concurring in part, dissenting in part and concurring in the judgment.
Sometimes hard cases make bad law. And sometimes easy cases make bad law. Only rarely, however, do easy cases make bad law by overruling good law.
This is an easy case, and it is one that we readily can resolve without overruling one of our precedents — United States v. Foreman, 436 F.3d 638 (6th Cir.2006). Since Foreman, the Supreme Court has decided three crime-of-violence cases, and none of them compels us to give the last rites to Foreman. One applies the same approach as Foreman. See James v. United States, 550 U.S. 192, 207-09, 127 S.Ct. 1586, 167 L.Ed.2d 532 (2007). And the other two bolster the result in Foreman by making it more difficult to establish that a defendant’s prior conviction is a crime of violence. See Chambers v. United States, — U.S. -, 129 S.Ct. 687, 690-92, 172 L.Ed.2d 484 (2009); Begay v. *382United States, 553 U.S. --•, 128 S.Ct. 1581, 1586, 170 L.Ed.2d 490 (2008). Why the majority must overrule Foreman, as opposed to registering disagreement with it, is a mystery.
This premature overruling of one of our precedents stands in stark contrast to a recent decision of the court. In resolving a difficult death-penalty case, Getsy v. Strickland, all three members of the panel registered their disagreement with the leading precedent on point, Cooey v. Strickland, 479 F.3d 412, 422 (6th Cir. 2007) (Cooey II). See 577 F.3d 309, 313 (6th Cir.2009); id. at 313-14 (Moore, J., concurring); id. at 314-16 (Merritt, J., dissenting). Yet, to its credit, the panel stood by Cooey II, allowing that precedent to survive, even when it meant that Mr. Getsy might not. See Getsy v. Strickland, No. 09-5935, — U.S.-, 130 S.Ct. 40, 174 L.Ed.2d 618, 2009 WL 2490098 (Aug. 17, 2009). Perhaps I am missing something, but I fail to understand why we cannot take the same path here. This, indeed, is the easier case: There is less at stake, and a faithful adherence to Foreman does not even alter the result.
Consider this straightforward way to resolve the case, one that does not require us to overrule Foreman and one that still allows the majority to conclude that Young’s fleeing-and-eluding conviction was a crime of violence. As the majority correctly notes, everything in this case depends on whether fleeing and eluding under Michigan law falls within ACCA’s catchall clause for crimes that “otherwise involve[] conduct that presents a serious potential risk of physical injury to another.” 18 U.S.C. § 924(e)(2)(B)(ii). And, as the majority correctly adds, a crime falls within that clause if (1) it presents a serious potential risk of injury akin in degree to the enumerated offenses — burglary, arson, extortion or the use of explosives, see 18 U.S.C. § 924(e)(2)(B)(ii) — see James, 550 U.S. at 202-09, 127 S.Ct. 1586, and (2) it involves the same kind of “purposeful, violent, and aggressive” conduct as the enumerated crimes, Begay, 128 S.Ct. at 1586; see also Chambers, 129 S.Ct. at 692.
In three eases, we have applied this framework to Michigan’s fleeing-and-eluding statute. In the first, United States v. Martin, we held that convictions involving either of two aggravators — that the “violation results in a collision or accident,” Mich. Comp. Laws. § 257.602a(3)(a), or “a portion of the violation occurred in an area where the speed limit is 35 miles an hour or less,” id. § 257.602a(3)(b) — present a serious potential risk of injury. See 378 F.3d 578, 582-84 (6th Cir.2004). We reasoned that both the flight itself and the eventual apprehension “create[ ] a conspicuous potential risk of injury” and committing the offense through either enhancement “confirms the palpable risk.” Id. at 582. We also analogized fleeing and eluding to escape' — which we previously recognized as a “crime of violence” under the Sentencing Guidelines, see United States v. Harris, 165 F.3d 1062, 1068 (6th Cir.1999), modification recognized in United States v. Ford, 560 F.3d 420, 423 (6th Cir.2009)— as both involve heightened emotions and usually end with a confrontation between the officer and the suspect. Martin, 378 F.3d at 582-83. In the end, we concluded that fleeing and eluding poses a greater risk than escape because fleeing and eluding always involves “flight [from officers] and the inherent third-party risks that such conduct entails.” Id. at 583.
In the second, Foreman, we addressed whether a fleeing-and-eluding violation amounted to a crime of violence if it did not involve any aggravating circumstances. We acknowledged that such conduct, like the conduct in Martin, creates a potential risk of physical injury to another. 436 F.3d at 642-43. But we decided it was *383only “the presence of ... fleeing in a low speed zone or causing an accident” that enabled us categorically to characterize the fleeing-and-eluding violation in Martin as ordinarily posing “a serious potential risk.” Id. at 643 (emphasis in original). We therefore held that the risk of injury was not ‘^serious” enough to amount to a crime of violence.
In the third, United States v. LaCasse, we considered whether the Supreme Court’s intervening decisions in Begay and Chambers altered these analyses. See 567 F.3d 763, 764 (6th Cir.2009). At issue was whether a typical third-degree fleeing-and-eluding conviction involves the type of purposeful, violent or aggressive conduct targeted by ACCA. Id. at 765-67. As laid out in the current statute, a third-degree offense involves the same conduct as a fourth-degree offense plus the presence of one of three aggravators: the two present in Martin or a prior conviction for fleeing and eluding. See Mich. Comp. Laws. § 257.602a(3) (2008). Michigan’s third-degree fleeing-and-eluding statute, we reasoned, requires affirmative, aggressive conduct distinguishable from the strict liability drunk-driving offense in Begay. LaCasse, 567 F.3d at 766. And Chambers, which held that a “failure to report” to prison was “a separate crime, different from escape” and not a “violent felony” under ACCA, 129 S.Ct. at 691-93, gave “added force” to this conclusion. LaCasse, 567 F.3d at 766-67. Chambers’ distinction between a prison breakout and a failure to report suggested that a traditional escape remained a “violent felony” involving the same type of purposeful, violent, or aggressive conduct targeted by ACCA. Id. at 767. And since “fleeing and eluding [is] but an attempt to escape,” we reaffirmed that a third-degree fleeing-and-eluding conviction remained categorically a “violent felony” under ACCA. Id. at 764, 767.
Taken together, these three decisions create the following line: A fourth-degree offense {e.g., a first-time offender) is not a crime of violence (Foreman); and at least two types of third-degree offenses {e.g., one that causes physical injury or occurs in a thirtyfive-mph zone) are crimes of violence {Martin and LaCasse). That leaves today’s question: Is the last type of third-degree offense, in which a defendant has committed fourth-degree fleeing and eluding more than once in five years, a crime of violence? Yes, I say, and the basis for saying so is straightforward and respectful of all three precedents: simply draw a line between run-of-the-mine violations of the statute (treated as non-violent crimes) and aggravated violations of the statute (treated as violent crimes). That is precisely what LaCasse seemed to do with the issue. And that, it seems to me, should end this case, particularly since it leads to the same result that the majority’s approach does. If the majority also wishes to register its disagreement with Foreman, it is free to do so and later panels (or the en banc court) can take that criticism into consideration. But nothing in this case presents a basis for overruling the decision, much less compels us to do so.
In maintaining that Foreman must be overruled, the majority notes that ‘Young” — the losing appellant in this case — “is correct that if Foreman remains good law, his sentence should be reversed.” Maj. Op. at 379 n. 4. That is not a recognized exception to stare decisis. A litigant’s position in a case, least of all the position of the losing litigant, does not require us to overrule this precedent, particularly since Young would lose whether Foreman remains good law or not.
Nor is Foreman inconsistent with Martin or with intervening Supreme Court precedent. Foreman’s holding does not contradict Martin’s holding. Third-degree *384and fourth-degree fleeing and eluding are separate offenses, and Martin addressed the only offense before it: third-degree fleeing and eluding. See 378 F.3d at 582.
The Supreme Court’s recent decisions also do not require the overruling of Foreman. Begay and Chambers both make it harder, not easier, to classify an offense as a crime of violence, because they both held (as Foreman held) that the offenses before them were not crimes of violence. See Begay, 128 S.Ct. at 1586, 1588; Chambers, 129 S.Ct. at 692-93. And Foreman applied the categorical approach, including the requirement that courts look to the ordinary manner in which an offense is committed, adhered to in James and required by Taylor v. United States, 495 U.S. 575, 110 S.Ct. 2143, 109 L.Ed.2d 607 (1990), which predates Foreman. Nothing, in short, justifies abandoning our custom of respecting the holdings of the panels that come before us. See Blair v. Henry Filters, Inc., 505 F.3d 517, 526 (6th Cir.2007).
If there was a time to overrule Foreman (and I don’t think there was), it was in LaCasse, not this case. LaCasse itself suggests, if not explicitly adopts, the same line I propose. If we now have the right to overrule Foreman, so did the LaCasse panel. But it left the decision as is. We instead should leave all three decisions (Foreman, Martin and LaCasse) where we found them, standing by the sensible line they draw: between fleeing-and-eluding offenses that involve an aggravating circumstance and those that do not.
I also see no cognizable reason to treat Young’s recidivist offense differently from the other aggravating circumstances. On the contrary: there is good reason to think that it poses more risk of physical injury and is more violent than at least one of the others: the thirty-five-mph-zone aggravator. Many people, I suspect, would prefer to take their chances in the vicinity of a first-time thirty-five-mph violator than in the vicinity of someone who has already been arrested for fleeing and eluding before and thus knows that what they are doing is wrong and has embraced the risks nonetheless.
Surely, as a general matter, it also makes sense to treat repeat offenders differently from first-time offenders. A central premise of the “career offender” component of ACCA — the law we interpret today — -is that recalcitrant law-breakers pose risks to the public that are different in kind and blameworthiness from initial lawbreakers. In the absence of that policy judgment, ACCA would not exist.
Stare decisis provides one reason for the approach I recommend. The reality that the outcome of this case would be the same, whether Foreman is good law or bad, provides another. But it makes the least sense to overrule a precedent gratuitously in such an evolving area of federal law. “[T]he mire into which ACCA’s draftmanship and Taylor1 s ‘categorical approach’ have pushed us,” Chambers, 129 S.Ct. at 694 (Alito, J., concurring), is consuming an increasing portion of our docket. This is the fourth case in five years dealing with one statute from one State, with each case asking us to slice the statute into thinner analytical pieces. Outside our circuit, a deep split has arisen over whether similar fleeing-and-eluding offenses present the requisite risk of injury, compare United States v. Kendrick, 423 F.3d 803, 809 (8th Cir.2005), and United States v. Howze, 343 F.3d 919, 921-22 (7th Cir.2003), with United States v. Kelly, 422 F.3d 889, 893-95 (9th Cir.2005), or the requisite type of aggressive conduct, compare United States v. West, 550 F.3d 952, 968-71 (10th Cir.2008), and United States v. Spells, 537 F.3d 743, 752-53 (7th Cir. 2008), with United States v. Harrison, 558 F.3d 1280, 1295-96 (11th Cir.2009). It is *385one thing to enter this quagmire because a case requires it; it is quite another to re-shift the shifting sand when nothing demands it.
There is one nuance about Foreman that deserves mention, but it does not require its overruling. When we hold that an offense is not categorically a crime of violence, we often remand the case to the district court so that the Government may establish through Shepard sources that the defendant’s conduct was a crime of violence, as we did in Foreman, 436 F.3d at 643, and as we have done in other cases, see, e.g., United States v. Mosley, 575 F.3d 603, 608 (6th Cir.2009); United States v. Baker, 559 F.3d 443, 455 (6th Cir.2009); United States v. Armstead, 467 F.3d 943, 948-50 (6th Cir.2006). This type of remand should not be done reflexively, however, for it does not always make sense, and in some settings it may even suggest that the court is embracing a non-categorical approach to the issue. If, say, the state law offense has just one category, or two categories that both amount to crimes of violence, nothing productive can come from a Shepard inquiry. By contrast, when an offense is divisible into two or more categories, and at least one category is a “crime of violence” and at least one is not, there is something the Shepard sources can reveal: that the defendant committed a type of offense that is a crime of violence. See Shepard v. United States, 544 U.S. 13, 19-21, 125 S.Ct. 1254, 161 L.Ed.2d 205 (2005); Chambers, 129 S.Ct. at 690-91.
The Shepard remand in Foreman illustrates the point. There are three ways to commit fourth-degree fleeing and eluding: “increasing the speed of the vehicle, extinguishing the lights of the vehicle, or otherwise attempting to flee or elude.” See Mich. Comp. Laws § 257.602a(l). Nothing in Foreman suggests that any one of these categories is a crime of violence. The point of the opinion was to say that all three offenses were not crimes of violence. Only confusion, I submit, comes from a Shepard remand in these circumstances. For in this setting, a remand suggests that we are asking the district court to do something that Taylor prohibits: engaging in a fact-specific inquiry about how this defendant committed this offense. Taylor, 495 U.S. at 600-02, 110 S.Ct. 2143. It also inadvertently suggests to district courts that we ourselves will ignore the categorical approach — and, worse, that we may reverse them for not doing so. As I see it, a Shepard remand should be issued only when the crime at issue contains at least two categories, at least one way of committing the offense is a crime of violence, and at least one way of committing the offense is not. But none of this justifies overruling Foreman, which at worst made a mistake that many of our precedents in this area have made. All we should do is sound a word of caution about Shepard remands and leave Foreman as we found it.
One final note. A court of this size cannot function if it does not stand by its prior decisions — if each panel is a court unto itself in deciding whether to adhere to a prior decision. Respect for our precedents, respect for each other and respect for the district court judges and litigants who must live with our decisions demands that each panel treat the overruling of precedents as a measure of last resort — an unavoidable act that intervening Supreme Court precedent (or something similar) compels. By any measure, that did not happen here. The majority seeing things differently, I respectfully dissent from this aspect of its decision and otherwise join the rest of it.