Treating the petition for rehearing en banc as a petition for panel rehearing, the petition for panel rehearing is DENIED. The court having been polled at the request of one of the members of the court, and a majority of the judges who are in regular active service not having voted in favor (see Fed. R. App. P. 35 and 5th Cir. R. 35), the petition for rehearing en banc is DENIED.

JERRY E. SMITH, Circuit Judge, with whom JONES and DeMOSS,Circuit Judges, join, dissenting from the denial of rehearing en banc:

The refusal of the en banc court to rehear this case en banc is unfortunate, for this is an opportunity to revisit the questionable practice of denying precedential status to unpublished opinions. Although I believe the panel reached a correct result, based primarily on the test set forth in Clark v. Tarrant County, 798 F.2d 736 (5th Cir. 1986), I respectfully dissent from the denial of rehearing en banc, which would have given this court an opportunity to examine the question of unpublished opinions generally, an issue that is important to the fair administration of justice in this circuit.

Consider what has occurred in this case. In Anderson v. DART, 180 F.3d 265 (5th Cir.) (per curiam) (unpublished) (table), cert. denied, 528 U.S. 1062 (1999), a panel of this court affirmed, "for essentially the reasons stated by the district court in its comprehensive and well-reasoned opinion," the judgment in Anderson v. DART, No. CA3:97-CV-1834-BC, 1998 U.S. Dist. LEXIS 15493 (N.D. Tex. Sept. 29, 1998), in which the magistrate judge had held that "DART is a political subdivision of the state of Texas, and is therefore immune from suit under the Eleventh Amendment." Id. at 1998 U.S. Dist. LEXIS 15493 at *24.

If the Anderson panel had published its opinion, it would have been binding on the panel in the instant case--Williams--and the result here would have been different. Based, however, on the mere fortuity that the Anderson panel decided not to publish, our panel in Williams was free to disagree with Anderson and to deny to DART the same immunity that Anderson had conferred on it less than two years earlier.

What is the hapless litigant or attorney, or for that matter a federal district judge or magistrate judge, to do? The reader should put himself or herself into the shoes of the attorney for DART. That client is told in May 1999, by a panel of this court in Anderson, that it is immune, on the basis of a "comprehensive and well-reasoned opinion." Competent counsel reasonably would have concluded, and advised his or her client, that it could count on Eleventh Amendment immunity.

Then, in March 2000, in the instant case, a federal district judge, understandably citing and relying on the circuit's decision in Anderson, holds that "[i]t is firmly established that DART is a governmental unit or instrumentality of the state of Texas." In February 2001, however, a panel, containing one of the judges who was on the Anderson panel, reverses and tells DART that, on the basis of well-established Fifth Circuit law from 1986, it has no such immunity. One can only wonder what competent counsel will advise the client now.

Fifth Circuit Rule 47.5.4 specifies that "[u]npublished opinions issued on or after January 1, 1996, are not precedent, except under the doctrine of res judicata, collateral estoppel or law of the case . . . ." This court's primary asserted justification for issuing unpublished opinions is efficiency.1 Indeed, efficiency may be the only justification for the practice of issuing unpublished opinions, but one that cannot be gainsaid. Danny J. Boggs & Brian P. Brooks, Unpublished Opinions & the Nature of Precedent, 4 Green Bag 2d 17, 19-21 (2000).

The justification for refusing to confer precedential status on such opinions is more tenuous. It maybe that, when this court promulgated rule 47.5 in 1995, the relative unavailability of unpublished opinions rendered their use as precedent fundamentally unfair. Today, however, that proposition is untenable: "Between Lexis and Westlaw, Internet sites maintained by universities and some of the circuit courts of appeals, and networks of attorneys practicing in particular fields, it is the rare opinion that is not disseminated for mass consumption." Boggs & Brooks, supra, at 18.2

The primary justification for refusing to grant unpublished opinions precedential weight--at least in this circuit--is, however, that an unpublished opinion will not "in any way interest persons other than the parties to [that particular] case," because the opinion neither establishes a new rule of law, modifies an existing rule of law, applies an existing rule to distinct facts, nor concerns any issue of significant public interest. 5th Cir. R. 47.5.1. Thus, at least theoretically, because an unpublished case does nothing new, an older case easily can be cited for the same proposition, rendering citation to the unpublished opinion unnecessary.

To the contrary, however, there are opinions that, though unpublished, do establish a new rule of law or apply existing law to distinct facts. In Christie v. United States, No. 91-2375MN (8th Cir. Mar. 20, 1992) (per curiam) (unpublished), the court, addressing the issue as res nova, held that the "mailbox rule" does not apply to refund claims against the Internal Revenue Service. In Anastasoff v. United States, 223 F.3d 898 (8th Cir.), vacated as moot, 235 F.3d 1054 (8th Cir. 2000) (en banc), another panel of the Eighth Circuit faced the same issue.

In Anastasoff, the taxpayer argued that Christie, the only Eighth Circuit decision on point, should be ignored because of its unpublished disposition. The court refused to overlook Christie, however, declaring that the circuit's local rule restricting the precedential value of unpublished opinions is unconstitutional. Id. at 899. As I will discuss, the court's holding on the constitutional question is at least arguable; Christie, moreover, is instructive as a textbook example of an unpublished opinion that in fact does announce a new rule of law. As I have explained, the same phenomenon occurred in the instant case.

Empirical evidence suggests that cases such as Christie and Anderson are more common than one might think. One study of unpublished opinions found "a surprising number of reversals, dissents, and concurrences. . . . suggest[ing] that panels authoring unpublished opinions reach some results with which other reasonable judges would disagree. . . . [F]ailing to give unpublished opinions precedential effect raises the very specter described by the Eight Circuit: that like cases will be decided in unlike ways . . . ." Deborah Jones Merritt & James J. Brudney, Stalking Secret Law: What Predicts Publication in the United States Courts of Appeals, 54 Vand. L. Rev. 71, 119 (2001) (citing Anastasoff, 223 F.3d at 901, 905).3 In the first half of 2001, this circuit has declined to publish at least four opinions in which a judge dissented,4 indicating that at least one of our number felt that each of those cases was not an easy application of existing law to indistinguishable facts.

For this reason, among others, the issuance of unpublished, non-precedential opinions recently has come under more intense scrutiny.5 In Anastasoff, Judge Richard Arnold, writing for the panel that declared the practice unconstitutional, questioned whether Article III's "judicial power" includes the power to disregard precedent, citing several historical bases for the conclusion that, when the Constitution was framed, "[t]he duty of courts to follow their prior decisions was understood to derive from the nature of the judicial power itself and to separate it from a dangerous union with the legislative power," Anastasoff, 223 F.3d at 930, despite the fact that "[b]efore the ratification of the Constitution, there was almost no private reporting and no official reporting at all in the American states," id. The panel therefore decided that Article III does not empower a court to ignore any of its prior cases, unpublished or not. Id.

Anastasoff has generated substantial controversy, and its historical research and conclusions have been criticized.6 Even if one disagrees that non-precedential opinions violate the Constitution, however, the issue is close enough for this court to give it en banc consideration.

In addition to the constitutional issue, the practice of refusing to recognize unpublished opinions as precedent raises important prudential questions. As I have said, one assumption on which the practice is based--that such opinions create no new law--is dubious. Moreover, in this particular case, one must question the fundamental fairness of having two opinions from the same court, one giving DART Eleventh Amendment immunity and the other, less than two years later, taking that immunity away.

The law is supposed to inform the choices of potential litigants. How can this circuit's decisions do so, if they carry no predictive effect?

Moreover, the justification for unpublished opinions--that, as a matter of efficiency, the court should not publish redundant opinions--provides no support for the proposition that such opinions should carry no precedential weight. If the opinion is a mere restatement of existing law (as it must be, if it is accorded unpublished status), what is the harm in viewing it as precedent?

There are powerful arguments both for and against the policy of giving precedential effect to unpublished opinions. Given the attention garnered by Anastasoff, both before and after that litigation was discontinued for mootness, and given the untoward inconsistency in result that I have pointed out from our circuit, this matter surely is important enough for the court to reconsider en banc.

Even without en banc consideration, I hope the court will reconsider the efficacy of rule 47.5.4 Because of its refusal to do so in the form of an en banc rehearing of the Williams case, I respectfully dissent from the denial of rehearing en banc.7