because most sections contain several subsections with materially different elements.

The third caution is about discipline. Once the elements are identified the analysis is closed, and information about what happened, however compelling, does not enter it. Volunteering that information at a hearing puts facts into a record that has no use for them and can damage the discretionary assessment that comes later.

Only after all of that does anything about the person's actual conduct matter, and often it never does. This is why a conviction that sounds serious can be immaterial and one that sounds trivial can be disqualifying, and why the analysis is done before removability is conceded rather than after, as set out in pleading to the allegations and conceding removability. Where the statute is close to the line, obtaining the right records and reading the state case law is detailed work, and a categorical approach immigration attorney is generally the difference between an argument and a concession.