We are called upon in this case to consider an issue of first impression in our Circuit: whether a state conviction for being a minor in possession of alcohol is "countable" under § 4A1.2(c) of the United States Sentencing Guidelines ("the Guidelines") for purposes of calculating a federal defendant's criminal history score. This question, in turn, requires us to adopt for the first time a standard for determining when a defendant's prior conviction is "similar" to an offense listed under § 4A1.2(c) such that the prior conviction may not be counted towards his criminal history score. We conclude that being a minor in possession of alcohol is similar to a juvenile status offense and cannot be counted, and we therefore VACATE the sentence of Defendant Charles Rogers Cole and REMAND the case to the district court for resentencing.

The factual and procedural history of this case is quite straightforward. Charles Rogers Cole ("Cole") was charged in a two-count indictment on September 16, 2003, in the Western District of Michigan. The first count alleged that Cole conspired to distribute and to possess with intent to distribute Ecstasy, in violation of 21 U.S.C. § 846, 21 U.S.C. § 841(a)(1), and 21 U.S.C. § 841(b)(1)(C). The second count alleged that Cole possessed with intent to distribute Ecstasy, in violation of 21 U.S.C. § 841(a)(1) and 21 U.S.C. § 841(b)(1)(C).

Pursuant to a plea bargain, Cole pleaded guilty to Count I, conspiracy to distribute and to possess with intent to distribute Ecstasy, and Count II was dismissed. The pre-sentence report prepared by a probation officer determined that under the United States Sentencing Guidelines, Cole's criminal history score was zero, with a resultant criminal history category of I. The government filed objections to the calculation of the criminal history category score, arguing that four state court convictions for being a minor in possession of alcohol were countable under the Guidelines and that Cole's criminal history score should therefore be four, placing him in criminal history category III.

At the sentencing hearing on May 24, 2004, the district court agreed with the government that Cole's convictions for being a minor in possession were countable under the Guidelines. However, the district court also concluded that assigning Cole to criminal history category III would "overrepresent[] the nature of these misdemeanor convictions" and determined that it was more appropriate to place Cole in criminal history category II. Cole was sentenced to sixty months imprisonment. He filed a notice of appeal to this Court on June 30, 2004, raising as his only issue the district court's use of his minor in possession convictions in calculating his criminal history score.1

We review a district court's legal interpretation of the Guidelines de novo. United States v. Edwards, 272 F.3d 812, 815 (6th Cir.2001) (citing United States v. Hurst, 228 F.3d 751, 756 (6th Cir.2000)). This standard of review is not altered by the decision of the Supreme Court in United States v. Booker, ___ U.S. ___, 125 S. Ct. 738, 160 L. Ed. 2d 621 (2005). United States v. Davidson, 409 F.3d 304, 310 (6th Cir.2005) ("We continue, in reviewing individual Guidelines determinations, to apply the standards of review we applied prior to Booker.")

II. Whether state court convictions for being a minor in possession of alcohol are countable under § 4A1.2(c) of the United States Sentencing Guidelines for purposes of calculating a defendant's criminal history score.

Section 4A1.2(c) establishes that sentences for "misdemeanor and petty offenses" shall be counted when determining a defendant's criminal history score, with two categories of exceptions, which are enumerated as follows:

(1) Sentences for the following prior offenses and offenses similar to them, by whatever name they are known, are counted only if(A) the sentence was a term of probation of at least one year or a term of imprisonment of at least thirty days, or (B) the prior offense was similar to an instant offense:

Driving without a license or with a revoked or suspended license

Hindering or failure to obey a police officer

Local ordinance violations (excluding local ordinance violations that are also criminal offenses under state law)

Trespassing.

(2) Sentences for the following prior offenses and offenses similar to them, by whatever name they are known, are never counted:

Vagrancy.

U.S. SENTENCING GUIDELINES MANUAL § 4A1.2(c).

As "minor in possession of alcohol" is not among the enumerated offenses listed under either § 4A1.2(c)(1) or (2), the task in this case is to determine whether it is "similar" to any of them. The Guidelines do not define "similar," and our Circuit, unlike several other circuits, has not previously considered how a court is to determine whether an offense for which a defendant was convicted is similar to any of the listed offenses. Cole presents two arguments regarding this issue: 1) he contends that being a minor in possession of alcohol is similar to a juvenile status offense, under § 4A1.2(c)(2), and 2) he argues that being a minor in possession of alcohol is similar to disorderly conduct or disturbing the peace, under § 4A1.2(c)(1). Because we conclude that being a minor in possession is similar to a juvenile status offense and therefore may never be counted in the calculation of a defendant's criminal history score, we decline to consider whether it is also similar to any of the offenses enumerated in § 4A1.2(c)(1).

Again, this Circuit has not prescribed a test to be applied in determining whether an offense for which a defendant has previously been convicted is similar to one of the offenses enumerated in § 4A1.2(c). The approaches taken by the other federal courts of appeals to determine whether an offense for which a defendant has previously been convicted is similar to one of the offenses listed under § 4A1.2(c) can be roughly categorized as three distinct tests, although we note that these tests are sometimes conflated and that some courts have taken different approaches in different cases. We will discuss each of these three tests in turn.

The Fifth Circuit enunciated a "multi-factor approach" in United States v. Hardeman, 933 F.2d 278 (5th Cir.1991). Pursuant to this approach, a district court endeavoring to determine whether an offense of prior conviction is similar to one of the offenses listed under § 4A1.2(c) should apply "common sense" in "rel[ying] on all possible factors of similarity, including a comparison of punishments imposed for the listed and unlisted offenses, the perceived seriousness of the offense as indicated by the level of punishment, the elements of the offense, the level of culpability involved, and the degree to which the commission of the offense indicates a likelihood of recurring criminal conduct." 933 F.2d at 282. The court concluded "if these tests for similarity, taken as a whole, indicate that the offense is, like the listed offenses, neither particularly serious nor likely to indicate recurring criminal conduct, then the defendant's prior offense should be excluded." Id.

A similar approach has been employed in at least one case in the Ninth Circuit, United States v. Martinez (Clyde), 905 F.2d 251 (1990). In that case, the Ninth Circuit instructed district courts to consider whether the offense of prior conviction provides a basis for predicting future criminal activity, whether the conduct involved is uniformly criminalized, and the degree of the severity of the penalty usually imposed for such conduct. 905 F.2d at 253.

The Second Circuit has adopted a multi-factor approach even broader than that announced in Hardeman. Concluding that in "using an indeterminate word like `similar,' the Commission signaled its intent to sacrifice some certainty for a commensurate amount of flexibility and resulting fairness," the court instructed district courts "to read `similar' to encompass the factors listed in Hardeman as well as any other factor the court reasonably finds relevant in comparing prior offenses and Listed Offenses." United States v. Martinez-Santos, 184 F.3d 196, 205 (2nd Cir.1999).

Finally, the First Circuit seems to have moved towards a multi-factor approach. Although its first pronouncement on this issue appeared to limit the similarity analysis "to the substance of the underlying offense," United States v. Unger, 915 F.2d 759, 763 (1st Cir.1990), the court more recently explained that Unger does not limit district courts to consideration of "the statutory elements of an offense as compared to those of an exception [listed under § 4A1.2]." Rather, the court explained that "[t]he statutory elements and the punishment provided for each offense are among the factors that deserve comparison in the assessment of similarity." United States v. May, 343 F.3d 1, 9-10 (1st Cir.2003) (emphasis added). For that proposition, the court cited to United States v. Reyes-Maya, 305 F.3d 362, 366 (5th Cir.2002), which applies the Hardeman approach.

The multi-factor approach has been the subject of criticism, both on the ground that it is difficult to apply many of the various factors enumerated by those courts that follow it, and on the broader philosophical basis that it has the effect of supplanting the Guidelines. The Fourth Circuit examined and rejected the multi-factor approach in United States v. Harris, 128 F.3d 850 (4th Cir.1997), specifically critiquing two factors. The court noted that in circuits that consider whether offense conduct is universally regarded as culpable, there is no clear answer to the question of "how many jurisdictions must regard the conduct as culpable or to what degree." Harris, 128 F.3d at 854. The court also took issue with the factor requiring examination of the likelihood of recurring criminal conduct indicated by the prior offense, noting that the "cases do not offer any unifying principle for how one offense, but not another, indicates a likelihood of future criminal conduct." Id. at 855.

With respect to a third factor used in the multi-factor approach, requiring a comparison of punishments, the Fourth Circuit had a broader critique: "To define `similar' as `similar punishments' would rewrite the Guidelines and bypass the framework created by the [Sentencing] Commission." Id. In this observation, the court echoed a concurring opinion by Judge Wallace in the Ninth Circuit's Martinez (Clyde) opinion. Judge Wallace wrote, "[a]s the majority sees it, the Commission enumerated the specific offenses merely to state a general principle that offenses which are not uniformly culpable or predictive of future criminal behavior are not counted. As I see it, however, the listed offenses serve a different purpose: they dictate that the sentencing court, in each case, will decide whether the defendant's prior offense is an offense identical or similar to each listed offense." Martinez (Clyde), 905 F.2d at 255 (Wallace, J., concurring). Judge Wallace noted that if the Sentencing Commission had wanted to announce a general principle like the one he ascribed to the majority, it could have simply done so rather than setting forth the lists contained in § 4A1.2(c)(2). Judge Wallace concluded that the majority's approach "writ[es] the [§ 4A1.2(c)(2)] list out of the Guidelines." Id.

In contrast to the very flexible multi-factor approach, two circuits have utilized a rigid "elements" test. As the Third Circuit has explained, pursuant to this approach, a court assessing the similarity between a prior offense and an offense listed under § 4A1.2(c) "focus[es] only upon the elements of the offenses as statutorily defined." United States v. Elmore, 108 F.3d 23, 27 (3d Cir.1997). The Fourth Circuit conducts that inquiry by asking whether the elements of the prior offense are "`nearly corresponding' [to] or `resembling in many respects'" the elements of the listed offense. Harris, 128 F.3d at 854 (quoting Black's Law Dictionary 1240 (5th ed.1979)). See also United States v. Tigney, 367 F.3d 200 (4th Cir.2004).

The elements approach was criticized by the Second Circuit in Martinez-Santos as "too rigid." 184 F.3d at 205.

A third approach, which we adopt today, is known as the "essential characteristics of the crime" test. United States v. Perez de Dios, 237 F.3d 1192, 1197-99 (10th Cir.2001). Although some observers seem to conflate this test with the elements test,2 as articulated by the Tenth Circuit, there is an important and obvious distinction. While both tests (appropriately, in our view) restrict analysis to a comparison of the actual conduct underlying the offenses being compared, the essential characteristics approach is not limited to an examination of the technical elements making up the offenses, but rather, simply considers whether the underlying conduct of the offenses is "`[n]early corresponding; resembling in many respects; somewhat alike; having a general likeness.'" Perez de Dios, 237 F.3d at 1198 (quoting United States v. Mitchell, 941 F.2d 690, 691 (8th Cir.1991)). The Tenth Circuit explained its decision to depart in this respect from the elements test: "the key terms of § 4A1.2(c) — `similar' and `by whatever name they are known' — give courts more latitude than merely comparing legal elements of offenses." Id.

Similar to the Tenth Circuit's "essential characteristics" test is a test adopted by the Ninth Circuit, distinct from its iteration of the multi-factor test, which "focus[es] on whether the activity underlying [the prior offense] is similar to the activities underlying the listed offenses." United States v. Martinez (Carlos), 69 F.3d 999, 1000 (9th Cir.1995). In subsequent cases, the court has applied both standards. See, e.g., United States v. Sandoval, 152 F.3d 1190 (9th Cir.1998).

As noted, we adopt today the approach articulated by the Tenth Circuit in Perez de Dios. In explaining our rationale for that decision, we will first explain why both the multi-factor approach and the elements test are, in our view, unsatisfactory.

We agree with the criticisms of the multi-factor approach discussed above. Specifically, we share the concern of the Fourth Circuit that the approach is, in some respects, unworkable. For example, in the absence of hard data relating to recidivism rates for various offenses, we would be hard-pressed to weigh the likelihood of future criminal conduct indicated by a prior offense without resort to speculation and guess-work. We also agree with Judge Wallace of the Ninth Circuit that the multi-factor approach essentially supplants the lists set forth at § 4A1.2(c), thus "depart[ing] from the basic canon of statutory interpretation that a court should not interpret a statute to render meaningless certain parts of that statute." Martinez (Clyde), 905 F.2d at 255 (Wallace, J., concurring) (citations omitted).

While we find that the multi-factor approach is overly flexible, we also regard the elements test as excessively rigid and technical. Just as we presume that the Sentencing Commission would have instructed courts to examine, for example, the respective punishments available for the prior and listed offenses if it had deemed such a consideration relevant, we see no reason that the Sentencing Commission would not have explained that "similar" meant "similar elements" if that is what it intended. We also note a practical problem in the application of the elements test: how many elements must be similar in order to conclude that the offenses are similar? Suppose, for example, that the prior offense consists of five elements and the listed offense of four, with a total of three shared elements. Does that suffice to deem the offenses similar? What if only two elements are common to both offenses? We simply do not know how this question is to be answered by courts following the "elements" approach. Finally, we are concerned that there may be cases in which a strict comparison of the elements of the offense of prior conviction and of the offense listed under § 4A1.2(c) may not reveal substantive similarities that exist between the two offenses, or may wrongly suggest a similarity that is more apparent than real.3

We prefer the "essential characteristics of the crime" approach because it preserves a focus on the activity underlying the offenses, which we regard as required by the language of § 4A1.2(c), while permitting a court to take a common sense approach to determining when the conduct underlying the prior offense is similar to the listed offense. We hold that district courts considering the issue of similarity under § 4A1.2(c) should consider the similarity between the "essential characteristics" of the activity underlying the offense of prior conviction and those underlying the listed offense.

We now must consider whether, under the test just enunciated, being a minor in possession of alcohol is similar to a juvenile status offense. Our fellow circuits appear to be in agreement that in comparing offenses, a court should look to the state statutory definition of the offense of prior conviction, but consider federal law in defining the listed offense under § 4A1.2(c). See, e.g., Elmore, 108 F.3d at 25; Unger, 915 F.2d at 762-63; Martinez (Clyde), 905 F.2d at 253; Borer, 412 F.3d at 992; Tigney, 367 F.3d at 202.

With respect to the state definition of the prior offense, the Michigan offense of being a minor in possession of alcohol prohibits a person under the age of twenty-one from possessing or attempting to possess alcohol. M.C.L. § 436.1703(1). What constitutes a "juvenile status offense" is a more difficult question, as it appears not to refer to a specific offense but to a group of offenses. This Court (in rejecting a defendant's argument that manslaughter and assault, if committed as a juvenile, constituted juvenile status offenses) has looked to Black's Law Dictionary in considering this question, citing the following definition of "status crime": a "`class of crime which consists not in proscribed action or inaction, but in the accused's having a certain personal condition or being a person of a specified character. An example of a status crime is vagrancy.'" United States v. Williams, 176 F.3d 301, 312 (6th Cir.1999). The First Circuit, in considering what constitutes a juvenile status offense, offered the following examples of status offenses: "hitchhiking, truancy, loitering, or vagrancy." Unger, 915 F.2d at 763. Given that all four of those offenses are listed separately from juvenile status offenses under § 4A1.2(c)(2), and none of them specifically references an offender's age, we think that "juvenile status offense" must entail something more, and that it must specifically relate to a defendant's age.

The Seventh Circuit has determined that "the obvious meaning [of juvenile status offense] is conduct that would be lawful for an adult and is unlawful solely by virtue of the defendant's juvenile status," giving as examples "the purchase of liquor or cigarettes." United States v. Ward, 71 F.3d 262, 263 (7th Cir.1995). Citing in part to Ward, the First Circuit concluded that a crime "constitutes a juvenile status offense only if three elements coalesce": 1) the defendant was a juvenile (i.e., under the age of eighteen) at the time of the offense; 2) the "conduct would have been lawful if engaged in by an adult"; and 3) "the offense is not serious." United States v. Correa, 114 F.3d 314, 318-19 (1st Cir.1997).

The Eighth Circuit, which to our knowledge is the only Circuit to have addressed the issue of whether a minor in possession conviction constitutes a juvenile status offense, applied the three-factor test set forth in Correa and concluded that a defendant's conviction for "prohibitions" under Ohio law (effectively the same crime as Michigan's "minor in possession of alcohol") was not a juvenile status offense because the defendant was nineteen at the time the offense was committed. United States v. Webb, 218 F.3d 877, 880 (8th Cir.2000). Interestingly, and for reasons that are not clear, the court did not consider whether the defendant's conviction for "prohibitions" at the age of nineteen was similar to a juvenile status offense, despite the fact that § 4A1.2(c)(2) bars the use of a juvenile status offense or an offense similar to a juvenile status offense in determining a defendant's criminal history score.

We are persuaded by Cole's argument that his convictions for being a minor in possession of alcohol, all obtained when he was nineteen years old, were similar to a juvenile status offense. Applying the definition offered in Correa, we note that a juvenile status offense is an offense which is illegal only because of the offender's age (under eighteen) and which is not serious. Under Michigan law, being a minor in possession of alcohol is an offense which is criminalized only because of the offender's age (under twenty-one). It is not a "serious" offense.4 The crux of both offenses is that they criminalize behavior for people of a certain age that is legal for people of a certain (greater) age. Indeed, in cases involving defendants under the age of eighteen, being a minor in possession of alcohol would not simply be similar to a juvenile status offense, it would constitute a juvenile status offense. We find it difficult to imagine two distinct offenses with more similar "essential characteristics," and have no difficulty concluding that being a minor in possession of alcohol is similar to a juvenile status offense, under § 4A1.2(c)(2).

Because, under § 4A1.2(c)(2), criminal history points can never be assessed for offenses similar to juvenile status offenses, we VACATE Cole's sentence and REMAND this case to the district court for re-sentencing, with instructions that the court calculate Cole's criminal history score without counting his convictions for being a minor in possession of alcohol.