dissenting.
Under our system of law, a prosecutor has awesome power because of his largely unreviewable discretionary authority to bring charges. See Bordenkircher v. Hayes, 434 U.S. 357, 364, 98 S.Ct. 663, 54 L.Ed.2d 604 (1978); United States v. Cowan, 524 F.2d 504 (5th Cir. 1975), cert. denied sub nom. Woodruff v. United States, 425 U.S. 971, 96 S.Ct. 2168, 48 L.Ed.2d 795 (1976). For all the theoretical power we possess as federal judges, it is only in an unusual case that we can step in and curb a dubious prosecution. See Oyler v. Boles, 368 U.S. 448, 82 S.Ct. 501, 7 L.Ed.2d 446 (1962); United States v. Leggett and Platt, Inc., 542 F.2d 655, 658 (6th Cir. 1976), cert. denied, 430 U.S. 945, 97 S.Ct. 1579, 51 L.Ed.2d 792 (1977). As the Supreme Court noted in United States v. Dotterweich, 320 U.S. 277, 285, 64 S.Ct. 134, 138, 88 L.Ed. 48 (1943), “[o]ur system of criminal justice necessarily depends on conscience and circumspection in prosecuting officers . . ”
One of the few controls on prosecutorial discretion is the rule enunciated in Blackledge v. Perry, 417 U.S. 21, 94 S.Ct. 2098, 40 L.Ed.2d 628 (1974). The defendant in Blackledge was convicted on a misdemeanor assault charge in state court. When he exercised his statutory right of appeal to a trial de novo, the prosecutor obtained a felony assault indictment covering the same conduct. The Court held that since the prosecutor had “a considerable stake” in deterring such appeals, his action raised a spectre of prosecutorial vindictiveness *248which could not be squared with due process. The state was barred from bringing the heavier charge.
Blackledge’s proscription of prosecutorial vindictiveness followed the rule laid down in North Carolina v. Pearce, 395 U.S. 711, 89 S.Ct. 2072, 23 L.Ed.2d 656 (1969) relating to judicial vindictiveness. If a convicted defendant successfully won a new trial on appeal, it was impermissible for the trial judge on retrial to resentence him to a stiffer term absent on-the-record findings based on previously unknown facts.1
These authorities provide a sound rationale for a court to police prosecutorial abuse.2 Today, my colleagues interpret *249narrowly the principles in those decisions and severely limit a needed and vital control on prosecutorial discretion. Because I cannot agree with my colleagues’ abject capitulation to potentially abusive use of prosecutorial power, I respectfully dissent.
I Prosecutorial Vindictiveness — The Problem
Pearce and Blackledge stand for the general proposition that vindictive conduct by persons with the awesome power of judges and prosecutors is unacceptable in our jurisprudential system. However, there is no need to show actual vindictiveness,3 the appearance of vindictiveness is prima facie sufficient. As the Court noted in Blackledge, supra, 417 U.S. at 28, 94 S.Ct. at 2102:
The rationale of our judgment in the Pearce case, however, was not grounded upon the proposition that actual retaliatory motivation must inevitably exist. Rather, we emphasized that “since the fear of such vindictiveness may unconstitutionally deter a defendant’s exercise of the right to appeal or collaterally attack his first conviction, due process also requires that a defendant be freed of apprehension of such a retaliatory motivation on the part of the sentencing judge.” 395 U.S., at 725 [, 89 S.Ct. 2072], We think it clear that the same considerations apply here. A person convicted of an offense is entitled to pursue his statutory right to a trial de novo, without apprehension that the State will retaliate by substituting a more serious charge for the original one, thus subjecting him to a significantly increased potential period of incarceration.
The concepts which underlie the Pearce and Blackledge decisions, then, are 1) forestalling actually vindictive behavior, because it is morally wrong4 and 2) ensuring that the appearance of vindictive behavior does not deter criminal defendants from exercising their rights. Applying these concepts in varying situations, the Supreme Court has determined what “prophylactic rule” 5 if any, need be applied.
In Pearce, supra, the Court imposed a two-fold requirement on a judge to justify *250a seemingly vindictive resentencing. First, there must be an on-the-record showing why the judge increased the sentence the second time around. Second, the reasons for the increased sentence “must be based upon objective information concerning identifiable conduct on the part of the defendant occurring after the time of the original sentencing proceeding.” Id., 395 U.S. at 726, 89 S.Ct. at 2081.6
The problem is that in Blackledge, the Court did not indicate what formal standard was to be applied to test the seemingly vindictive conduct of a prosecutor. Analogizing to Pearce, defendants urge that in a case such as this, the conduct be deemed to have been vindictive unless there are new, after-the-fact reasons on the record why the prosecutor “upped the ante.” The government’s view is that Blackledge should be applied gingerly (if at all) in pre-trial situations and that it should be able to advance reasons such as inadvertence and neglect to justify its conduct. The majority opinions more or less adopt the government’s position.
The other courts of appeals which have considered this issue have advanced varying standards. The Ninth Circuit has ruled that where a prosecutor increases the charges in apparent response to the defendant’s exercise of a procedural right, the prosecutor bears a “heavy burden” of dispelling the resultant presumption of vindictiveness. It is vital to dispel the appearance of vindictiveness as well as actual vindictiveness. See United States v. Groves, 571 F.2d 450 (9th Cir. 1978) (presumption arose when prosecutor brought added charges in response to defendant’s motion to dismiss on statutory speedy trial grounds); United States v. Alvarado-Sandoval, 557 F.2d 645 (9th Cir. 1977) (presumption arose when defendant refused to plead to misdemeanor charge and indicated that pre-trial motions would be raised. “Immaterial that, due to a failure of communication within the office, the assistant U.S. Attorney who initially appeared was not personally aware of [defendant’s prior] record.”); United States v. DeMarco, 550 F.2d 1224 (9th Cir.) cert. denied, 434 U.S. 827, 98 S.Ct. 105, 54 L.Ed.2d 85 (1977) (presumption arose when defendant insisted on his statutory venue rights); United States v. Ruesga-Martinez, 534 F.2d 1367 (9th Cir. 1976) (“heavy burden” of offsetting presumptive prosecutorial vindictiveness not offset by showing of prosecutor’s inexperience).
The Fifth Circuit has ruled in one case that a prima-facie presumption of vindictiveness arises when a prosecutor apparently acts to chill the exercise of a right which can be dispelled if the prosecutor can give non-vindictive reasons for his conduct. Hardwick v. Doolittle, 558 F.2d 292 (5th Cir. 1977) , cert. denied, 434 U.S. 1049, 98 S.Ct. 897, 54 L.Ed.2d 801 (1978). Hardwick's concern over the existence of actual vindictiveness should be contrasted with the balancing test advanced in Jackson v. Walker, 585 F.2d 139 (5th Cir. 1978) where the court stated that: “In deciding whether to require a showing of actual vindictiveness or merely a showing of reasonable apprehension of vindictiveness, a court must weigh the extent to which allowing the second indictment will chill the exercise of the defendant’s appeal rights against the extent to which forbidding the second indictment will infringe on the exercise of the prosecutor’s independent discretion.” Id. at 145. See also United States v. Thomas, 593 F.2d 615, 624 (5th Cir. 1979); Miracle v. Estelle, 592 F.2d 1269 (5th Cir. 1979). The other courts of appeals have not specifically articulated the requirements for dispelling the appearance of prosecutorial vindictiveness but have dealt with the issue on an ad hoc basis.7
*251Questions of prosecutorial vindictiveness are not easy to resolve. Some Courts have emphasized the need to deter actual retaliatory vindictiveness, e. g. Hardwick v. Doolittle, supra; while other courts have emphasized the need to avoid the appearance of vindictiveness which deters others from exercising their rights, e. g. United States v. Ruesga-Martinez, supra. The first view emphasizes the moral consideration of preventing a prosecutor from retaliating because a troublesome defendant insists on his rights. The second view emphasizes the necessity of preventing other defendants from being apprehensive about exercising statutory or constitutional rights.
There is universal agreement in the courts that a finding of actual vindictiveness warrants immediate imposition of Blackledge’s proscriptions.8 The disagreement is over dealing with situations where there is an appearance of vindictiveness from a defendant’s viewpoint, but where this appearance is not strong9 or where the prosecutor simply made an honest mistake. In those cases, there exists the strong countervailing policy of giving leeway to prose-cutorial discretion and letting a prosecutor bring charges where there exists probable cause that the defendant violated the law.
An additional complicating factor is the difficulty of showing actual vindictiveness, which involves a determination of the prosecutor’s state of mind. Rare is the prosecutor who will openly admit that he added on charges in retaliation against the defendant. Sometimes it is clear that a retaliatory motive is the only basis for a prosecutor’s actions,10 just as it is sometimes clear that there does not exist even an appearance of *252vindictiveness when a prosecutor added charges.11 The typical case,' however, is in between these extremes. A prosecutor’s actions will appear vindictive, but the prosecutor will claim inadvertence or mistake. To reject these explanations in every case is to intrude deeply into prosecutorial discretion. Yet, ready acceptance of glib prosecutorial explanations for the seeking of heavier charges could make a mockery of Blackledge, which focused on the need to thwart the appearance of vindictiveness and the concomitant chill on other defendants’ exercise of their rights. As I will explain below, this is precisely what is wrong with my colleagues’ views.
II Prosecutorial Vindictiveness — A Suggested Approach
I would adopt an overall balancing test somewhat similar to the balancing test advanced in Jackson v. Walker, 585 U.S. 139 (5th Cir. 1978) as providing the best resolution of these difficulties. A court should first decide as a threshold matter whether a prosecutor’s action in seeking a heavier second indictment appears to be vindictive. If so, the court should examine the facts and weigh the extent to which allowing the second indictment will chill defendant’s exercise of the right in question with the extent to which forbidding the second indictment infringes on the prosecutor’s charging authority. If the balance falls in favor of the defendant, then the government would have the heavy burden of offsetting the appearance of vindictiveness. If the balance favors the government then there would arise a prima-facie case of vindictiveness, but all the government need do is provide neutral explanations to demonstrate that it did not, in fact, act vindictively-
The balancing approach makes a great deal of sense when one realizes that there are degrees of apparent vindictiveness which have varying effects on defendants and their propensity to exercise their rights. Thus, if a prosecutor seeks to augment the charges when a defendant exercised his right to a trial de novo, as occurred in Blackledge, the appearance of vindictiveness is so great that almost no explanation will justify it.12 On the other hand, if a prosecutor seeks a superseding indictment on retrial which substitutes a lesser charge which happens to carry a greater potential prison term, an appearance of vindictiveness certainly exists, but defendants’ apprehension of vindictiveness will not be great.13 See Jackson v. Walker, supra.
The balancing approach also takes into account varying degrees of prosecutorial in*253terest. For example, a prosecutor should have greater latitude to augment charges when dealing with multiple criminal acts or a crime spree, see Jackson v. Walker, supra 585 F.2d at 145-46; Hardwick v. Doolittle, supra 558 F.2d at 301, than when he is dealing with a single criminal act. See Miracle v. Estelle, supra 592 F.2d at 1276. Similarly, a prosecutor should have broader leeway to add charges before an initial trial than in a case where a defendant is to be tried a second time. See United States v. Thomas, supra 593 F.2d at 624. In each case, there would be a factual determination to be made, necessarily on an ad hoc basis.14
Ill The Instant Case
In this case, the district court correctly made the threshold determination that the government’s seeking of an additional charge after the defendants “won” a hotly contested bail motion implicated analysis under Blackledge. The government sought to show that the reason it acted the way it did was that the Assistant U.S. Attorney was inexperienced and that the conspiracy count would have been added earlier but for factors beyond her control — an office moratorium on grand jury indictments because of legal challenges to grand jury composition, a grand jury recess during the new year’s holidays and preemption of her grand jury time by a Senior Assistant U.S. Attorney. Relying on United States v. De Marco, supra and United States v. Ruesga-Mar-tinez, supra, the district court concluded that these explanations were facially insufficient to meet the government’s “heavy burden” of dispelling the appearance of vindictiveness. Applying the overall balancing test I advance in this opinion, I agree with the district court.15
Applying the balancing test advanced in this dissent, I think it clear that on these facts, the chilling effect on defendant’s freedom to seek release on bail outweighed the need to give prosecutors freedom to decide whom to prosecute. The critical factor in this case was that the government strongly opposed defendants’ release on bail. Despite the fact that defendants were arrested in November of 1975 and indicted in November of 1976, it was only after the defendants successfully appealed their bail determination from the magistrate to the district court in January of 1977 that the prosecutor added the conspiracy count. In addition, no crime spree was involved; the government added extra charge encompassing the same underlying conduct. These circumstances create a strong appearance of vindictiveness strikingly similar to that in Blackledge where the defendant appealed to a trial de novo. Concomitantly, the inevitable effect of this is to create a strong apprehension of vindictiveness among other defendants in the same situation.
I recognize that the government’s interest in having broad charging discretion in a pretrial situation such as this one is substantial. This case presents an instance where application of the balancing test reveals both a strong defendant’s interest and a strong prosecutor’s interest.16 On these *254facts, I think that the balance should be struck in favor of the defendant; the appearance of vindictiveness was simply too great. Accordingly, the test which the government should meet is dispelling the appearance of vindictiveness and not the lesser standard of offsetting an inference of actual vindictiveness.17
I agree with the district court that the government’s proffered explanations, are insufficient to dispel the apprehension of vindictiveness which resulted here.18 I do not think that defendants should have to worry that successful appeal of a bail motion will result in their facing additional charges.
IV The Views of My Colleagues
My colleagues have written two separate opinions in this case. Their views on the prosecutorial vindictiveness issue are quite different from mine. Unfortunately for the law of this circuit, each of their opinions cannot be reconciled with the other.
A fundamental premise underlying both of my colleagues’ opinions is that Borden-kircher v. Hayes, 434 U.S. 357, 98 S.Ct. 663, 54 L.Ed.2d 604 (1978) significantly limits Blackledge v. Perry. Indeed, Judge Merritt takes the position that Bordenkircher effectively overrules Blackledge. I cannot agree. Both of my colleagues fall into the trap of relying on hyperbole in the dissent in Bordenkircher. The danger of relying on Supreme Court dissents should be obvious. I would look to the language in the majority opinion in Bordenkircher which emphasizes the limited nature of the Court’s holding. The Court spoke at length of plea bargaining and the idea that “there is no such element of punishment or retaliation so long as the accused is free to accept or reject the prosecution’s offer.” 434 U.S. at 363, 98 S.Ct. at 668. Indeed, the Court stated that “We hold only that the course of conduct engaged in by the prosecutor in this case which no more than openly presented the defendant with the unpleasant alternatives of forgoing trial or facing charges on which he was plainly subject to prosecution, did not violate the Due Process Clause of the Fourteenth Amendment.” 434 U.S. at 365, 98 S.Ct. at 669.
Bordenkircher presents a case where the prosecutor acted vindictively, but where an independent factor — the defendant’s ability to choose — eliminated any due process problems. The Court was careful to contrast the give and take of plea bargaining with the “unilateral imposition of a penalty upon a defendant who had chosen to exercise a legal right.” 434 U.S. at 362, 98 S.Ct. at 667. It is unquestionably true that there exists tension between Bordenkircher and Blackledge. However, I would take Justice Stewart at his word and limit Bordenkircher to its facts pending further guidance from the Supreme Court.
Although Judge Merritt seems to think that I have misperceived Bordenkircher, I think that the opposite is true. No other court has advanced the view that Bordenkircher limits Blackledge. Some courts have not even bothered to cite Bordenkircher when discussing prosecutorial vindictiveness issues outside of the plea bargaining context, e. g., Jackson v. Walker, supra. Those courts which have cited Bordenkircher have advanced the view which I hold. *255See United States v. Jones, 587 F.2d 802, 805 n. 2 (5th Cir. 1978); United States v. Groves, 571 F.2d 450, 455 (9th Cir. 1978). See also Watkins v. Solem, 571 F.2d 435, 436 (8th Cir. 1978) (Bordenkircher holding “very narrow”). See generally Smaltz, Due Process Limitations on Prosecutorial Discretion in Re-Charging Defendants : Pearce to Blackledge to Bordenkircher, 36 Wash & Lee L.Rev. 347 (1979).
Instead of limiting Bordenkircher to its facts, as I would, my colleagues have seized upon that decision to limit Blackledge and Pearce to their facts. I think that my colleagues views are unsound, both as a matter of precedent and as a matter of public policy. Because their opinions are so different, I will discuss them separately.
A.
Although Judge Green purports to adopt a balancing test similar to the one I advance or the one Judge Goldberg advanced in Jackson v. Walker, supra; in reality he does no such thing. Judge Green breaks down vindictiveness issues into three factual situations: 1) substitution of charges, 2) addition of charges in a crime spree situation, 3) addition of charges for the same basic criminal acts (the situation here). He would bar charges in situation # l.19 Regarding situations # 2 or # 3, however, he would mechanically apply a prima-facie vindictiveness test, and allow the government to rebut with glib explanations such as inadvertence or mistake in all cases.20
The problem with this approach is that it ignores Blackledge’s underlying concern with the appearance of prosecutorial vindictiveness and concomitant chilling effect on defendants.21 It simply gives too much deference to prosecutorial discretion in most situations and effectively limits Blackledge to its narrow facts.
Instead of making Judge Green’s artificial distinctions, I would recognize the pros-*256ecutorial interest in situations # # 2 and 3 and weigh it against the defendant’s interest in different fact settings. That is a true balancing test and gives proper weight to both defendants’ and prosecutors’ interests.
The facts of this case illustrate how tenuous Judge Green’s distinctions really are. Because the prosecutor only added a conspiracy count, he allows the government to advance explanations such as inadvertence and neglect to rebut a prima facie case of vindictiveness. Had the prosecutor substituted a more severe charge, Judge Green would bar it. I do not think that this per se distinction is warranted. Instead, I would look to all of the circumstances. As I have previously indicated, what made the prosecutor’s conduct so apparently vindictive here is the timing of the addition of charges, immediately after the government lost a hard-fought bail motion. It is the high chilling effect of this conduct which should result in a barring of the charge. In this case, whether the government substituted or added the charge is not very important. Yet, under Judge Green’s view, it is dispositive.22
B.
Judge Merritt’s opinion is certainly unique. His position is that no matter why a prosecutor brings added charges, there is no problem of vindictiveness unless a second trial is implicated. Assume that a prosecutor openly threatens a defendant, in advance, with more charges if the defendant brings a motion for bail or exercises some other procedural right not involving a second trial. To Judge Merritt, why, this is nothing more than good old “regular prose-cutorial hard nosedness.”
Not only does Judge Merritt benignly, approve “regular prosecutorial hard nosedness,” he lives in deathly fear that to interfere with a prosecutor’s decisionmaking is “unmanageable” and “unworkable.” Judge Merritt’s opinion accords almost mystical deference to prosecutorial discretion, yet advances no policy arguments to support this position. In my view, there are compelling reasons to reject it.
The only reason why prosecutorial vindictiveness problems exist in the first place is that the prosecutor’s office bungled things. There is no reason why the defendants could not have been charged with all offenses from the very beginning.23 Where a prosecutor makes a reasoned decision to bring certain charges in the beginning, there never exists any question of vindictiveness. It is only when a prosecutor starts adding charges later on that could have been brought earlier that vindictiveness questions arise. Since vindictiveness issues come up only if a prosecutor does not do his job right in the first place, I see no reason not to carefully examine them. To take the opposite approach (that of Judge Merritt) is to abdicate completely to prose-cutorial discretion. I do not think that our system of criminal justice should tolerate the prosecutorial abuse which Judge Merritt’s views would facilitate. A person may be prosecuted for many reasons, but he should never have to face added charges because he exercised his rights.
*257Judge Merritt properly notes the serious problems which vindictiveness issues present for a prosecutor. Prosecutors are necessarily opposed to criminal defendants, and prosecutors make honest mistakes. Unlike Judge Merritt, I do not think that these problems are insurmountable. Nor do I think that we should shirk from our duty to police vindictiveness and apply clear' Supreme Court precedent.
I recognize the difficulties in dealing with actual fact situations such as this one.24 I believe that a balancing test provides a good resolution of them. The irony is that the balancing test I advance recognizes the strong prosecutorial interests which underlie Judge Merritt’s views. The difference is that it does not abdicate to these interests. The peculiar facts of this case are unique in that it is a rare instance where the balancing test results in the barring of additional charges brought before trial. In most such situations, the prosecutorial interest will be so strong and the appearance of vindictiveness so slight that the prosecutor would not be barred from bringing additional charges.
Judge Merritt would look to the double jeopardy clause for guidance and would extend Bordenkircher to all pre-trial situations.25 The problem with this idea is that there exists no judicial support anywhere that I have seen for it. Indeed, both Black-ledge and Bordenkircher specifically rely on the due process clause alone and do not speak of double jeopardy “values.” I am of the view that prosecutorial vindictiveness encompasses far more than situations involving a second trial.
V Conclusion
In conclusion, I would accord far greater scrutiny to prosecutorial vindictiveness questions than either of my colleagues. Judge Green’s opinion is not unreasonable but allows prosecutors too much leeway. Judge Merritt’s opinion is a simplistic overreaction to a complex problem. It is unfortunate that he reaches the ostrich-like conclusion that because prosecutorial vindictiveness is difficult to police, it should be ignored whenever possible.
While each of the opinions in this case will doubtless be read with interest by prosecutors and defense counsel alike, neither they nor the district judges of the circuit will have definitive guidance. Given the chaotic nature of the law in the circuits regarding prosecutorial vindictiveness, we can expect further guidance on this question from the Supreme Court in the near future. It is regrettable that, until then, my colleagues have passed up an opportunity to impose needed controls on potentially abusive prosecutorial conduct.
. See also Colten v. Kentucky, 407 U.S. 104, 92 S.Ct. 1953, 32 L.Ed.2d 584 (1972) (where increased sentence was imposed by different judge after appeal to a trial de novo, possibility of vindictiveness was minimal); Chaffin v. Stynchcombe, 412 U.S. 17, 93 S.Ct. 1977, 36 L.Ed.2d 714 (1972) (minimal possibility of vindictiveness where different jury imposed stiffer sentence on retrial).
There exists some tension between Pearce, Colten and Chaffin, as is inevitable when a court formulates rules to try to control subjective behavior. See Hardwick v. Doolittle, 558 F.2d 292, 299 (5th Cir. 1977), cert. denied, 434 U.S. 1049, 98 S.Ct. 897, 54 L.Ed.2d 801 (1978). Nonetheless, “[t]he lesson that emerges from Pearce, Colten, and Chaffin is that the Due Process Clause is not offended by all possibilities of increased punishment upon retrial after appeal, but only by those that pose a realistic likelihood of vindictiveness.” Blackledge v. Perry, 417 U.S. 21, 27, 94 S.Ct. 2098, 2102, 40 L.Ed.2d 628 (1974).
. Neither of my colleagues addresses the government’s threshold arguments that Blackledge should be completely inapplicable here. Judge Green’s opinion implicitly rejects these claims since he does apply a form of Black-ledge analysis. I agree that the government’s arguments are frivolous and that this case must be analyzed under Blackledge.
The government claims that Blackledge is inapplicable on the facts of this case and advances three arguments in support. First, the government asserts that what occurred here was so trivial that there existed no real possibility of vindictiveness. The government correctly states that Blackledge only applies where the prosecutor has a “considerable stake” in deterring the exercise of an established right. The government contends that it does not have such a considerable stake in chilling a defendant’s right to seek admittance to bail. It bases this argument on the notion that a bail hearing does not involve a burdensome expenditure of prosecutorial resources, nor the potential release of a convicted felon. The government would have us restrict the Blackledge rule to cases where the right exercised by the defendant carries the potential for retrial.
The government’s argument that prosecutors are not much concerned whether defendants are released on bail can only be described as surprising to anyone familiar with our criminal justice system. An eager-to-convict prosecutor has every reason to favor keeping a defendant in jail pending trial. A jailed defendant has great difficulty cooperating with his attorney in building a defense. Incarcerated defendants are under more pressure to plea-bargain. Further, in the eyes of a maliciously vindictive prosecutor who is bent on punishing a defendant, time spent in jail awaiting trial offers great satisfaction, even if the defendant is ultimately acquitted.
The facts of this case demonstrate a prosecutor’s considerable stake in the granting or denial of bail to criminal defendants. After the arrest of Thurston Brooks, Tallice Andrews and Fannie Braswell, Fannie Braswell turned state’s evidence and offered to testify against her two co-defendants. A confidential informant told the assistant U.S. Attorney that the defendants had threatened Ms. Braswell with bodily harm. In response the government placed Ms. Bras-well in the custody of the Federal Witness protection program and strenuously argued to both the Magistrate and the district court that the defendants should not be released on bail. It is difficult to imagine a case in which the prosecutor’s stake in deterring the exercise of a right is more clearly presented.
The government’s attempt to limit Blackledge to situations involving a full retrial has no basis in logic. While it is true that a prosecutor always has a considerable stake in avoiding the cost and inconvenience of a trial, a prosecutor often has a similar motivation to deter the exercise of Constitutional or statutory rights where something other than a full retrial is involved. To the extent that the exercise of a right in some way decreases the likelihood of conviction or makes a prosecutor’s job more burdensome, the prosecutor has an incentive to deter the exercise of that right.
I recognize that there may exist situations where a court could find that a prosecutor was utterly uninterested whether a defendant was released on bail. However, there would be a clear record of this in that the government would not have opposed bail in the first place.
The government’s second argument is that it did not substitute a more severe charge, as occurred in Blackledge; it merely added the extra conspiracy charge. I do not see why this distinction should make any difference. As *249this court noted in United States v. Sturgill, 563 F.2d 307 (6th Cir. 1977), the important question is whether a defendant is subject to the possibility of enhanced punishment. This is what motivates the vindictive prosecutor and chills the defendant’s exercise of his rights. See Miracle v. Estelle, 592 F.2d 1269, 1275-76 (5th Cir. 1979); Jackson v. Walker, 585 F.2d 139, 146 (5th Cir. 1978); United States v. Jamison, 164 U.S.App.D.C. 300, 308, 505 F.2d 407, 415 (D.C. Cir. 1974).
The government thirdly argues that the danger of prosecutorial vindictiveness is minimized here by the presence of an independent grand jury. This argument misperceives Blackledge. All that the grand jury did here was determine whether there was probable cause for the charges. The grand jury did not address why charges were brought or whether the bringing of charges would deter defendants from exercising their rights. Further, the notion that a grand jury operates as a significant check on a prosecutor is rather remarkable. Although the “runaway” grand jury is not unknown in our history, anyone with any knowledge of our criminal justice system knows that grand juries routinely return whatever indictment the prosecutor requests. Given the fact that the prosecutor orchestrates the grand jury proceedings and that no opposing attorneys are allowed to participate in them, any other result would be surprising. I note that in Blackledge itself, supra, 417 U.S. at 23, 94 S.Ct. 2098, the prosecutor also sought — and received — a grand jury indictment for a greater offense when the defendant exercised his right to a new trial. Thus, Blackledge implicitly rejects the government’s argument as well.
. By “actual vindictiveness” I do not mean that the prosecutor is motivated by malice. Rather, I mean that the prosecutor is motivated by a desire to retaliate against a pesky defendant who is making the former’s job more difficult for him/her.
In the case at bar, the district court did not question the prosecutor’s integrity or motivation. 1 have no reason to do so. On the contrary, I have high regard for this prosecutor and the U.S. Attorney’s Office for which she works.
. See North Carolina v. Pearce, 395 U.S. 711, 723-24, 89 S.Ct. 2072, 23 L.Ed.2d 656 (1969). Compare id. at 738-39, 89 S.Ct. 2072 (Opinion of Black, J.). Although this view is based on Aristotelian and Kantian notions of “fairness,” one might add that it also serves to uphold the legitimacy of the system in the eyes of the public and thus can be justified under utilitarian analysis as well.
. Colten v. Kentucky, 407 U.S. 104, 116, 92 S.Ct. 1953, 32 L.Ed.2d 584 (1972).
. Compare Justice White’s view that would “authorize an increased sentence on retrial based on any objective, identifiable factual data not known to the trial judge at the time of the original sentencing proceeding.” North Carolina v. Pearce, supra at 751, 89 S.Ct. at 2089.
. See United States v. Johnson, 537 F.2d 1170 (4th Cir. 1976) (Defendant successfully overturned guilty plea on appeal. Court of Appeals accepted prosecutor’s explanation that increased charges on remand for trial were added simply to create a stronger case for the govern*251ment, but held that the defendant’s apprehension of vindictiveness was sufficient to bar them); United States v. Ricard, 563 F.2d 45 (2d Cir. 1977) (apprehension of prosecutorial vindictiveness successfully rebutted by showing that a different U.S. Attorney reviewed the case before trial and made the determination that augmented charges were warranted); United States v. Partyka, 561 F.2d 118 (8th Cir. 1977) (any fear of vindictiveness rebutted because reversal of conviction on appeal disclosed informant’s identity and thus removed government’s reason for not earlier seeking prosecution on stiffer charges); United States v. Preciado-Gomez, 529 F.2d 935 (9th Cir.) cert. denied, 425 U.S. 953, 96 S.Ct. 1730, 48 L.Ed.2d 197 (1976) (defendant’s overall conduct, which lead to an increased charge after a mistrial was declared, was not fully known by the prosecutor until after mistrial was declared); United States v. Jamison, 164 U.S.App.D.C. 300, 505 F.2d 407 (D.C. Cir. 1974) (increase in charges could be justified by intervening events or by discovery of new evidence of which the government was excusably unaware. Court did not undertake “to anticipate definitively all those circumstances which might be thought to warrant a prosecutor’s escalation of an initial indictment”).
. See Miracle v. Estelle, 592 F.2d 1269 (5th Cir. 1979) (“. . vindictiveness is always an illegitimate consideration in prosecutorial decisionmaking”); Hardwick v. Doolittle, 558 F.2d 292, 299-300 (5th Cir. 1977), cert. denied, 434 U.S. 1049, 98 S.Ct. 897, 54 L.Ed.2d 801 (1978) (vindictiveness is an improper factor in the prosecutorial process). In Bordenkircher v. Hayes, 434 U.S. 357, 98 S.Ct. 663, 54 L.Ed.2d 604 (1978), the Court found no due process violation even when the prosecutor admitted that he acted out of vindictiveness. As I will discuss below, however, I think that Borden-kircher presents a unique factual situation which is inapplicable here.
. See, e. g. United States v. Ricard, 563 F.2d 45, 48 (2d Cir. 1977) (prosecutor brought stiffer superseding indictment after a conversation with defense counsel in which the latter indicated that the defendant was prepared to go to trial. Court questioned whether this was enough to raise a likelihood of vindictiveness). See also Jackson v. Walker, 585 F.2d 139, 146-148 (5th Cir. 1978) (overall circumstances of the case presented a “remote” likelihood of prosecutorial vindictiveness).
. See e. g. Hayes v. Cowan, 547 F.2d 42, 43 n.2 and 45 (6th Cir. 1976), rev’d sub nom. Bordenkircher v. Hayes, 434 U.S. 357, 98 S.Ct. 663, 54 L.Ed.2d 604 (1978) (prosecutor admitted vindictive motive). See also North Carolina v. Pearce, 395 U.S. 711, 715, 726, 89 S.Ct. 2072, 23 L.Ed.2d 656 (1969) (district court had found the conclusion “inescapable” that state trial judge had increased sentence in retaliation for an initial successful appeal); United States v. McFadyen-Snider, 590 F.2d 654 (6th Cir. 1979) (perjury indictment would probably not have been brought had defendant not succeeded in her appeal from an initial conviction); United States v. Groves, 571 F.2d 450, 454 (9th Cir. 1978) (“The conclusion is inescapable on this record that the government brought the marijuana charge in retaliation for the appellant’s exercise of his statutory rights on the cocaine charge.”)
. See Chapman v. Estelle, 593 F.2d 687 (5th Cir. 1979) (defendant succeeded in having his bargain-for guilty plea set aside. Prosecutor refused to bargain further and brought enhanced charges. Not vindictive for prosecutor to refuse to reoffer same plea bargain which defendant had gotten set aside); United States v. Snell, 592 F.2d 1083 (9th Cir. 1979) (substituted charge carried same maximum sentence); United States v. Thurnhuber, 572 F.2d 1307 (9th Cir. 1977) (where mistrial was declared by the trial judge, prosecutor’s action in adding additional charges prior to retrial was not vindictive since it was not done in response to defendant’s assertion of a procedural right). Martinez v. Estelle, 527 F.2d 1330 (5th Cir. 1976) (defendant agreed to waive a jury trial in exchange for prosecutor’s waiver of a habitual offender count and recommendation of a 15 year sentence in the event of conviction. Conviction and 15-year sentence resulted, but reversed on appeal. No vindictiveness where, at retrial, defendant was convicted and sentenced or original, stiffer charges after rejecting the same agreement he had been offered earlier); United States v. Anderson, 514 F.2d 583 (7th Cir. 1975) (defendant agreed to plead guilty in exchange for a lowering of the charges against him. When plea bargain was set aside, original charges were reinstated.)
. In Blackledge itself, the Court indicated that had the state shown that it had been previously impossible to proceed on the more serious charge, then no vindictiveness problem would exist. See Blackledge, supra, 417 U.S. at 21 n. 7, 94 S.Ct. 2098.
. This approach is akin to that used in United States v. McFadyen-Snider, 590 F.2d 655 (6th Cir. 1979). There, this court affirmed a factual determination that the bringing of a perjury indictment after a successful appeal had the effect of punishing the defendant for exercising her rights.
In that case, the factual determination was one of actual vindictiveness — the prosecutor probably would not have brought the perjury charge had the defendant lost her appeal.
. Obviously, the length of additional potential incarceration is a significant factor in determining the extent to which the prosecutor’s conduct chilled defendants in the exercise of their rights. Cf. United States v. Sturgill, 563 F.2d 307 (6th Cir. 1977).
. The district court suggested that only two explanations for apparently vindictive behavior would suffice in all cases: 1) lack of existence of essential elements of the increased offense at the time of the original indictment; 2) discovery of new evidence of which the government was previously unaware. See 444 F.Supp. at 1244. I agree with the district court that this type of objective explanation is what is required for the prosecutor to meet his heavy burden of offsetting the appearance of vindictiveness.
I disagree with the notion that this type of explanation is required in all cases. Should application of the balancing test I advance in this dissent result in a presumption of actual vindictiveness, then neutral explanations such as those advanced by the prosecutor in this case would suffice, if credited, to offset the presumption. See Hardwick v. Doolittle, supra, 558 F.2d at 301.
. Contrast Miracle v. Estelle, 592 F.2d 1269 (5th Cir. 1979) (substantial defendant’s interest and weak prosecution interest. In addition, state did not try to explain its conduct.); Jackson v. Walker, 585 F.2d 139, 148 (5th Cir. 1978) *254(“. .a very limited due process interest balanced against a moderately weighty prose-cutorial independence interest.”)
. The government also argued that a defendant is so interested in getting out on bail that even a seemingly vindictive increase in charges would not deter the exercise of that right and that thus only a finding of actual vindictiveness should be required.
Given the complex and subjective nature of deterrence, it cannot be said that no deterrent effect results in situations such as this. In addition, there exists independent value to minimizing the perception of defendants — and the public — that vindictive prosecutorial behavior takes place. Further, I do not see any significant difference between a defendant’s interest in being released on bail and a defendant’s interest in having his conviction reversed on appeal (Pearce) or securing a trial de novo (Blackledge).
. The prosecutor in Blackledge did not try to justify his conduct. However, I do not think, that the Supreme Court would have ruled differently had the prosecutor advanced reasons such as inexperience or neglect.
. I agree with Judge Green that in situation # 1, where the government increased the severity of the charges arising from a single criminal act, Blackledge applies. Put in terms of the balancing test I advance, the appearance of vindictiveness is very strong and the government’s interest is slight. I also agree with Judge Green that explanations such as inadvertence or neglect never can be sufficient to justify the government’s conduct in such a situation. In such a situation, only new information or previous legal impossibility should justify increasing the charges.
. Thus, although purporting to follow Jackson v. Walker, Judge Green, in reality, extends the lenient standard of Hardwick v. Doolittle to all fact situations involving the addition (as opposed to the substitution) of charges.
In contrast, I would extend the balancing test of Jackson v. Walker, to all situations. Under my suggested approach, a district court would balance the competing interest of the prosecutor and the defendant in all cases to arrive at one of two possible standards. Under the first standard, a prima-facie case of vindictiveness arises which can be easily offset if the prosecutor convinces the court that he really wasn’t retaliating against the defendant. Explanations such as inadvertence or inexperience are fine. Under the alternate standard, as I phrased it, the government must offset any “appearance of vindictiveness.” To do this, explanations such as inadvertence or inexperience are not fine. The only acceptable explanations are those advanced by the district court, lack of knowledge on the government’s part or legal impossibility. I thus propose a two-tiered standard. The first tier is identical to the standard used by the Fifth Circuit in Hardwick. The second tier is identical to the standard used by the Ninth Circuit. I would affirm the district court because application of the balancing test should result in application of the more severe standard, which the government cannot meet.
. I fail to see why Judge Green is concerned by the phrase “appearance of vindictiveness.” The fact that the Ninth Circuit cases with which Judge Green disagrees also use this expression should not make it unsuitable. (I note that this expression was also used in Jackson, 585 F.2d at 144). As I read Pearce and Blackledge they are concerned: 1) with forestalling prosecutorial behavior which is actually vindictive; 2) with ensuring that prosecutorial behavior which appears to be vindictive (even if it’s really not) does not create an apprehension of vindictiveness which deters defendants from exercising their rights. The Supreme Court in Blackledge specifically used the phrase “apprehension” of vindictiveness. The phrase “appearance of vindictiveness” is useful because it serves to distinguish situations where actual vindictiveness occurred, as in Bordenkircher, from situations where no one claims that the prosecutor acted in bad faith. I see no real difference between that phrase and expressions such as “unexplained addition of charges,” or the Supreme Court’s language of defendants’ “apprehension of vindictiveness.”
. Curiously, Judge Green does not cite United States v. McFadyen-Snider, 590 F.2d 654 (6th Cir. 1979) where this court barred the bringing of a perjury indictment on a finding that it probably would not have been brought had the defendant not been successful in a separate appeal. Judge Green’s narrow view of prose-cutorial vindictiveness is difficult to reconcile with McFadyen-Snider. Also, I would accord no significance to the vague dicta in United States v. Sturgill, 563 F.2d 307, 309 (6th Cir. 1977).
. Both Judge Green, ante at p. 241 n. 7 and Judge Pratt below, 444 F.Supp. at 1243 n. 4 have commented on the problem of prosecuto-rial overcharging as a possible consequence of stringent judicial views of vindictiveness questions. Ideally, a reasonable prosecutor, aware of his awesome charging authority, will make a careful judgment of what charges should be brought against a given defendant. Unfortunately, the reality of prosecutorial decision-making is far removed from this ideal scenario. Given the large amount of overcharging which goes on now, 1 see little danger that it will increase if we act firmly when dealing with vindictiveness issues.
. Judge Merritt’s rhetorical questions, I think, are answered in this dissent. A balancing test can weigh each of the factors he cites on either side of the plaintiff-defendant equation. I note once again that vindictiveness issues are only present if a prosecutor adds charges which he could have, but did not bring earlier. Any resultant problems, in effect, are of the prosecutor’s own making.
. Judge Merritt’s suggested limitation of Blackledge to situations after the first trial is over is even a broader argument than that made by the government which would limit Blackledge to situations involving a trial or retrial. See n. 2, supra.