The principal question presented by this appeal is the voluntariness of a confession. After selling a modest quantity of cocaine
*1128
(9.47 grams, 83 percent pure) to an undercover agent, Roger Rutledge was indicted by a federal grand jury and then arrested by federal officers and placed in custody in the federal courthouse in Chicago to await his initial appearance before a magistrate. While they were waiting, the officers gave Rutledge, both orally and in writing, the warnings required by
Miranda v. Arizona,
Pursuant to a plea agreement in which Rutledge agreed to cooperate with the government and the government agreed to drop one of the charges against him and to bring his cooperation to the attention of the sentencing judge, Rutledge pleaded guilty to the sale for which he had been arrested. The confession he had made after being arrested, however, had revealed that he had dealt in a substantially larger quantity of drugs than the government had suspected or had charged him with dealing, and this revelation that he had engaged in criminal conduct related to the offense to which he had pleaded guilty increased his sentencing exposure under the Sentencing Guidelines from between 12 and 18 months in prison to between 57 and 71 months. The judge sentenced him to 71 months. The appeal challenges the propriety of the judge’s relying on the confession to jack up Rutledge’s sentence.
We are bound by the judge’s finding of who said what; that is a determination of credibility, and can rarely be disturbed on appeal — and not in this case. There is nothing inherently incredible about the officer’s version of what was said, though it required some prompting to get him to say it. But we make an independent judgment as to whether, all circumstances considered, the confession that Rutledge made after waiving his right to remain silent was voluntary and therefore admissible.
Colorado v. Connelly,
United States v. Hawkins,
But this is all by the way. The government (as if oblivious to our opinions even when they are favorable to it) does not question the validity of
Hawkins,
and we must therefore make an independent determination of the voluntariness of Rutledge’s statement. It is a task complicated by the fact that courts have not been successful in devising a standard that will determine in a consistent fashion when confessions should be excluded on grounds of involuntariness. Of course if the confession is
unreliable,
it should go out, along with other unreliable evidence. It is on this basis that confessions extracted by torture are excluded. But in most cases in which a confession is sought to be excluded because involuntary, there is little likelihood that the inducements placed before the defendant were so overpowering as to induce an untrue confession. The courts in such cases retreat to the proposition that a confession, to be admissible, must be the product of a free choice.
Watts v. Indiana,
An alternative approach, which is implied by
Connelly
and may well describe the courts’ actual as distinct from articulated standard, is to ask whether the government has made it impossible for the defendant to make a
rational
choice as to whether to confess — has made it in other words impossible for him to weigh the pros and cons of confessing and go with the balance as it appears at the time. This approach, sketched in
Weidner v. Thieret, supra,
If the officers, fully intending to use anything Rutledge said against him, had said to him, “Tell us all you know about the drug trade, and we promise you that nothing you tell us will be used against you,” then he would have a strong argument that any ensuing confession had been extracted by fraud and was involuntary.
Quartararo v. Mantello,
At the other extreme, if the officers had merely promised Rutledge to inform the prosecutors of his cooperation, or had stated that, other things being equal, cooperation is helpful to an accused, or had reminded him that while cooperation might help him the government would not hesitate to use anything he said against him, there could be no serious argument that he had been coerced to confess. Our case is in between. The challenged statement was not quite a promise not to use anything Rutledge said against him; at the same time, as the judge found, it was not quite truthful.
Rutledge does not deny having been given the
Miranda
warnings and having understood them; the voluntariness of his waiver of his
Miranda
rights, as distinct from the voluntariness of his confession, is not contested. One of the warnings is that anything you say may be used against you. When moments after signing his waiver of
Miranda
rights Rutledge asked whether cooperation would help him, he could not have meant by this — having just been told that anything he said could be used against him — to be asking whether he could confess with impunity to mass murder. He could not have meant to be asking whether, if he confessed to being the head of a Colombian drug cartel, the government would do its best to get him a lighter sentence for the crimes for which he had been indicted. He may have been asking just whether the
fact
of his confessing would receive consideration as cooperation. The officers said it would. This was not, so far as appears, a lie when made (which means, it was not a lie at all). The government later agreed to drop one of the charges against Rutledge and it informed the judge of his cooperation. Government is not forbidden to “buy” information with honest promises of consideration.
Fare v. Michael C.,
An alternative interpretation of the officer’s statement, however, is that it promised Rutledge a
net
benefit from spill
*1131
ing the beans. If this was the promise, it is unlikely that the officer intended to keep it; and if he did not, then the statement was fraudulent. But it was the sort of minor fraud that the cases allow. Far from making the police a fiduciary of the suspect, the law permits the police to pressure and cajole, conceal material facts, and actively mislead — all up to limits not exceeded here.
United States v. Guerrero,
There was no net benefit to Rutledge, clearly; was there any benefit? We have said there was, yet against this it can be argued that the agreement to drop one of the charges against Rutledge and to bring his cooperation to the judge’s attention was in exchange not for his confessing to a larger role in the drug trade but simply for his forgoing his right to a trial. If so, then it can be argued that the confession helped him not at all — the government had reneged on its promise that his cooperating would help him. The defendant’s forgoing his right to trial is an important consideration in virtually every plea bargain, because the government does not have the resources to try every criminal defendant. But cooperation in the investigation of other offenders is also an important consideration in plea bargaining, and there is no reason to doubt that it was important here. A better argument is that the written plea agreement recites as the consideration for dropping the charge that was dropped Rutledge’s cooperation in testifying against and otherwise helping the government bring his accomplices to justice; the agreement makes no mention of the confession. But the confession was the beginning of the cooperation, for it was the confession that revealed that Rutledge was involved in more than bush-league drug dealing. Had Rutledge not revealed his part in a larger drug operation his cooperation would have been less valuable — would not have commanded as high a “price” in the plea-bargaining market.
Although we have been treating this case so far as one in which the defendant’s confession is used as evidence to convict him, in fact Rutledge was convicted only of the sale to the undercover agent— not of the broader activities to which he confessed. (The count that was dropped was conspiracy to distribute the small quantity of drugs sold to the undercover agent.) The confession was used to jack up his sentence. As an original matter one might argue that an involuntary confession could be used for
this
purpose, since a wide variety of otherwise inadmissible evidence is traditionally usable when it comes to deciding on the proper sanction for the crime of which the defendant has been duly convicted. But the Supreme Court has scotched the possibility,
Estelle v. Smith,
The sentencing context is relevant, however, because Rutledge is trying to gain mileage from section lB1.8(a) of the Sentencing Guidelines, which provides that “where a defendant agrees to cooperate with the government by providing information concerning unlawful activities of others,
and the government agrees that self-incriminating information so provided will not be used against the defendant,
then such information shall not be used in determining the applicable guideline range.” The section is inapplicable, because the condition we have italicized was not satisfied. The officer who questioned Rutledge after he was arrested promised (if it was a promise) that, if he cooperated, his cooperation would be helpful to him; it did not promise him that if in cooperating he confessed to other crimes the information would not be used against him.
United States v. Hallam,
We are inclined to think, by the way, that the italicized language in the Guidelines should be interpreted narrowly. The provision on agreements is addressed to prosecutors rather than to police, and it would *1132 wreak much havoc with prosecutorial prerogatives if casual inducements by police officers were treated as enforceable plea agreements. It is true that the provision speaks of “the government” rather than of the Department of Justice, and that it is not limited to formal plea bargaining. Yet, although Rutledge had been indicted, the interrogation of Rutledge followed immediately upon his arrest and before there were even the glimmerings of plea negotiations. So far as appears, no prosecutors were present, and it is certain that Rutledge was questioned not by prosecutors but by ordinary drug enforcement agents. It is certain too that the interrogating agents were not mindful of the possible interplay between statements that they might make to him and his entitlements under the Sentencing Guidelines.
If all we have said thus far is wrong, there is still the language of the plea agreement to be considered. The agreement states that “no promises or representations have been made, nor agreements reached, other than those set forth in this agreement.” This is the equivalent of an integration clause in a regular contract, and negates any effort by Rutledge either to undo the terms of the plea agreement by pointing to an alleged promise, made before the agreement was signed, to reward him for his cooperation, or to use that promise as the basis for an agreement that his related criminal conduct would not be used to increase his sentence. So even if the confession was extracted by a promise, that promise merged into the plea agreement; and the government did not violate the plea agreement.
Hartman v. Blankenship,
There is one other issue. In order to use the drug sales to which the defendant confessed to increase his sentence for the sale to the undercover agent, the judge had under the Guidelines to find that they “were part of the same course of conduct or common scheme or plan as the” sale of which he was convicted. § lB1.3(a)(2). He did find this, and his finding is not clearly erroneous and therefore binds us.
United States v. Vopravil,
Affirmed.
