concurring.
Although I agree with the majority that appellant is entitled to a new trial, my reasons for reaching that conclusion are sufficiently narrower and distinct as to warrant a separate opinion. I also write separately to emphasize the unusual circumstances of this case and to address further the question of whether and when guilty plea negotiations with, or promises to, a cooperating co-defendant that fall short of a formal plea agreement must be disclosed.
*471In finding a violation of Brady v. Maryland, 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963), the majority’s primary focus is upon a letter sent by co-defendant James Alexander’s counsel to the District Attorney of Luzerne County, Robert J. Gillespie, Jr. The majority construes this letter as proof of a guilty plea “understanding” between the Commonwealth and Alexander. The majority derives further support for its belief that the letter proves a specific “understanding” from the fact that the sentence Alexander ultimately received was only four months longer than the sentence his counsel had proposed in the letter. Slip Op. at 10-11. In my view, the majority’s overemphasis on the letter is mistaken. The Commonwealth clearly had no obligation under Brady to disclose this letter, or the other letters the majority discusses, since the credited testimony below established that the plea negotiations to which they refer, occurring nearly a year before appellant’s trial, did not result in any actual plea agreement.
Instead, the due process duty to disclose here arose from the separate fact, unequivocally testified to by former District Attorney Gillespie at the PCRA hearing, that although no specific plea agreement was ever reached, Gillespie nevertheless had advised Alexander’s attorney that Alexander would receive consideration and be treated fairly if he cooperated (N.T. PCRA at 576-83). When Alexander squarely and repeatedly insisted at trial that no “promises,” “agreements,” or “arrangements” for “favorable treatment” had been made in connection with the prosecution pending against him (T.T. 1253, 1303-04), the Commonwealth was under a duty to disclose the promise of consideration and fair treatment made by Gillespie. Giglio v. United States, 405 U.S. 150, 153-55, 92 S.Ct. 763, 31 L.Ed.2d 104 (1972) (United States Attorney’s office obliged to disclose promise of non-prosecution in exchange for cooperation made by assistant United States attorney at grand jury stage; “evidence of any understanding or agreement as to a future prosecution” is relevant to credibility).1
*472Instead of doing so, the trial prosecutor, who apparently was unaware of the previous representation made by his superior, denied the existence of any arrangement (T.T. 1201, 1499). Under Giglio, it does not matter that the trial prosecutor was unaware of the prior promise made by his superior:
[Wjhether the nondisclosure was a result of negligence or design, it is the responsibility of the prosecutor. The prosecutor’s office is an entity and as such it is the spokesman for the Government. A promise made by one attorney must be attributed, for these purposes, to the Government.
Giglio, 405 U.S. at 154, 92 S.Ct. 763. The failure to disclose the District Attorney’s promise of consideration in the face of Alexander’s denial at trial of any arrangement implicates Brady/Giglio. The unconsummated guilty plea correspondence is of secondary importance at best; i.e., it is relevant only insofar as it suggests what it was that Alexander and his counsel might have understood the general promise of consideration/fair treatment to entail.
In its Brady analysis, the majority overstates the content and importance of the guilty plea correspondence, while it fails to fully appreciate the significance of the District Attorney’s testimony. Like appellant, Alexander was charged with first degree murder and kidnapping. Alexander, however, cooperated with authorities, beginning with his assisting police in locating the victim’s body; indeed, his cooperation predated any involvement of the District Attorney’s office. Before appellant was brought to trial, District Attorney Gillespie and Alexander’s counsel engaged in plea negotiations that were memorialized to some extent in a December 6, 1983, letter from Alexander’s counsel to Gillespie. The majority characterizes this letter as referring to “an earlier conversation with Gillespie regarding a deal for Alexander to receive a two year jail term upon a plea to the pending murder charges.” Majority op. 563 Pa. at 464, 761 A.2d at 1172. Later on, the majority goes so far as to state that the letter revealed that the Commonwealth “had offered Alexander a sentence of two years. on the charges of murder and kidnapping, pending *473information on his prior record.” Id. 563 Pa. at 467-68, 761 A.2d at 1174. These characterizations are inaccurate.
In fact, the December 6th letter did not specify whether it was referring to the murder charge or the kidnapping charge,2 nor did it make any reference to a “deal” or a sentencing “offer.” Instead, Alexander’s counsel stated in the letter that he had received a copy of Alexander’s prior record and then discussed how that record might affect a calculation of the sentencing guidelines — again, without specifying a particular offense. Counsel noted that, even if Alexander’s prior record score was “zero,” the mitigated minimum sentence under the guidelines would be thirty-six months. Counsel then stated that, “[previous to our discussion concerning the sentencing guidelines, you had indicated to me a sentence of two years.” Counsel further noted that he understood that the guidelines “could present a problem,” but felt that a sentence in excess of 36 months would be inappropriate in light of Alexander’s cooperation.
The majority’s characterization of the letter as referring to a “deal” “offered” by the Commonwealth for a two year sentence also ignores the PCRA testimony of Gillespie, who stated that the reference to “two years” could just have easily concerned a mere proposal by the defense, or “I could have said that the last time we had cooperation somebody received two years” (N.T. PCRA at 592). In addition, there is no evidence that Gillespie wrote back to Alexander’s counsel on the matter, although Gillespie did forward a copy of the letter to the state trooper spearheading the murder investigation, along with a notation that it was time to “firmly discuss our deal with Alexander.”
Conflicting evidence was presented at the PCRA hearing concerning whether these “negotiations” resulted in an actual plea agreement with Alexander before appellant’s trial ten *474months later. Alexander, who by the time of the PCRA hearing had served his entire sentence, testified, as co-felons so often do at that point, in favor of appellant. Alexander insisted that such a deal had been reached. Specifically, he claimed that he was promised that he would receive a “minimum sentence” in exchange for his testimony. Alexander further claimed that he was told by “somebody” in the district attorney’s office, whom he could not identify, to deny that the deal existed if he was asked about it at trial because “it could mess everything up.” Alexander claimed that he had deliberately perjured himself at trial when he stated that he was not testifying pursuant to an agreement (N.T. PCRA at 14-16, 29).
Alexander’s account that he testified pursuant to a plea arrangement and perjured himself at trial was squarely contradicted by every other PCRA witness who would have been in a position to know about such an arrangement — ie., his counsel (who would have to have actively suborned the perjury if Alexander was telling the truth), Gillespie, and the two assistant district attorneys who actually prosecuted appellant. That evidence established that there were plea negotiations, but no concrete plea agreement, and Alexander was never told to lie at trial. The testimonial conflict was resolved by the PCRA court, which had an opportunity to observe the demeanor of the relevant witnesses, in favor of the Commonwealth, with the court specifically discrediting Alexander’s testimony. PCRA slip op. at 20. That credibility finding, of course, is binding on this Court. Commonwealth v. Nelson, 484 Pa. 11, 14, 398 A.2d 636, 637 (1979) (“It is up to the trial court to judge the credibility of [recantation testimony]”); accord Commonwealth v. Henry, 550 Pa. 346, 706 A.2d 313 (1997).
The fact that no plea agreement was reached, however, does not end the Brady/Giglio inquiry, for the United States Supreme Court has recognized that promises or understandings as to the potential future prosecution of a co-defendant that fall short of a finalized plea bargain may also sufficiently implicate the witness’s credibility that due process requires disclosure. The constitutional underpinning for the require*475ment that such exculpatory evidence be disclosed was traced by the Giglio Court as follows:
As long ago as Mooney v. Holohan, 294 U.S. 103, 112, 55 S.Ct. 340, 342, 79 L.Ed. 791(1935), this Court made clear that deliberate deception of a court and jurors by the presentation of known false evidence is incompatible with ‘rudimentary demands of justice.’ This was reaffirmed in Pyle v. Kansas, 317 U.S. 213, 63 S.Ct. 177, 87 L.Ed. 214 (1942). In Napue v. Illinois, 360 U.S. 264, 79 S.Ct. 1173, 3 L.Ed.2d 1217 (1959), we said, ‘[t]he same result obtains when the State, although not soliciting false evidence, allows it to go uncorrected when it appears.’ Id., at 269, 79 S.Ct., at 1177. Thereafter Brady v. Maryland, 373 U.S. [83,] 87, 83 S.Ct. [1194,] 1197, 10 L.Ed.2d 215 [(1963)], held that suppression of material evidence justifies a new trial ‘irrespective of the good faith or bad faith of the prosecution.’ ... When the ‘reliability of a given witness may well be determinative of guilt or innocence,’ nondisclosure of evidence affecting credibility falls within this general rule. Napue, supra, at 269, 79 S.Ct., at 1177.
Giglio, 405 U.S. at 153-54, 92 S.Ct. 763. Indeed, since Giglio, some courts have recognized that an unconsummated agreement can create a greater incentive for a witness to testify in a manner that he perceives to be favorable to the government. E.g. State of Louisiana v. Lindsey, 621 So.2d 618 (La.Ct.App. 1993) (promise of favorable consideration in exchange for testimony deemed credible gave witness “a direct personal stake” in defendant’s conviction; “The fact that a specific reward was not guaranteed through a promise or a consummated plea agreement, but was expressly contingent on the state’s good faith and satisfaction with [the witness’s] testimony, served only to strengthen any incentive to testify falsely in order to secure [the defendant’s] conviction.”). Cf. United States v. Bagley, 473 U.S. 667, 683, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985) (plurality opinion by Blackmun, J.) (“The fact that the stake was not guaranteed through a promise or binding contract, but was expressly contingent on the Government’s *476satisfaction with the end result, served only to strengthen any incentive to testify falsely in order to secure a conviction”).3
Here, the plea negotiations with Alexander did not break down or terminate.4 Rather, they were deliberately left open in part because District Attorney Gillespie, as a matter of policy, sought to avoid plea agreements with cooperating co-defendants until after the cooperation had occurred and could be assessed for truthfulness. There is nothing wrong with such a policy which, as Gillespie testified, he deemed a practical necessity since, “you always have the possibility of getting burned into a plea agreement and then having the defendant not cooperate” (N.T. PCRA at 577). An additional legitimate concern weighing in favor of not committing to plea agreements before cooperation was the recognition that such agreements were relevant to bias/credibility and thus could be used to undermine the witness’s testimony (id. at 578, 589-90).
Gillespie further testified, however, and it this testimony that implicates Brady/Giglio, that it would have been his normal practice to indicate to Alexander’s counsel that Alexander’s truthful cooperation “would get consideration” and that, in point of fact, “I can tell you that I am sure we told [Alexander’s counsel] that his client would be treated fairly if he cooperated” (id. at 577, 582). Gillespie elaborated that what fair treatment would entail was a “case by case” matter, ie., the Commonwealth would “look at the circumstances of, one, his cooperation, what it meant to the trial of the case; two, look at his participation in the crime; and three, then look and see what justice demanded. And I can tell you that takes many forms” (id. at 576, 582-83). Finally, Gillespie noted that his county was comparatively small, that he had a relationship of “mutual respect and trust” with most defense counsel, including Alexander’s lawyer, and thus he expected *477that Alexander’s counsel would accept that the Commonwealth would honor its word (id. at 573-74, 577, 579-80).
Gillespie’s promise to Alexander’s counsel, rather than the plea negotiation letter, implicates due process. It was that promise that had to be disclosed once Alexander flatly denied at trial that any “promises,” “agreements,” or “arrangements” for “favorable treatment” had been made in connection with his open case. The plea negotiation letter, though important, was derivative. It suggests what Alexander had reason to believe would be the “consideration” he would receive for his cooperation. Alexander, like appellant, was facing charges of first degree murder; if convicted, he faced a sentence of life imprisonment at best. The fact that the Commonwealth was willing to discuss a plea agreement that would result in a significantly lesser term of years and remove the prospect of life imprisonment was significant in and of itself, irrespective of the precise sentence Alexander ultimately would receive. Alexander’s credibility and potential interest and bias were a central issue at trial. Particularly in light of the nature of the negotiations that preceded it, Gillespie’s promise provided Alexander with a powerful incentive to testify favorably for the Commonwealth and unquestionably was a matter that was relevant to attack Alexander’s credibility. Such impeachment evidence, no less than exculpatory evidence, is covered by the Brady rule. See Bagley, 473 U.S. at 676, 105 S.Ct. 3375.
The remaining question is whether the undisclosed promise was material for Brady purposes and whether there is a “reasonable probability” that the outcome of the trial would have been different if Alexander had been subject to cross-examination on the point. Id. at 677, 105 S.Ct. 3375; see also id. at 685, 105 S.Ct. 3375 (White, J., concurring).5 As to this point, I agree with the majority’s analysis. Alexander was not just a key witness against appellant, but an indispensable one. This trial, moreover, squarely pitted his credibility against that of appellant, who also testified. Cross-examination of a *478cooperating co-defendant like Alexander, whose testimony is obviously motivated to some degree by self-interest, can be effective even where the witness has received no “deal” or “promise” in exchange for his testimony. Here, however, there was an existing promise, although a vague promise, and Alexander’s true status was not disclosed to appellant. Moreover, appellant’s efforts to impeach Alexander on grounds of expected favorable treatment from the Commonwealth were stymied when that claim was met by a denial not only from the witness but also from the trial prosecutor, an officer of the court. The jury was thus erroneously led to believe that Alexander had no personal stake in the trial’s outcome beyond a vague hope for leniency. Suggesting to the jury that a witness’s testimony was influenced by a unilateral hope for mercy is far less effective than concrete evidence that the witness was given an actual assurance of fair treatment and consideration in a case where any such consideration meant the prospect of avoiding a life sentence. Given the importance of Alexander’s testimony, and the strength of the undisclosed impeachment, I agree that a new trial is warranted.
Not all negotiations between the prosecution and a cooperating co-defendant witness falling short of an actual agreement must be disclosed to the defense. It is only where some actual promise of favorable treatment in the witness’s own prosecution has been made, and that fact becomes material at trial, that disclosure is required. Because I am satisfied that those contingencies exist here, I concur in the mandate of a new trial.
Justice NEWMAN joins this concurring opinion.. Although Giglio involved a federal prosecution, it was based upon the due process clause and the Brady line of cases and thus is binding on state prosecutions.
. Since the letter spoke of a sentencing range short of life imprisonment, if it was referring to the murder charge, it had to be contemplating either third degree murder or a manslaughter charge. Notably, since both third degree murder and kidnapping are graded as felonies of the first degree, 18 Pa.C.S. §§ 2502(c), 2901(b), they would result in the same sentencing guideline ranges.
. This portion of Justice Blackmun’s opinion in Bagley was joined only by Justice O'Connor.
. Alexander’s counsel testified that, if plea negotiations had broken off, he would have advised Alexander to cease cooperating with the Commonwealth (N.T. PCRA at 727). Alexander, of course, did continue to cooperate and testified against Appellant.'
. Bagley was a majority opinion as to the propositions cited in the text above but was a plurality opinion as to its application of ihe Brady materiality standard to the facts of that case.