Web Content Display Web Content Display


The Army Lawyer


A New Paradigm for Plea Agreements Under the 2016 MJA



  PDF Version
(Credit: istockphoto.com/Michal Chodyra)

Web Content Display Web Content Display

The Military Justice Act of 2016 (2016 MJA)1 made sweeping changes to the Uniform Code of Military Justice (UCMJ). One of the most significant procedural revisions came from new statutory plea agreement authority. While prior plea agreements relied on a convening authority’s promise to exercise clemency in exchange for the accused’s plea, the 2016 MJA provides a convening authority, pursuant to a plea agreement, the ability to limit the sentence adjudged directly.2

The concept of a clemency-based agreement bolsters much of the robust plea agreement case law developed over the past sixty years. But the near complete evisceration of clemency powers accomplished by the 2016 MJA, together with new plea agreement authority, threatens to disrupt decades of military justice practice in subtle, yet profound ways. In light of this coming revolution in practice, new theoretical guideposts are necessary.

This article aims to provide a new operational paradigm, based on contractual performance analysis, to expose the depth and breadth of changes to plea agreement practice. Rather than a simple critique of the law, this article endeavors a systematic analysis and comparison of the legacy and the new plea agreement systems through the prism of performance order, giving the reader a robust mental approach to stitch together the many procedural changes into a single theoretical tapestry. This article will also walk through a fictitious scenario to demonstrate how different decisions by the government and the accused can produce different issues and outcomes.

This article will take a close look at the legacy system through the lens of clemency authority, revealing how clemency has affected our understanding of the plea agreement system and presaging how a move away from a clemency-based system may influence practice under the 2016 MJA. Provisions of the 2016 MJA plea agreement system will also be laid out. This article will deep-dive into the new system, demonstrating predictable friction points, and offering ways to avoid or resolve them. Finally, it will conclude with advice to convening authorities and judges on how to navigate within the new paradigm.

Evaluation of the Legacy System: Performance by the Accused, then Performance by the Convening Authority

In the legacy military justice system, an agreement between the convening authority and the accused attaches upon beginning of the accused’s performances—generally through the entry of a guilty plea.3 After accepting the plea, the court-martial conducts presentencing proceedings and sentences the accused.4 After sentencing, the case returns to the convening authority, who performs by disapproving any portions of the sentence that exceed the limitations specified in the agreement.5

The preceding paragraph presents a simplified view of trial procedure to make a point on performance order: first the accused, then the convening authority. This order of performance arose by virtue of a statutory scheme that gave the convening authority the last say over an accused. It also gave rise to five significant corollaries that are now doctrine. They are discussed in turn below.

Independent Sentencing, then Clemency

Perhaps most significantly, the legacy system order of performance enables the military judge to sentence independent of the agreement.6 Because the convening authority performance is in the form of clemency, the court is free to sentence as the evidence and conscience dictate. Under this model, the accused obtains the benefit of the lesser of the court’s sentence and the pretrial agreement. For example, if the pretrial agreement contains a clause in which the convening authority agrees to approve no more than two years of confinement, but the accused receives a sentence to three years of confinement, then the convening authority may approve only two years of the sentence. The accused benefits from a one year reduction in sentence. On the other hand, if the same case results in a sentence to only one year, then the accused will serve only one year. Clemency action would be irrelevant, because there is no sentence in excess of the agreement.

Because the entirety of the convening authority performance takes place after the court-martial, the court need not know the contents of any sentence limitation until after the sentence is announced. After the sentence is rendered, the judge will briefly inquire into the limitation in the agreement to ensure the accused’s understanding and a meeting of the minds about the limitation.7

Protection for an Accused Once Performance Begins

The legacy system order of performance dictates additional protections for an accused who has initiated, but not completed performance.8 A convening authority may withdraw at any point, but once the accused “begins performance of promises contained in the agreement,”9 the convening authority is no longer free to withdraw from the agreement.10

The agreement between the accused and the convening authority creates a problem seen in unilateral contracts—ones that can only be accepted by performance.11 These types of contracts create an issue where the offeree (in this case, the accused) has started performance, but has not fully completed performance. In a true unilateral contract, the offeror would be free to withdraw from the agreement even if the offeree had nearly completed performance.12 This legal “problem” occurs in every legacy plea agreement because it is inherent in the system: the convening authority must perform last, after the accused has completed performance. There is no provision for a convening authority to perform stepwise.

If the accused has started performance—perhaps by entering into a stipulation of fact—but has not completed performance on other terms of the agreement, what protections does an accused have from a convening authority who would withdraw from the agreement? The law protects a performing accused by binding the convening authority to perform once the accused begins performance.13

Built in Protection for the Convening Authority upon the Accused’s Withdrawal

The flip side of the unilateral contract problem is that it provides built in protections for the convening authority where the accused abandons the agreement. The accused may withdraw from the agreement at any point, even after acceptance of plea.14 The convening authority generally is entitled to retain the benefits of the accused’s performance if the accused withdraws.15 While the accused may withdraw from the agreement at any time, the accused’s options for “undoing” his past performance begin to narrow as time goes on. For example, if an accused testifies against another Soldier, pursuant to an agreement, but then later decides to withdraw from the agreement, the convening authority would still have the benefit of the earlier testimony. Thus, an accused who abandons the agreement will generally leave the convening authority in no worse position than without the agreement.

Perhaps the most significant alteration of the accused’s ability to “undo” the previously taken actions occurs at announcement of sentence.16 If the accused successfully withdraws from the agreement before announcement of sentence, he would ordinarily be permitted to withdraw his plea as well, and the case would be set for trial.17 However, if the accused withdraws from the agreement after announcement of sentence, the accused would not be able to withdraw the plea.18 Regardless when (or why) the accused withdraws from an agreement, the withdrawal releases the convening authority from the obligation to grant clemency under the agreement.19 Thus, an accused is likely to withdraw from an agreement only where undesirable terms remain to be performed,20 or where there is nothing to be gained from remaining in the agreement.21

While the convening authority obtains the benefit of whatever the accused has performed before withdrawal, the convening authority also loses nothing from an accused’s withdrawal, no matter when it occurs. Even if the accused withdraws from the agreement while the ink is still wet, the convening authority “loss” is mere prospective performance. In the context of a court-martial, this is no loss at all. While an accused’s withdrawal may make the case more difficult to prove, the government’s ultimate interest in ensuring justice through due process remains unharmed.

Protection for the Convening Authority on the Accused’s Failure to Perform

The legacy system order of performance also protects a convening authority if the accused simply fails to perform. A failure of performance is qualitatively different than a formal withdrawal, though the protections for the convening authority are similar. In a failure to perform, the accused breaches the agreement because of inability or unwillingness to complete the terms of the agreement. Protections of the convening authority extend even when the failure to perform is after the sentence has been adjudged. For example, consider an accused who agrees to conform his behavior to certain standards while awaiting the convening authority’s clemency decision. If the accused were to engage in misconduct after the court-martial was complete but before the convening authority’s clemency action, the convening authority would be free to withdraw from the agreement because the accused is in breach.22

At this point, the independent sentence adjudged by the court is a significant protection for the convening authority. There need not be any re-sentencing hearing; rather, the convening authority may simply decline to exercise clemency, and the sentence as adjudged stands.23 Thus, the order of performance protects the convening authority’s interest in judicial economy.

Post-trial Agreements Possible

Finally, the order of performance makes possible a post-trial agreement between the accused and the convening authority. While such agreements are not common, several cases have demonstrated the legal possibility of post-trial agreements. In United States v. Dawson, the accused released the convening authority from the pretrial agreement, in exchange for a new agreement in which the convening authority agreed not to prosecute new charges.24

After trial, and before action, the accused may modify the pretrial agreement, “so long as the accused has the assistance of counsel, the modification is the product of a fully informed and considered decision, and it is not the product of a coercive atmosphere.”25 While it is unclear how far Dawson extends, it is clear that the order of performance is what makes such agreements possible.

The five features of the legacy system discussed above do not exist in a vacuum; rather, they exist because the statutory clemency system creates an environment conducive to their existence. The legacy system provides no independent authority for a convening authority to direct the sentence of a court-martial. Therefore, the only bargaining chip a convening authority has is clemency action. Moreover, a judge must sentence in a vacuum because the law provides no legitimate authority for a convening authority to direct the sentence of a court-martial; in fact, if anything, Article 37, UCMJ, would prohibit influence over the judicial acts of a court-martial.26 Once the agreement is formed, the law protects the accused, who performs first, and must be able to rely on the promises of the convening authority. In turn, the convening authority is assured of obtaining the benefit of the bargain because if the accused withdraws or otherwise fails to perform after announcement of sentence, the convening authority generally gets to keep the benefit.

Plea Agreements Under the 2016 MJA

A Brief History of How We Got Here

In recent years, the once robust powers of clemency have suffered a near total collapse that started with the case of United States v. Wilkerson and concluded with the 2016 MJA changes. In Wilkerson, a convening authority overturned the sexual assault conviction of an Air Force lieutenant colonel.27 The case enjoyed significant publicity,28 and resulted in a 2014 revision to the UCMJ that sharply limited clemency authority in non-plea agreement cases.29

While the amended clemency rules allowed a “carve out” specifically for pretrial agreements, the damage was done.30 The 2016 MJA replaces the post-trial clemency process with a sleek new “entry of judgment” model that gives judges control over the post-trial process, and convening authorities optional, token clemency authority.31

In place of clemency powers, the 2016 MJA supplies the convening authority a powerful new tool to issue a direct limit on the sentence of the court-martial. The new system requires the convening authority to perform first—before trial—by issuing a direct limitation effective against the court-martial deliberative process.

Convening Authority Power to Limit the Court-Martial—The New Statutory Authorities

With the brief history in place, we now turn to a discussion of the 2016 MJA provisions governing plea agreements. Article 53a of the 2016 MJA allows a convening authority to agree to a “limitation[] on the sentence” a court-martial may adjudge.32 The language of the statute as implemented in Rule for Court-Martial (RCM) 705 allows the convening authority and the accused to agree to “limitations on the maximum punishment, . . . the minimum punishment, and . . . the minimum and maximum punishments.”33 In order to facilitate the limitation, sentencing occurs with knowledge of the terms of the agreement.34

To facilitate this, the military judge will be informed of the plea agreement terms, to include any sentencing limitation.35 In cases where sentencing is to be before a panel, the plea agreement is not disclosed to the members unless the accused requests disclosure.36 However, just as in the case of the military judge, the sentence the members may impose will be limited—the members “shall vote on a sentence in accordance with the sentence limitation.”37 In practice, the members may simply be informed of any maximum or minimum without referring to the agreement itself.

The plea agreement limits are not binding until after the military judge has accepted the agreement.38 Once the military judge accepts the agreement, it is binding on all parties and the court-martial.39

The new Article 53a, UCMJ, requires military judges to exercise a gatekeeping role. The judge is obligated to reject agreements when the terms have not been accepted by both parties,40 or when the accused does not understand the agreement.41 The military judge does have limited authority under the accompanying rules to reform such agreements,42 but “may not participate in discussions” regarding the agreement.43 Moreover, the military judge must reject terms that are “prohibited by law”44 or are “contrary to, or inconsistent with, a regulation prescribed by the President . . . .”45

Finally, the military judge must reject terms that violate mandatory minimum provisions for certain sex-related offenses.46 The relevant offenses are listed at Article 56(b)(2), UCMJ, and include rape,47 sexual assault,48 rape of a child,49 sexual assault of a child,50 as well as attempts,51 and conspiracies to commit such offenses.52 The relevant mandatory minimum for these offenses is a dismissal or dishonorable discharge.53

Two significant exceptions allow judges to accept plea agreements for less than the mandatory minimum. First, the military judge may accept a plea agreement that provides for a bad conduct discharge.54 Second, the military judge may accept a plea agreement free of any mandatory minimum where the trial counsel recommends the sentence in exchange for “substantial assistance by the accused in the investigation or prosecution of another person . . . .”55

Evaluation of the 2016 MJA System: Performance by the Convening Authority, Then Performance by the Accused

Even with the thumbnail sketch presented so far, it should be obvious that the new plea agreement rules represent a significant paradigmatic shift from the legacy system. In the legacy system, the convening authority performed last—i.e., by granting clemency as agreed. The new system requires the convening authority to perform first—before trial—by issuing a direct limitation effective against the court-martial deliberative process.

What may not be obvious is how this shift affects other aspects of performance by the government and the accused. First, the convening authority now has a heightened obligation to understand trial process and the rights of the accused in order to set up the deal correctly. Second, the convening authority’s new powers over sentencing means that the convening authority will need to observe some reasonable boundaries to avoid intruding upon the province of the court. Third, and finally, the military judge will have greater gatekeeping and policing obligations to ensure the agreement does not fail for public policy reasons. Each will be discussed in turn below.

The Convening Authority’s Obligation to Understand Trial Rights and Processes

Under the legacy system, the convening authority did not need a sophisticated understanding of the accused’s rights or trial processes. After hearing the sentencing evidence, the court-martial adjudged an “all in” punishment—thus providing a scaffold upon which clemency could act. While some understanding of the judge’s sentencing options was important, the plea agreement and the sentencing procedure ran parallel and independent of each other. The court-martial did not need to work with the convening authority, and the convening authority did not need to work with the court-martial. Thus, while it was possible a convening authority could enter into a poorly negotiated agreement, the risk of entering into a legally insufficient agreement was attenuated.

Under the 2016 MJA, the convening authority now has much more opportunity to botch the agreement because the convening authority may directly limit56 the maximum punishment, the minimum punishment, or both.57 The practical effect of prospective performance is that the convening authority must have a solid understanding of the 2016 MJA trial processes to ensure that the limitations agreed to will be effective.

The Convening Authority Must Understand the Impact of Forum Choice

The first significant process the convening authority must understand is forum selection. Forum selection occurs when the accused selects by whom he wishes to be tried. The two possible forums are before military judge and before a military panel. The right to select forum is common to both the 2016 MJA and the legacy system; however, under the 2016 MJA, the election has a much more significant effect on the sentencing procedures. 58

Under the 2016 MJA, if the accused elects sentencing by a military judge, the sentence will be segmented. This means that the judge must determine any appropriate confinement or fines on a per specification basis.59 If there is no confinement or fine to be awarded for the specification, the judge will so state.60 Moreover, once all confinement has been adjudged, the military judge must determine which confinement terms will be served consecutively and which will be served concurrently.

The sentencing rules provide certain factors for the judge to consider in making that determination: whether the specifications “involve[] the same victim and the same act or transaction”; whether the plea agreement speaks to the issue; whether the specifications are unreasonably multiplied; whether the total sentence is justified on the merits of the case;61 and whether, in a special court-martial, the sentence must be adjusted to meet the maximum confinement authorized. Given these considerations, confinement for some specifications may run concurrently, while for other specifications it may run consecutively.

While sentences to a fine and confinement are segmented, the remainder of the sentence is unitary.62 This means that the remaining portions of the sentence will be adjudged on an “all in” basis, rather than on a specification-by-specification basis. For example, if the accused were to receive a punitive discharge, the military judge would not specify the offense(s) upon which the discharge was adjudged.

If the accused instead elects sentencing by a panel, the procedures are much simpler. As under the legacy system, the panel will simply adjudge a “single sentence” for all the specifications of which the accused is found guilty.63 There is no sentence segmentation, and there is neither authority nor procedure for a panel to adjudge confinement in consecutive or concurrent terms.

The Convening Authority Must Understand When Sentencing by a Judge Is Available

The default sentencing forum is before military judge.64 The accused is entitled to elect a panel for sentencing only if he has been convicted of an offense by a panel.65 Otherwise, the accused will be sentenced by military judge. Rule for Court-Martial 1002(b) obliges the military judge to determine “[i]n a court-martial consisting of a military judge and members, upon the announcement of findings . . . whether the accused elects sentencing by members in lieu of sentencing by the military judge.”66 While the Rule goes on to note that the “military judge shall determine the sentence” unless a timely election has been made,67 the Rule does not re-state the explicit requirement under Article 53(b)(1)(A), UCMJ, that the military judge is the default sentencing authority, and does not re-state the specific conditions under which the accused may elect sentencing by a panel.

Unfortunately, the ambiguous wording in RCM 1002 appears to have complicated the plain language of the statute. In fact, the Army’s current Military Judge’s Benchbook provides that an accused may elect sentencing by a panel regardless whether the accused has been convicted of any offense by the panel.68 Both the confused wording of RCM 1002 and the Benchbook provisions appear to rest on a legacy-influenced understanding that a court-martial must assemble as a pre-requisite to sentencing the accused.69

This article will endeavor to provide a more nuanced understanding of how the 2016 MJA rules governing assembly, forum, and court-martial composition makes clear that the option to elect sentencing by a panel is only available where a panel has convicted an accused of an offense.

We begin with assembly. Under both legacy and 2016 MJA systems, assembly is the bookend to the convening action. When an authorized person70 selects members, the court-martial does not exist until those members have been assembled by the military judge to hear charges referred to it.71 Upon assembly, the court may be said to be “composed” of members or, alternately, a military judge. In the context of an accused’s elections on court-martial composition, the term “forum” is frequently used.

Article 16, UCMJ, defines the court-martial in terms of the membership required to compose a court-martial.72 The compositional requirement is so vital that both the legacy and the 2016 MJA systems provide a default forum if the accused should fail to elect one. Under the legacy system, the default forum was trial before a court-martial consisting of officers.73 Thus, for example, if the accused were to voluntarily absent himself after arraignment, and before electing a forum, the trial could continue in absentia before a panel consisting of officer members. By contrast, under the 2016 MJA, while there is no default to a panel composed of officers,74 if there is no election by the accused, the court-martial will consist of the members selected by the convening authority.75 Under both systems, once the members are assembled, the members are required to be at every session of court, except as outlined in Article 39(a), UCMJ.76

Article 39(a), UCMJ, provides authority for the military judge to receive the plea of the accused, and to conduct an inquiry into the plea outside the presence of the members.77 Further, provided the accused is provident, the military judge will generally enter the finding during the Article 39(a) session.78

Under the legacy system, the accused could not be sentenced without assembling the court, because Article 39(a), UCMJ, did not provide any authority for a military judge to do so. Instead, even in cases where the accused pleaded guilty to all specifications, the accused was required to make a forum election that carried through to sentencing. For example, consider an accused who elected a court-martial before enlisted members at the arraignment, and then immediately entered a plea of guilty. In that circumstance, the military judge would hold an Article 39(a) session outside the presence of the members, during which he would conduct an inquiry into the plea and enter findings. After entry of the findings, the military judge would call for the members and assemble them to determine the sentence. By contrast, if the accused elected trial before military judge, then the military judge would assemble himself as the court-martial prior to sentencing the accused. In either scenario, the accused’s entry of a plea waived trial on those specifications, and it was not until sentencing that the forum election had any effect. A legacy system influenced understanding of the requirements for assembly and forum election—perhaps coupled with a desire to extend every possible protection to the accused—thus, likely accounts for a muddled application of the sentencing forum rules under the 2016 MJA.

An amendment to Article 39(a) under the 2016 MJA upends the traditional guilty plea requirement for assembly prior to sentencing. The amendment provides that the judge can sentence the accused at an Article 39(a) session.79 The significance of this provision is hard to overstate; in a case where the accused has pleaded guilty to all offenses, the military judge will sentence the accused without a forum election from the accused. By contrast, where a court-martial has been assembled with members and has found the accused guilty, the accused’s sentencing election effectively provides that the panel not be released before sentencing and adjournment of the court-martial.

The new Article 39(a), UCMJ, authority is especially significant when read in concert with Article 53(a), UCMJ. Article 53(a), UCMJ, is clear that an accused will be sentenced by a judge alone, unless the accused is “convicted of an offense in a trial by general or special court-martial consisting of members” and elects a panel. The converse is that if the accused pleads guilty to all offenses, there will be no trial on any issue of guilt, much less a trial before members.80 Under the 2016 MJA, the entire availability of a panel hinges solely on the question of whether there is a triable issue of guilt. If there is, and the panel finds the accused guilty, then the accused may elect to be sentenced by that panel. If, however, there is no triable issue of guilt, then the accused must be sentenced by military judge.

This reading is in concert with the language of the statute originally proposed by the Military Justice Review Group (MJRG). That language would have made judge-alone sentencing mandatory in every non-capital case.81 While there is no meaningful legislative history to explain why Congress retreated from this position in the 2016 MJA, it may be inferred by the plain language actually adopted in Articles 39(a) and 53, UCMJ, that Congress intended the exception for a narrow subset of facts: where a panel finds the accused guilty of an offense and the accused requests sentencing by that panel.

The accused may elect to continue a contested case before the same panel that has found him guilty. But if there was no trial (e.g., because the accused pleads guilty), then the court will never assemble and the accused will be sentenced by the judge who received his pleas.

Given the state of the law, and the currently fraught application of it, practitioners would be wise to ensure the plea agreement contains a mutual understanding of forum selection. The exact form of the agreement is less important than is language wherein all parties understand that sentencing will be before military judge alone. Second, while it may be technically possible in some circumstances that a panel could be legally available in a plea agreement case—for example, in a mixed plea scenario—the form of the agreement and the sentencing rules differ so drastically based on forum that it would be advisable to ensure that sentencing by judge alone is specifically stated in the agreement.

Which Sentencing Regime Will Be in Effect?

Whether the legacy or 2016 MJA rules apply depends on the date of commission of the crimes to be sentenced. The simplest cases will be ones where all specifications occurred either before or after the effective date of 1 January 2019. In these cases, the relevant procedures in effect at the time of the commission of the offenses will apply.

When the accused faces a mix of specifications committed both before and after the effective date, or if there are continuing offenses alleged that “bridge” the effective date, the rules provide that the legacy procedures will apply, unless the accused elects the 2016 procedures.82 Only one sentencing regime will apply, regardless when the offenses were committed.83

In bridging cases, practitioners should include a term addressing whether the legacy or 2016 MJA rules will apply. If there is to be no election of the 2016 MJA procedures, the best practice would be to include a provision for the accused to affirmatively waive the election. The form of the agreement should include a unitary sentence agreement, and the case should proceed entirely under the legacy clemency based rules.84

If the parties wish to proceed under the 2016 MJA sentencing procedures, the accused should agree to make that election as part of the agreement, and the agreement format should reflect the election. The agreement should tailor the sentence limitations on a specification-by-specification basis if the forum will be judge alone, while the agreement should include a unitary sentence if the accused is entitled to and elects a panel.

While the recommendations expressed above may sound straightforward to apply, the application can become quite complicated, depending on the case. To help flesh out the types of issues that may arise in “thinking forward” to sentencing, we will walk through different permutations of a hypothetical case: United States v. Smith.

United States v. Smith: A Fictitious Scenario

Suppose we have a case where Private Smith leaves Fort Swagg without authority. About six months later, while driving on post, he is apprehended by a military police officer (MP), and returned to military control. During the traffic stop, the MP asks him if he was “military.” Private Smith tells him, “Not now, I’m not.” The trial counsel drafts charges with one specification of absent without leave (AWOL) terminated by apprehension, and one specification of false official statement. While the defense agrees the accused was AWOL, they do not agree that Private Smith gave a false statement. Rather, the defense contends he was commenting on his unfortunate circumstance of having been apprehended, and expressing his expectation that he would be disciplined and thrown out of the military. Unable to reach an agreement on the false official statement, the government and the defense nevertheless agree to a guilty plea on the AWOL offense, and to a sentence of no fewer than two months and no more than six months of confinement. The agreement does not specifically address the false official statement, or whether the sentence limitation applies to the false official statement. The agreement also has no term regarding the election of a military judge.

A trial proceeds, the accused is found provident in his plea, and the judge accepts the agreement. The government now elects to proceed on the remaining charge.

Under the legacy system, an agreement to “leave out” forum selection and a contested specification would have been an unremarkable (though inefficient) arrangement. However, under the 2016 MJA, a partial plea agreement is not only inefficient, it may fail. To illustrate, let us consider the different permutations that could occur due to the agreement’s silence.

First permutation: The accused elects trial by judge alone on the contested charge, and is found not guilty of the false official statement. Under these facts, the judge will sentence the accused solely on the offenses subject to the agreement.85 On its surface, the agreement appears sufficient because it tells the judge all he needs to know—in this case, the minimum and maximum term of confinement. However, on closer inspection, the agreement is silent as to other possible terms of the punishment. For example, may the judge sentence the accused to forfeitures? Is the agreement a limitation only on the confinement term such that any other authorized punishment may be adjudged? For example, is the judge free to sentence the accused to a punitive discharge? Or is the agreement a limitation on the entire sentence such that no other authorized punishment may be adjudged? Under that reading, the accused may, at a maximum, be sentenced to no more than six months of confinement, with no other punishment. If the parties do not agree on how to read the agreement, the agreement fails for a meeting of the minds.

Second permutation: The accused elects trial by judge alone on the contested charge, but this time he is found guilty of the false official statement. The only new question presented here is whether the confinement limitation applies only to the AWOL offense, or if it applies to the total sentence for both offenses. Under the 2016 MJA, the judge must segment confinement and fine terms for each individual specification.86 Additionally, the judge must determine whether confinement is to run consecutively or concurrently,87 but the agreement is unclear about what the judge should do. If the agreement applies only to the AWOL offense, then the judge could sentence the accused for the false official statement completely free of the plea agreement limitations, exposing the accused to far more punitive liability. If the agreement applies to both offenses, then the judge would need to reverse-engineer the sentence to ensure that confinement is no fewer than two months, and no more than six.88 To the extent inquiry into the agreement discloses controverted terms, the judge must either conform the agreement to the understanding of the accused (with agreement of the trial counsel), or else permit the accused to withdraw the plea.89

Third Permutation: In this variation, the trial counsel again elects to prosecute the contested specification. This time, the accused elects trial by a panel, and is found not guilty of the false official statement. Under the 2016 MJA, the accused is ineligible to elect panel sentencing because he has not been found guilty of an offense by a panel.90 The default rule of military judge sentencing would apply. In this case, the judge would sentence the accused under the terms of the agreement, but the same structural concerns as in the first permutation above would come into play.

Fourth Permutation: In this scenario, the accused again elects trial by a panel. This time, he is found guilty of the false official statement. If the accused declines to elect sentencing by a panel, the default sentencing rules would apply and the military judge would sentence the accused. The same considerations listed above would apply.

Alternately, the accused could elect sentencing by the panel following separate procedures. Unlike a judge, the panel has authority only to deliver a “single sentence.”91 The panel cannot segment the sentence, and therefore the panel need not determine whether the confinement is to be served consecutively or concurrently.

In the Smith hypothetical, despite the apparent simplicity of the agreement, the judge would still need to determine whether the agreement represents the total limitation on the sentence, or whether the parties intended it to be only a limitation on the AWOL term. In the event the term was intended to apply solely to the AWOL term, the term would likely be unenforceable, because there is no mechanism to determine how much of a “single sentence” adjudged by the members was attributable to the AWOL. For example, if the panel adjudged eight months of confinement, there would be no way to determine how much confinement was attributable to the AWOL (which was required to be no less than two months, and no more than six), and how much was attributable to the false official statement.

Recap: Principles to Consider in “Thinking Forward” Toward Sentencing

The hypothetical case of United States v. Smith exposes many issues that could trap the unwary. Unfortunately, it likely represents just a few of the potential issues that could arise. While it would be impossible to forecast all possible “issue” scenarios, several easily foreseen issues with corresponding “best practices” follow:

First, practitioners should include a term as to forum selection. Sentencing before a military judge raises very different issues than sentencing before a panel. While the scenario in United States v. Smith illustrates one way an unexpected forum selection may complicate practice under the 2016 MJA, it is by no means the only way. A discussion of the 2016 MJA panel selection and empanelment rules is beyond the scope of this article; however, the practitioner should know the new rules would pose an unwelcome additional cost to judicial economy.

Second, practitioners should carefully consider any mandatory minimums that may apply and specifically address those in the agreement. Article 56, UCMJ, provides that punishment for certain sex offenses must include a dismissal or dishonorable discharge.92 However, Article 53a, UCMJ, provides limited circumstances under which a plea agreement may mitigate the mandatory minimum.93 For example, if the trial counsel intends to recommend no mandatory minimum on the basis of substantial assistance by the accused, then the agreement should specify the parties’ agreement as to what the government’s obligation is.94 Even where the parties do not contemplate that the agreement could affect the mandatory minimum, a term to that effect may be a best practice.95

Third, practitioners should take care to ensure that, for each offense, every lawful punishment has been addressed. To avoid unwieldy agreements, one possible practice would be to lay out the desired limits on a per specification basis, and then address all remaining punishment terms with a blanket statement such as, “[a]ny other lawful punishment may be adjudged in the discretion of the military judge.” This term would enable a judge more freedom to sentence an accused in addition to whatever else was in the agreement.96 Failure to include such a term could result in a situation similar to that arising in the Smith hypothetical, where it was unclear from the agreement whether the judge had any limit on his discretion with regard to unaddressed, but otherwise authorized, punishments.

Another possible practice would be to address all remaining terms with a blanket statement preventing the judge from adjudging any other punishment. If this provision were to be used, however, the agreement would need to address any applicable mandatory in order to avoid any argument about whether the government has obligated itself to perform on a limitation of the mandatory minimum.97

Fourth, practitioners should be careful to ensure that any conditions on referral do not impact the jurisdiction of the court-martial. Article 18(c), UCMJ, provides that a general court-martial has jurisdiction over certain sex offenses, and other attempts to commit such offenses.98 An agreement to refer any of the covered offenses to a special court-martial would fail for lack of jurisdiction.99

Fifth, practitioners should exercise particular care in cases involving conspiracies to commit named sex-related offenses. Article 56(b)(2), UCMJ, contains a list of offenses subject to mandatory minimums.100 This list is identical to that in Article 18(c), UCMJ, except that it includes the additional offense of conspiracy to commit the named sex-related offenses. Because the relevant mandatory minimum in such a case (dishonorable discharge) can only be adjudged by a general court-martial, the case must be referred to a general court-martial.101 However, in cases where the only offense giving rise to a mandatory minimum is a conspiracy to commit a covered sex offense, and the agreement includes a sentence limitation specifying a bad conduct discharge, then the case could be referred to a special court-martial.102

Sixth, practitioners should include a term addressing an election on sentencing rules where there are offenses subject to both legacy and 2016 MJA rules. As discussed earlier in this article, under RCM 902A, a court-martial cannot have mixed sentencing rules. In cases where the convening authority has referred offenses subject to legacy rules along with other offenses subject to the 2016 MJA rules, the legacy sentencing rules will apply unless the accused elects the new rules.103 Given that the election of the new rules would result in an agreement that would look very different from one under the old rules,104 the agreement should expressly address the election and structure the agreement to accommodate it.

To conclude, practitioners must “think forward” to sentencing because the plea agreement is a direct limitation on the sentence the court delivers. Unlike the legacy system, where the plea agreement is for a clemency modification of the sentence, practitioners under the new system will need to ensure that the convening authority’s sentencing direction reflects the understanding of the parties and is legally possible. Next, discussion will turn to how a direct sentencing limitation intersects with current statutory and case law designed to ensure justice and keep improper influence out of the tribunal.

Unlawful Command Influence and the Problem of the “Empty Ritual”

Under the 2016 MJA, the convening authority has the power to limit a judge’s sentencing authority, but such power runs into limits imposed of other doctrines. Recall that under the 2016 MJA, a plea agreement can provide a limit on the maximum sentence, the minimum sentence, or both. Based on a plain reading of the statute, a convening authority and an accused may bargain for an exact sentence which would apply without regard to the sentencing case.

To illustrate how this could work, we return to the Smith hypothetical. Assume that the parties reach an agreement to dismiss the false official statement in exchange for a term of confinement on the AWOL offense that is no more than six months and no less than six months. Additionally, the parties agree that a bad conduct discharge will be adjudged,105 and that no other punishment may be adjudged in the case. Would an agreement such as this raise any legal issues?

Two doctrines may have an impact on this agreement: unlawful command influence (UCI) and the doctrine requiring a “complete presentencing proceeding.”106 Each will be handled in turn.

Unlawful Command Influence

The possibility of direct limitation of the sentence sets up an interesting question about how it would be interpreted in light of the doctrine preventing adjudicative phase UCI. 107

The prevention of UCI is crucial to the military justice system.108 Article 37, UCMJ, states that “[n]o person . . . may attempt to coerce, or, by any unauthorized means, influence the action of a court-martial . . . in reaching the findings or sentence in any case.”109 The cases have interpreted this clause to prohibit improper influence during the adjudicative phase of trial, whether the influence actually occurs or occurs only in appearance.110 Apparent UCI exists when an objective, disinterested observer with knowledge of all the facts would harbor a significant doubt about the fairness of the court-martial proceedings.111 Once evidence of UCI has been raised, the burden falls on the government to prove beyond a reasonable doubt either that the predicate facts did not occur, the facts do not constitute UCI, or the facts do not prejudice the proceeding.112

It may be tempting to dismiss the application of Article 37, UCMJ, to plea agreements because the language of Article 53a explicitly allows a convening authority to directly limit the sentence of the court-martial. Thus, a strict reading appears to allow sum certain agreements such as the one in the Smith hypothetical. Indeed, one might even argue that part of the purpose of Article 53, UCMJ, was to give convening authorities more direct control over the sentencing process.

A quick dismissal of UCI concerns would be a mistake for three reasons. First, Article 37, UCMJ, is a systemic check on other articles. Under the legacy system, Article 37, UCMJ, was deployed in the plea agreement context to protect the sentencing authority from influence by a convening authority’s agreement to exercise Article 60, UCMJ, clemency powers.113 Under the legacy system, the convening authority’s exercise of power was purely retrospective, and acted upon an adjudged sentence. Yet, Article 37, UCMJ, ensured judicial independence in formulating the sentence, and insulated the judge from the future acts of the convening authority by preventing disclosure of the agreement until after the sentence was announced. While Article 53a, UCMJ, changes the role a convening authority plays in the plea agreement context, there is no reason to believe that Article 37, UCMJ, will play any less systemic role than it historically has. All of the same doctrines of fairness—and the appearance of fairness—will likely apply even more forcefully under the 2016 MJA, for the military judge will now have direct knowledge of the limitations issued by the convening authority. It is even more likely that Article 37, UCMJ, will be applied more strictly in the 2016 MJA system because the order of performance has changed—the convening authority performs in directing the court even before it has heard the evidence or deliberated on a sentence.

Second, the language of Article 53a, UCMJ, is unlikely to be interpreted by an appellate court without at least some historical analysis of the text. The original formulation proposed by the MJRG would have provided much stronger objective guidance—in the form of sentencing parameters—to a military judge in evaluating a plea agreement. The original formulation provided that “the military judge may reject [the plea agreement] only if it proposes a sentence that is both outside the sentencing parameter and plainly unreasonably.”114 This provision would have operated as a limit on the military judge, but also on convening authority power. In essence, the convening authority would know that to be accepted, the agreement would need to fall within the parameters, if any, and must not be plainly unreasonable.

Neither the “plainly unreasonable” language nor the sentencing parameters made it into the 2016 MJA; however, the rest of the statute authorizing direct limitations did. In place of that language, courts are instead to look to whether the agreement is unlawful or violates the promulgated rules.115 Appellate courts seeking to determine just what Congress intended when striking those limits, and in what way violates the law, will (and should) most certainly turn to Article 37, UCMJ, for guidance.

Finally, practitioners would do well to realize that the body of legal thought underlying plea agreements contains much disagreement over the proper role of such agreements in society. The bodies of thought can be loosely lumped into two categories: those who believe that plea agreements should be limited or abolished because they are a fundamentally unfair exercise of state power,116 and those that support plea agreements as an exercise in freedom of contract.117 While these two bodies of thought may exist at polar extremes, they are likely to inform case law evaluating Article 53a, UCMJ.

Few would disagree that United States v. Care118 and its progeny skew towards a paternalistic view of the law as regards a Soldier attempting to enter a guilty plea.119 With Article 53a, UCMJ, now providing the convening authority direct opportunity to limit the sentencing authority, provided the accused agrees, it is likely that the appellate case law will reflect a struggle between those jurists who subscribe to a freedom of contract paradigm, and those who believe the judiciary should be a check on command authority. It is likely that the strong streak of pre-existing (paternalistic) legal thought is likely to play a role in couching Article 53a, UCMJ, within the broader context of Article 37, UCMJ, limitations against UCI.

Returning to how the Smith hypothetical might be evaluated in light of UCI concerns, this deal may raise an issue of apparent unfairness. Essentially, the agreement between the convening authority and the accused have fixed the sentence in such a way that the judge is powerless to further shape it, even after receiving all the evidence at the sentencing hearing.

The court may then look at whether the facts constitute UCI and whether they prejudice the proceeding. The government will no doubt contend that the convening authority could not have committed UCI because Article 53a, UCMJ, authorizes such deals. The defense will argue, however, that while a convening authority certainly has the authority to limit the court-martial, Article 37, UCMJ, still acts as a limit on the convening authority’s exercise of that power. Secondly, the defense will argue that the agreement prejudiced the accused because it prevented the accused from benefiting from an independent military judge because the judge would have examined the limitation before having heard the sentencing evidence or argument.

Ultimately, a court is likely to evaluate the deal under the rubric of apparent UCI—what would a reasonable member of the public think about the proceeding? Courts would likely look at whether the crime was of such a magnitude, or the aggravation evidence so strong, as to demonstrate the agreement was a good one for the accused.

Of course, there are two ironies likely to limit such cases: first, the government would be unlikely to deal for a low “sum certain” cap in a strong case; and second, supposing there were such an agreement, the government might lose whatever efficiency-related benefits it hoped to achieve by a “sum certain” cap in having to justify the agreement with evidence entered into the record.120

The “Empty Ritual” Problem and “Complete Presentencing Proceedings”

A second problem with a “sum certain” type plea agreement is that it has a tendency to render the trial “an empty ritual.”121 Although the doctrine appears facially similar to UCI, this doctrine has its foundation in public policy concerns rather than Article 37, UCMJ. In United States v. Davis, the plea agreement called for the “appellant to request trial by military judge alone, enter into a confessional stipulation, [and] ‘call no witnesses and present no evidence . . . during the case on the merits.’”122 The Court of Appeals for the Armed Forces (CAAF) found that the agreement violated public policy, but found that the accused was not entitled to relief because the accused had no evidence to present anyway.123

Although the holding in Davis may appear limited only to cases involving confessional stipulations, the court noted that “[a] fundamental principle underlying this Court’s jurisprudence on pretrial agreements is that ‘the agreement cannot transform the trial into an empty ritual.’”124 The court pointed out that RCM 705 prohibits terms in a plea agreement that “deprive[] the accused of . . . the right to complete presentencing proceedings”125 derived from the basic public policy protections against an “empty ritual” announced in United States v. Allen.126

Applying this rationale to the Smith hypothetical, there are multiple issues that a “sum certain” agreement could raise. Suppose that the parties arrived at the bad conduct discharge and the six-month confinement agreement after reviewing what they concluded would be admissible in the presentencing proceedings. Suppose, for example, the government’s evidence against Private Smith included a string of failures to report,127 for which he had received punishment under Article 15, UCMJ.128 Suppose the accused had also received several counseling statements for disrespecting his superior officers.129 The rules would permit the government to introduce the circumstances of the apprehension, to include the alleged false statement, on sentencing.130 Finally, suppose the government had evidence that Private Smith’s unit deployed shortly after the accused went AWOL, and that the person who replaced Private Smith died conducting a mission that Private Smith otherwise would have performed. The government believes this would be admissible as aggravation evidence.131

Suppose the defense has evidence that the accused only left because his mother was sick and had recently lost her job. Further, they could call members of the unit to testify that he was a good Soldier, and that they would serve with him again if they had the chance. The accused got a job as a construction worker while he was AWOL, evidence which the defense believes will show the accused has strong rehabilitation potential.

Suppose that both the government and defense offer and obtain admission of all evidence discussed above. As agreed, the military judge sentences the accused to six months of confinement and a bad conduct discharge.

This scenario appears to provide a full sentencing proceeding—after all, the accused got what he had bargained for, and there was a full sentencing case presented to the judge. And yet, on the basis of Davis, the accused may nonetheless have a strong appellate argument that the sentencing proceeding was an “empty ritual.” After all, he had no reasonable expectation that the judge would consider and be able to act on the mitigation evidence he presented. Thus, he may argue that the “no less than” limit and the bad conduct discharge requirement should have been held invalid because of their tendency to render the proceedings an “empty ritual.” The defense would argue that those terms were impermissible as a matter of public policy, and ask the court to negate those limits and then either reassess the sentence, or return the case for resentencing.

Suppose instead that the parties decide not to present any evidence at sentencing.132 The parties rely on the agreement and argue to the judge that, because he has accepted the agreement, he is bound to the terms whether or not the parties present any sentencing evidence. The judge agrees that this is a correct interpretation of Article 53a, UCMJ, and sentences the accused in keeping with the agreement.

On appeal, the accused may maintain a two-pronged argument. First, he may argue that his defense counsel was ineffective because he did not submit any evidence in mitigation. He may argue that his defense counsel should have told the judge he had authority to reject the agreement and sentence according to his conscience.133 Thus, he could argue that the agreement amounted to nothing more than a six-month cap on the sentence, for if the judge had rejected the agreement, he surely would have sentenced the accused to a lesser punishment. He may argue that but for the failures of his defense counsel, he would have been sentenced to a lesser punishment.134 He could also supplement his argument with the fact that the government put on no evidence.

Second, he could argue that the presentencing hearings were an “empty ritual” and that he was denied a complete presentencing hearing by operation of the agreement. Even though there was no evidence presented during the sentencing case, he may argue that he would have presented the evidence but for the “sum certain” agreement.

Is a “sum certain” agreement practically advisable? Given the potential for appellate issues, and the attendant loss of judicial economy, practitioners should avoid the allure of such agreements. Instead, practitioners should strive to ensure that the judge has a meaningful decision to make upon sentencing. Doing so eliminates the concern that the proceeding is an “empty ritual.”

Unfortunately, just how much sentencing leeway is necessary to avoid the UCI and the “empty ritual” problems is not something that can be decided by bright line rule. Nevertheless, this author proposes that a good indicator of whether the judge has a meaningful decision is if the agreement provides the defense sufficient motive and opportunity to influence the sentence in some significant way.

The wise practitioner will think carefully before agreeing to a “sum certain” deal, or one in which the agreement is narrowly prescribed. Practitioners might instead look to barter limitations on a per specification basis (e.g., “I will agree to limit Specification 1 to no less than x years confinement, if you agree to limit specification 2 to no more than y years confinement.”). Practitioners could also broaden a facially “sum certain” agreement by agreeing to wide judicial latitude on whether the confinement will be served concurrently or consecutively. Practitioners could also agree to a “sum certain” as to one specification, but agree that another specification may be sentenced according to broader limits. Practitioners might also agree to a “sum certain” as to a particular part of the sentence (e.g., forfeitures), while leaving open another part of the sentence (e.g., confinement). In all this, the practitioner must carefully “think forward” to how these limits may play out during sentencing, and to whether there is sufficient judicial freedom to avoid raising the UCI and “empty ritual” problems.

Practitioners should avoid the temptation to justify them. No doubt such agreements would seem to enhance judicial economy, and might even appear to enhance justice in cases where it is clear the accused is getting a good deal. Moreover, practitioners might point out that a “sum certain” agreement is similar to an agreement by a civilian prosecutor to recommend a certain sentence. These arguments are apt to lead astray. Our system is a statutory one in which the convening authority holds an immense power imbalance over the accused, and potentially over the court. Indeed, even the court-martial cannot exist apart from the authorization of the convening authority.135 With that kind of power already in the hands of the convening authority, the convening authority should be hesitant to exercise too much authority over the sentence where it might appear to undermine the independence of the trial judiciary.

Heightened Military Judge Responsibilities

The 2016 MJA introduces additional implied responsibilities on military judges. Just as the opportunity for direct limitation of the court-martial introduces the opportunity for convening authority overreach, so too does it introduce the requirement for the trial judiciary to zealously guard the province of the court. The prospective nature of a plea agreement under the 2016 MJA raises two specific issues the military judge should be tracking. First, the judge has an increased obligation to serve as gatekeeper. Second, the judge has a heightened obligation to police performance of the terms of the agreement.

Gatekeeping Role of the Military Judge

The gatekeeping role of the military judge is a traditional one. Judges have a “sua sponte duty to insure [sic] that an accused receives a fair trial.”136 In the plea agreement context, judges have the duty to ensure the accused is giving a plea that is “provident”—that is, wise.137 United States v. Care imposes a judicial obligation of inquiry that arises from the concern that the accused has limited bargaining power as against the convening authority.138 Care was decided in the context of a system in which the convening authority had no direct influence over the sentence adjudged; in a new system where the convening authority has authority over the sentence ab initio, the judge has that much more obligation to police the agreement.

Articles 53a(b)(4) & (5), UCMJ, impose a judicial obligation to reject agreements that are inconsistent with the law or with the Rules for Court-Martial. While this authority affirms the gatekeeping role of the military judge, it is not helpful to a judge trying to determine when to exercise that role.

Judges should inquire into plea agreements while keeping the provisions of Article 45, UCMJ, at the forefront of their thinking. Under Article 45, UCMJ, the accused may not enter a plea “improvidently.” In modern practice, the “providence” required in Article 45, UCMJ, and the Care inquiry have become synonymous. However, in light of practice in a new system in which judges will see terms never before tested, judges will have to return to the root meaning of “improvident”—unwise—and interpret the plea agreement itself within that framework. In short, judges should be alert to terms that would make it unwise for an accused to enter a plea.

In the legacy system, the judge’s inquiry as to the terms of clemency only looked at whether the parties agreed as to the meaning of the terms. Under the 2016 MJA, judges have an obligation to look more searchingly at the sentence limitation to determine whether the accused is pleading providently. In fact, it is entirely possible that the limitations themselves could make the plea improvident—i.e., unwise. Where there are potential issues, judges should take care to get the facts on the record, and rule as to any issues explicitly.

Judges also need to be on the lookout for terms that violate public policy.139 While there is no exhaustive list of provisions that would be void for public policy reasons, a good rule of thumb is that any agreement that results in unfairness, or the appearance of unfairness in the case, or any agreement which tends to undermine the ability of the court to administer justice, will be void for public policy reasons.140

Finally, under RCM 705(b)(1), judges are reminded to reject involuntary agreements and terms that deprive the accused of certain fundamental rights to counsel, due process, issuance of a challenge to jurisdiction; a speedy trial; complete presentencing proceedings; and complete and effective exercise of post-trial and appellate rights.

To exercise their gatekeeping role, RCM 910 provides the judge authority to reject agreements. Under that rule, a judge shall “issue a statement explaining the basis for the rejection; and allow the accused to withdraw any plea; and inform the accused that if the plea is not withdrawn the court-martial may impose any lawful punishment.”141 Because much of the plea and plea agreement inquiry aims to protect the accused from himself, it is likely that a rejection of the agreement may stem from terms the judge wishes not to enforce against the accused. Upon learning the reasons for rejection, the prudent accused may decide to persist in the plea and obtain the benefit of a more favorable sentencing outcome than contemplated in the now-rejected agreement.

Heightened Obligation to Police Performance

A new role arising under the 2016 MJA is the judge’s obligation to police any post-trial performance terms. Under the legacy system, performance failures of the accused could be addressed by the convening authority at action. For example, consider an agreement where the accused agreed to “good behavior” through convening authority action. If the accused got into trouble after the court-martial, the convening authority could simply refuse to exercise the clemency contemplated in the plea agreement, and approve the full sentence of the court-martial. The same would be true of any term in which performance was to occur after announcement of the sentence.142

Under the 2016 MJA, the sentencing relief is in the agreement itself. If the accused fails to perform on a post-trial obligation, the convening authority no longer has any means to protect himself. Unlike the legacy system, there is no “fall back” sentence of the judge that was not influenced by the agreement. The sentence of the judge now reflects the terms of the agreement, rather than what the accused would have gotten without the agreement.

The Rules for Court-Martial provide for post-trial motions and proceedings.143 There are no procedures specifically addressing how a judge is to proceed when an accused has breached a term post-announcement of the sentence. Nevertheless, if the government wishes to withdraw from the agreement based on a breach, the judge will need to conduct an Article 39(a) session to permit the government to withdraw from the agreement. If the accused persists in his plea of guilty, the judge may proceed to a rehearing on the sentence, this time free from the terms of the agreement.

To sum up, the military judge’s obligation to stand as gatekeeper must rise equal to the convening authority’s heightened power over the proceedings, just as the military judge’s obligation to police an accused’s performance after announcement of sentence must provide support to the convening authority in amounts equal to his reliance risk.


The 2016 MJA made significant revisions to the military justice system. Some of the changes are obvious. What is not so obvious is how those changes reflect fundamental shifts within the system itself. Practitioners who wish to succeed must recognize the limits of their experience under the legacy system—no matter how long and distinguished—for it is not simply the rules that have changed, it is the system itself. The changes to the plea agreement system represent a significant shift away from a clemency-based model to a prescriptive system in which the convening authority has power, pursuant to a plea agreement, to direct the sentence of the court-martial. The changes will have significant ripples. Some changes are easily understood and foreseen. Other changes are hidden in plain sight, waiting for enterprising lawyers to find them and shape the law for decades to come. TAL


LTC (Ret.) Bigler retired as an associate professor in the Criminal Law Department at The Judge Advocate General’s Legal Center and School, Charlottesville, Virginia. He is currently the Director, Legal and Contracts at Cardno GS, Inc.


1. National Defense Authorization Act for Fiscal Year 2017, Pub. L. No. 114-328, 130 Stat. 2000 (2016) [hereinafter 2016 MJA].

2. 10 U.S.C. § 853a (2016). Unless otherwise specified, all versions of the statutes and accompanying rules cited in this article are those passed under the 2016 MJA. The version in effect prior to passage of the 2016 MJA will be denoted as “Legacy.”

3. See Manual for Courts-Martial, United States, R.C.M. 705(d)(4) (2016) [hereinafter 2016 MCM] (preventing convening authority after the accused “begins performance of promises contained in the agreement”). See also, U.S. Dep’t of Army, Pam. 27-9, Military Judge’s Benchbook, Chapter 2 (2014) [hereinafter Benchbook] (providing a detailed trial script). The accused might also perform by entering into a stipulation of fact or by doing some other bargained-for action. This chronology purposely oversimplifies the process to make clear the order of performance.

4. See generally, 2016 MCM, supra note 3, R.C.M. §§ 1001-11 (establishing sentencing procedures).

5. See generally, 2016 MCM, supra note 3, R.C.M. §§ 1101-14 (establishing post-trial processing procedures leading up to clemency action).

6. See, e.g., 2016 MCM, supra note 3, R.C.M. 910(f)(3) (“[T]he military judge ordinarily shall not examine any sentence limitation contained in the agreement until after the sentence of the court-martial has been announced.”); 2016 MCM, supra note 3, R.C.M. 910(h)(3) (“After sentence is announced, the military judge shall inquire into any parts of the pretrial agreement which were not previously examined by the military judge.”).

7. Compare 2016 MCM, supra note 3, R.C.M. 910(f)(3) (requiring disclosure of the entire agreement, save any limitations, before the plea is accepted) with 2016 MCM, supra note 3, R.C.M. 910(f)(4) (obligating the judge to inquire to ensure the accused understands the agreement and the parties agree to the terms).

8. Significantly, whether the convening authority was bound at all by an accused’s performance was not a foregone conclusion in the early days of pretrial agreements. See, e.g., Major Charles W. Bethany, The Guilty Plea Program 1 (Apr. 1959) (unpublished LL.M. thesis, The Judge Advocate General’s School) (on file with The Judge Advocate General’s Legal Center and School Library) (posing the question of whether a convening authority is bound by the accused’s performance).

9. 2016 MCM, supra note 3, R.C.M. 705(d)(4)(B).

10. 2016 MCM, supra note 3, R.C.M. 705(d)(4)(B).

11. See, e.g., Restatement of Contracts § 12 (1932). See also, Carlill v Carbolic Smoke Ball Company1 QB 256 (1893). In that case, the court held that a contract was created where a company offered a sum of money to anyone who contracted the flu after completing their prescribed treatment. The claimant, who had completed the treatment, thereafter got sick with the flu and claimed the money.

12. The typical law school contracts hypothetical explains the issue thus: suppose someone offered you $100 to walk across the Brooklyn Bridge. By the nature of the offer, acceptance only occurs on performance. Suppose that you are now just yards away from completing the task when the offeror reconsiders and withdraws the offer. At that point, finishing those last few yards may give you some personal satisfaction, but it will not earn you the $100.

13. See, e.g., United States v. Dean, 67 M.J. 224 (C.A.A.F. 2009). In that case, the amendment of a witness list pursuant to an agreement, along with other things, constituted beginning of performance. Id.

14. 2016 MCM, supra note 3, R.C.M. 705(d)(4)(A).

15. If there has never been a meeting of the minds, the accused and convening authority will be returned to status quo ante. See, e.g., United States v. Dunbar, 60 M.J. 748 (2004) (finding no agreement where the government and accused disagreed as to the meaning and effect of the sentence limitation, and setting aside the findings and sentence). In this case, the government would not be able to retain any of the benefits of the accused’s guilty plea. See generally 2016 MCM, supra note 3, Mil. R. Evid. 410(a)(4) (prohibiting the use of a guilty plea that “did not result in a guilty plea or [was] later withdrawn”).

16. 2016 MCM, supra note 3, R.C.M. 910(h).

17. See, e.g., UCMJ art. 45(a) (2018) (“If an accused . . . sets up a matter inconsistent with his plea . . . a plea of not guilty shall be entered in the record and the court shall proceed as though he had pleaded not guilty.”); 2016 MCM, supra note 3, R.C.M. 910(h)(1) (the military judge may, as a matter of discretion, permit an accused to withdraw a plea before announcement of the sentence). An accused may withdraw a plea after announcement of sentence in certain limited circumstances casting doubt on the guilty verdict. See, e.g., United States v. Olson, 25 M.J. 293 (C.M.A. 1987).

18. Cf. United States v. Olson, 25 M.J. 293 (C.M.A. 1987) (holding the accused could withdraw his plea after announcement of sentence where it was clear from the case that his entry of plea was not knowing and voluntary due to a reasonable mistake as to the terms of the agreement).

19. See, e.g., United States v. Bulla, 58 M.J. 715 (C.G. Ct. Crim. App. 2003) (finding that on the facts of the case, the convening authority was justified in withdrawing from the agreement and approving the sentence as adjudged where the accused violated a misconduct provision).

20. An example may be where there is an agreement to cooperate in another case.

21. An example may be where the sentence as adjudged is less than the limitation expressed in the agreement.

22. Bulla, 58 M.J. at 715.

23. See, e.g., United States v. Tester, 59 M.J. 644 (A. Ct. Crim. App. 2003) (convening authority may decline to exercise clemency where accused violated post trial conduct provisions and the convening authority went through appropriate procedures).

24. United States v. Dawson, 51 M.J. 411 (C.A.A.F. 1999). Dawson relies heavily on the fact that the agreement concerned matters outside the control of the court-martial—namely the decision as to the disposition of new offenses, and the decision whether to conduct vacation proceedings. Dawson cites a “tunnel of power” metaphor proposed by Judge Sullivan in United States v. Boudreax, 35 M.J. 291 (CMA 1992) for the proposition that the authority of a convening authority over certain post-trial actions (in this case, vacation proceedings) fall within the “portion of [the tunnel] under the control of the command structure,” and that therefore a court-martial is not the deciding or reviewing authority. United States v. Dawson, 51 M.J. 411, 413 (C.A.A.F. 1999).

25. United States v. Parker, 62 M.J. 459 (C.A.A.F. 2006) (citing United States v. Pilkington, 51 M.J. 415, 416 (C.A.A.F. 1999)).

26. UCMJ, art. 37(a) (2016).

27. Court Reporter’s Chronology in the case of United States v. Lieutenant Colonel James H. Wilkerson, Information Access Policy & Compliance Branch, https://www.foia.af.mil/Portals/22/documents/Library/Investigations/Wilkerson-Case/AFD-130403-023.pdf?ver=2016-09-13-105133-637 (last visited Oct. 3, 2019).

28. See e.g., Robert Draper, The Military’s Rough Justice on Sexual Assault, N.Y. Times (Nov. 26, 2014).

29. See National Defense Authorization Act for Fiscal Year 2014, Pub. L. No. 113-66, § 1702, 127 Stat. 672, 954 (2013) (restricting power to grant clemency over certain offenses).

30. Military Justice Review Group, Report of the Military Justice Review Group, Part I: UCMJ Recommendations 486 (2015), http://ogc.osd.mil/images/report_part1.pdf [hereinafter MJRG Recommendations].

31. UCMJ, arts. 60a(a)-(b) (2018). Even then, the convening authority has limited clemency authority over sentences of confinement, discharge, or death.

32. UCMJ, art. 53a(a)(1)(B) (2018).

33. Manual for Courts-Martial, United States, R.C.M. 705(d)(1) (2019) [hereinafter 2019 MCM].

34. See generally, 2019 MCM, supra note 33, R.C.M. 910(f).

35. 2019 MCM, supra note 33, R.C.M. 910(f)(3) (“[T]he military judge shall require disclosure of the entire agreement before the plea is accepted.”).

36. 2019 MCM, supra note 33, R.C.M. 705(f).

37. 2019 MCM, supra note 33, R.C.M. 1009(d)(6).

38. Rule for Court-Martial 910(f)(5) does not specifically require acceptance of the plea prior to acceptance of the agreement; however, that should be understood from the context. 2019 MCM, supra note 33, R.C.M. 910(f)(5). Rule for Court-Martial 910(f)(5) provides that the “military judge shall announce on the record whether the plea and the plea agreement are accepted” after the plea agreement inquiry is complete. Id. The agreement inquiry requires only that the military judge has ensured “the accused understands the agreement and that the parties agree to the terms of the agreement.” 2019 MCM, supra note 33, R.C.M. 910(f)(4)(A). However, the acceptance of the plea agreement generally follows the factual basis inquiry. See 2019 MCM, supra note 33, R.C.M. 910(e); U.S. Dep’t of Army, Pam. 27-9, Military Judge’s Benchbook (1 Sept. 2014) (Unofficial Version 18.0, dtd 14 Feb. 2018), at 14-29. While the order of operations was not crucial under the legacy system, order is highly important under the 2016 MJA. That is because successful entry of the plea is required to give the agreement binding effect upon the military judge’s acceptance. Interpreting this provision otherwise would have the absurd result of binding the parties to perform under the explicit language of Article 53(d), UCMJ, even before the accused has admitted his guilt.

39. UCMJ, art. 53a(d) (2018); 2019 MCM, supra note 33, R.C.M. 1002(a)(2) (“If the military judge accepts a plea agreement with a sentence limitation, the court-martial shall sentence the accused in accordance with the limits established by the plea agreement.”).

40. UCMJ, art. 53a(b)(1) (2018).

41. UCMJ, art. 53a(b)(2) (2018).

42. 2019 MCM, supra note 33, R.C.M. 901(f)(4)(B).

43. UCMJ, art. 53a(a)(2) (2018).

44. UCMJ, art. 53a(b)(4) (2018).

45. UCMJ, art. 53a(b)(5) (2018). The provisions of Article 53(a)(b)(4) and (5) were adopted in Section 531(d), National Defense Authorization Act for Fiscal Year 2018. National Defense Authorization Act for Fiscal Year 2018, Pub. L. No.115-91, § 531(d), 131 Stat. 1283, 1384 (2017). While these amendments do not provide explicit interpretive guidance on what is “prohibited by law” or “inconsistent” with regulations, how it might be applied will be discussed later, infra Heightened Military Judge Responsibilities.

46. UCMJ, art. 53a(b)(3)-(c) (2018).

47. UCMJ, art. 56(b)(2)(A) (2018).

48. UCMJ, art. 56(b)(2)(B) (2018).

49. UCMJ, art. 56(b)(2)(C) (2018).

50. UCMJ, art. 56(b)(2)(D) (2018).

51. UCMJ, art. 56(b)(2)(E) (2018).

52. UCMJ, art. 56(b)(2)(F) (2018).

53. UCMJ, art. 56(b)(1) (2018). With the exception of conspiracy, only a general court-martial has jurisdiction to hear the listed offenses. UCMJ, art. 18(c) (2018). While not a focus of this article, the astute practitioner will recognize the apparent oversight and advise the convening authority to refer the case to a general court-martial where the charges warrant a court-martial. See UCMJ, art. 34 (2018); 2019 MCM, supra note 33, R.C.M. 601(d).

54. UCMJ, art. 53a(c)(1) (2018).

55. UCMJ, art. 53a(c)(2) (2018).

56. Both the legacy and the 2016 MJA systems use the same language of “limitation.” Compare, e.g., 2016 MCM, supra note 3, R.C.M. 910(f)(3) (preventing the military judge from reviewing any “sentence limitation”) with UCMJ, art. 53a(a)(1)(B) (2018) (discussing a “limitation on the sentence” as an appropriate subject of a plea agreement). However, a closer reading of the legacy rule makes it clear that the convening authority is not executing a direct limitation on the sentence. See UCMJ, art. 60(c)(4)(C) (2016) (discussing the convening authority’s powers to grant clemency if called for in a pretrial agreement).

57. 2019 MCM, supra note 33, R.C.M. 705(d).

58. The legacy rules provided a straightforward sentencing procedure that does not change regardless whether the accused elects sentencing by a military judge or by a panel. Under that system, the court-martial sentenced the accused to a unitary, all-in punishment containing the term of confinement, punitive discharge, forfeitures, or other parts of the sentence. 2016 MCM, supra note 3, R.C.M. 1002(b).

59. 2019 MCM, supra note 33, R.C.M. 1002(d)(2).

60. Id.

61. See 2019 MCM, supra note 33, R.C.M. 1002(f), which provides a list of factors to ensure that the “court-martial shall impose a sentence that is sufficient, but not greater that necessary, to promote justice and to maintain good order and discipline in the armed forces.”

62. 2019 MCM, supra note 33, R.C.M. 1002(d)(2)(C).

63. 2019 MCM, supra note 33, R.C.M. 1002(d)(1).

64. UCMJ, art. 53(b)(1)(A) (2018) (“Except as provided in subparagraph (B), . . . if the accused is convicted of an offense in a trial the military judge shall sentence the accused.”).

65. See UCMJ, art. 53(b)(1)(B) (2018) (“If the accused is convicted of an offense by general or special court-martial consisting of a military judge and members and the accused elects sentencing by members under section 825 of this title (article 25), the members shall sentence the accused.”); UCMJ, art. 25(d)(1) (2018) (“Except [in capital cases], the accused in a court-martial with a military judge and members may, after the findings are announced and before any matter is presented in the sentencing phase, request, orally on the record or in writing, sentencing by members.”) (emphasis added).

66. 2019 MCM, supra note 33, R.C.M. 1002(b)(1).

67. 2019 MCM, supra note 33, R.C.M. 1002(d)(2).

68. Benchbook, supra note 3 (Guilty Plea Inquiry under “Scripts”).

69. See 2019 MCM, supra note 33, R.C.M. 911 Discussion (“Assembly of the court-martial is significant because it marks the point after which: substitution of the members and military judge may no longer take place without good cause (see UCMJ art. 29, R.C.M. 505, 902, 912); the accused may no longer, as a matter of right, request trial by military judge alone or withdraw such a request previously approved (see Article 16, R.C.M. 903(d)); and the accused may no longer request, even with the permission of the military judge, or withdraw from a request for members (see Article 25(c)(2), R.C.M. 903(d)).”) (citations in original).

70. See UCMJ, art. 22 (2018) (granting general court-martial convening authority); UCMJ, art. 23 (2018) (granting special court-martial convening authority); UCMJ, art. 24 (2018) (granting summary court-martial convening authority).

71. The accused may always request trial by judge alone in lieu of trial before the members. See UCMJ, art. 16(b)(3) (2018); UCMJ, art. 16(1)(B) (2016).

72. See, e.g., UCMJ art. 16(1)(A) (2016) (defining a general court-martial as composed of a military judge and at least five members); UCMJ art. 16(2)(A)-(B) (2016) (defining a special court-martial as composed of at least three members or of a military judge and at least three members); UCMJ art. 16(b)(1) (2018) (defining a general court-martial as consisting of a military judge and eight members); UCMJ, art. 16(c)(1) (2018) (defining a special court-martial as consisting of a military judge and four members).

73. Legacy Article 25(a) provided that any commissioned officer is eligible to serve on a court-martial if the convening authority has selected the commissioned officer to serve as a member. See generally UCMJ, art. 16 (2016). Under Legacy Article 25(d)(1), an enlisted accused may elect enlist members, provided that he has done so prior to assembly of the court martial. UCMJ, art. 25(d)(1) (2016). Additionally, if the accused has elected trial by a military judge before the court-martial has assembled, then the military judge may hear the case if the request is approved. See 2016 MCM, supra note 3, R.C.M. 903(c)(2).

74. Under the 2016 MJA Article 25(c)(2), the accused may elect trial before a court-martial composed entirely of officers or composed of at least one-third enlisted members. UCMJ, art. 25(c)(2) (2018). If the accused makes no election, the default composition of the court-martial will be the members selected by the convening authority under Article 16, UCMJ. Id.

75. See MCM, supra note 33, R.C.M. 903(a)(2) Discussion.

76. Compare UCMJ, art. 39(a) (2018) (“[T]he military judge may . . . call the court into session without the presence of the members [in limited enumerated circumstances].”), with UCMJ, art. 39(a) (2016) (same).

77. Compare UCMJ, art. 39(a)(3) (2016) (permitting a military judge to hold an arraignment and receive the pleas of the accused); UCMJ, art. 39(b) (2016) (permitting a military judge authority to perform other procedural functions authorized by the President under Article 36 [the Rules for Courts-Martial]); 2016 MCM, supra note 3, R.C.M. 910(f) (establishing procedures for the military judge to receive the pleas); 2016 MCM, supra note 3, R.C.M. 910(g) (providing authority for the military judge to enter a finding of guilty on the basis of an accepted plea) with UCMJ, art. 39(a)(3) (2016) (providing a military judge authority to hold arraignment and receive the pleas of the accused); UCMJ, art. 39(a)(5) (2016) (providing a military judge authority to perform other procedural functions authorized by the President under Article 36 [the Rules for Courts-Martial]); 2019 MCM, supra note 33, R.C.M. 910(f) (establishing procedures for the military judge to receive pleas).

78. Compare 2016 MCM, supra note 3, R.C.M. 910(g) (providing authority for the military judge to enter a finding of guilty on the basis of an accepted guilty plea) with 2019 MCM, supra note 33, R.C.M. 910(g) (providing authority for the military judge to enter a finding of guilty immediately at an Article 39(a) session on the basis of a guilty plea).

79. UCMJ, art. 39(a)(4) (2018) (providing authority to “conduct[] a sentencing proceeding and sentenc[e] the accused in non-capital cases”). Not only does the provision contemplate the possibility of having a “court-martial” that never actually assembles, but in contested member cases, it also contemplates releasing the assembled court-martial before sentencing. Id. Under the legacy system, the release of the assembled members only occurred at adjournment.

80. 2019 MCM, supra note 33, R.C.M. 910(c) (requiring that the military judge inform the accused, among other things, of “the right to be tried by a court-martial,” and that “if the accused pleads guilty, there will not be a trial of any kind as to the offenses to which the accused has so pleaded”).

81. MJRG Recommendations, supra note 30, at 477-78.

82. 2019 MCM, supra note 33, R.C.M. 902A.

83. 2019 MCM, supra note 33, R.C.M. 902A(a).

84. Note that the version of Article 60, UCMJ, in use depends on the dates of the earliest offense for which the accused was sentenced. Exec. Order No. 13,825 § 6(b), 83 Fed. Reg. 9,889 (Mar. 1, 2018). Thus, where there are multiple offenses crossing the effective date, the astute reader will recognize that clemency powers remain available to the convening authority and may serve as a foundation for a plea agreement.

85. Article 53(b)(1)(A) provides a default rule that the military judge “shall sentence the accused.” UCMJ, art. 53(b)(1)(A) (2018). Under Article 53(b)(1)(B), the accused may elect sentencing by members where “the accused is convicted of an offense in a trial . . . by members.” UCMJ, art. 53(b)(1)(B) (2018).

86. 2019 MCM, supra note 33, R.C.M. 1002(d)(2)(A).

87. 2019 MCM, supra note 33, R.C.M. 1002(d)(2)(B).

88. The 2016 MJA provides new sentencing factors for the purpose of “imposing punishment that is sufficient, but not greater than necessary, to promote justice and to maintain good order and discipline in the armed forces,” UCMJ, art. 56(c) (2018); 2019 MCM, supra note 33, R.C.M. 1002(f). While “reverse engineering” is not one of the factors, the Rules for Court-Martial at least provides some authority for a judge to consider the plea agreement when determining whether confinement terms will run concurrently or consecutively. 2019 MCM, supra note 33, R.C.M. 1002(d)(2)(B). The problem raised in this scenario is that a judge would be determining confinement using an agreement that is facially collateral to at least one of the offenses. See 2019 MCM, supra note 33, R.C.M. 705(d)(2) (providing that a plea agreement with limitations on confinement “shall include separate limitations, as applicable, for each charge or specification”).

89. 2019 MCM, supra note 33, R.C.M. 910(f)(4)(B).

90. UCMJ, art. 53(b)(1)(B) (2018).

91. 2019 MCM, supra note 33, R.C.M. 1002(d)(1).

92. UCMJ, art. 56(b) (2018).

93. Article 53a(c)(1) provides that a “military judge may accept a plea agreement that provides for a sentence of bad conduct discharge.” UCMJ, art. 53a(c)(1) (2018). Additionally, pursuant to an agreement, the trial counsel can recommend less than the mandatory minimum in exchange for “substantial assistance” by the accused. UCMJ, art. 53a(c)(2) (2018).

94. For example, the agreement could provide that the trial counsel will recommend the judge not apply the mandatory minimum. In a different case, the agreement could include a stronger term specifying not only that the trial counsel will make the recommendation, but that the judge will sentence free of the mandatory minimum.

95. An example of such a term might provide: “No term in this agreement shall be interpreted as having any effect on a mandatory minimum punishment.” The agreement could also include a severability clause.

96. 2019 MCM, supra note 33, R.C.M. 1003(b)(1). Practitioners should understand that a term like this would also enable a judge to sentence an accused to more onerous punishments, like hard labor without confinement. See 2019 MCM, supra note 33, R.C.M. 1003(b)(6) (permitting sentence to confinement and hard labor without confinement in the same case, under certain circumstances). A more defense-friendly version of a blanket provision might state, “The military judge may adjudge no punishment except as expressly provided in this agreement.”

97. Cf, United States v. Lundy, 60 M.J. 52 (C.A.A.F. 2004) (the accused arguing that the punishments triggering suspension by operation of law must also be suspended, and the court finding that the accused was entitled to specific performance).

98. The offenses listed are: violation of Article 120(a) or (b), violation of Article 120b, or an attempt to commit such an offense. These offenses are rape or sex assault, where penetration has occurred.

99. In this case, an accused would be well served to argue for specific performance.

100. UCMJ, art. 56(b)(2) (2018).

101. UCMJ, art. 18(c) (2018).

102. UCMJ, art. 53a(c) (2018).

103. 2019 MCM, supra note 33, R.C.M. 902A(a).

104. Remember that under the old rules, an “all in” punishment term would apply regardless of forum election. However, under the new rules, the default is for sentencing by a military judge. The agreement would thus need to cover limitations for each and every specification if the accused elected the new rules.

105. This term would be unlawful under the legacy system. See, e.g., United States v. Libecap, 57 M.J. 611 (C.G. Ct.Crim. App. 2002) (agreement to request a bad conduct discharge was void as against public policy). It would also be a bad idea under the 2016 MJA. See infra The “Empty Ritual” Problem and “Complete Presentencing Proceedings.”

106. 2019 MCM, supra note 33, R.C.M. 705(c)(1)(B).

107. Several doctrines have arisen to help courts handle instances of unlawful command influence. The first doctrinal division concerns the phase of the case during which the impacts of the unlawful influence is felt. The two phases are known as the accusatory phase and the adjudicative phase. A classic example of accusatory phase unlawful command influence is when a superior influences a subordinate to forward a case that they may not have otherwise forwarded. The second doctrinal division concerns the type of harm done. Actual unlawful command influence is influence that results in an action violating the provisions of Article 37, UCMJ. UCMJ, art. 37 (2018). Apparent unlawful command influence concerns itself with whether a disinterested member of the public, knowing all the facts, would lose confidence in the justice system. United States v. Boyce, 76 M.J. 242 (C.A.A.F. 2017).

108. Unlawful command influence is a “mortal enemy of military justice.” United States v. Thomas, 22 M.J. 388, 393 (C.M.A. 1986).

109. UCMJ, art. 37 (2018).

110. See, e.g., United States v. Boyce, 76 M.J. 242 (C.A.A.F. 2017); see also Lieutenant Colonel John L. Kiel, Jr., They Came in Like a Wrecking Ball: Recent Trends at CAAF in Dealing With Apparent UCI, Army Law., Jan. 2018, at 19.

111. Boyce, 76 M.J. at 249.

112. United States v. Biagase, 50 M.J. 143, 150 (C.A.A.F. 1999); United States v. Simpson, 58 M.J. 368, 373–78 (C.A.A.F. 2003).

113. MJRG Recommendations, supra note 30, at 483 (noting that the “confluence . . . of Articles” 60 and 37 explain much of the legacy pretrial agreement system).

114. Id. at 487.

115. UCMJ, arts. 53a(b)(4)&(5) (2018).

116. See, e.g., Albert W. Alschuler, Straining at Gnats and Swallowing Camels: The Selective Morality of Professor Bibas, 88 Cornell L. Rev. 1412, 1414 (2003) (criticizing plea bargaining as “marvelously designed to secure conviction of the innocent”); Steven P. Grossman, An Honest Approach to Plea Bargaining, 29 Am. J. Trial. Adv. 101, 103 (2005) (arguing that “differential sentencing between defendants who plead guilty and those who go to trial is, in large part, punishment for exercising the right to trial.”); Ralph Adam Fine, Plea Bargaining: An Unnecessary Evil, 70 Marq. L. Rev. 615 (1987) (arguing to abolish plea bargaining).

117. See, e.g., Robert E. Scott & William J. Stuntz, Plea Bargaining as Contract, 101 Yale L.J. 1909, 1910 (1992) (arguing that plea agreements are “paradigmatic bargains of the sort we routinely enforce in other contexts”).

118. United States v. Care, 40 C.M.R. 247 (1969).

119. See, e.g., United States v. Parker, 10 M.J. 849, 851 (N.C.M.R. 1981) (likening the Care inquiry to “running the gauntlet”).

120. While the government might lessen this burden by entering into a stipulation of fact with the defense, the defense may not wish to stipulate to aggravating facts because they could harm the accused should the sentence be reassessed on appeal.

121. United States v. Davis, 50 M.J. 426 (C.A.A.F. 1999).

122. Id. at 427.

123. Id. See also United States v. Bertelson, 3 M.J. 314 (C.M.A. 1977) (holding that a confessional stipulation is acceptable where there is no agreement not to raise defenses or motions).

124. United States v. Davis, 50 M.J. 426 (C.A.A.F. 1999) (citing United States v. Allen, 8 U.S.C.M.A. 504, 507, 25 C.M.R. 8, 11 (1957)).

125. 2019 MCM, supra note 33, R.C.M. 705(c)(1)(B).

126. United States v. Allen, 8 U.S.C.M.A. 504 (C.M.A. 1957).

127. Essentially, the accused was late to work. This is a violation of Article 86, UCMJ. UCMJ, art. 86 (2018).

128. Admissible under R.C.M. 1001(b)(2). 2019 MCM, supra note 33, R.C.M. 1001(b)(2).

129. These statements would not be admissible, except perhaps to rebut something the defense put on. See 2019 MCM, supra note 33, R.C.M. 1001(b).

130. 2019 MCM, supra note 33, R.C.M. 1001(b)(4).

131. See 2019 MCM, supra note 33, R.C.M. 1001(b)(4).

132. Of these two scenarios, this seems the most likely to occur, given that it would yield the most efficient trial.

133. See UCMJ, arts. 53a(b)(4)-(5) (2018); 2019 MCM, supra note 33, R.C.M. 910(f)(7) (establishing rules for rejecting a plea agreement).

134. See United States v. Harpole, 77 M.J. 231 (C.A.A.F. 2017) (test for ineffective assistance of counsel requires deficient counsel performance and resulting prejudice).

135. See UCMJ, arts. 16, 25 (2018).

136. United States v. Watt, 50 M.J. 102, 105 (C.A.A.F. 1999).

137. See UCMJ, art. 45 (2018).

138. See generally, United States v. Care, 40 C.M.R. 247 (1969).

139. See, e.g., United States v. Edwards, 58 M.J. 49 (C.A.A.F. 2003) (“To the extent that a term in a pretrial agreement violates public policy, it will be stricken from the pretrial agreement and not enforced.”).

140. See, e.g., United States v. Weasler, 43 M.J. 15 (C.A.A.F. 1995) (knowing and intelligent waiver of accusatory phase UCI not against public policy); United States v. Smith, 44 M.J. 720 (A. Ct. Crim. App. 1996) (contingent sentences where accused would have to pay $100,000 fine or else face an additional fifty years of confinement was void for public policy); United States v. McFadyen, 51 M.J. 289 (C.A.A.F. 1999) (waiver of motion to obtain sentencing credit for unlawful pretrial punishment does not violate public policy); United States v. Edwards, 58 M.J. 49 (C.A.A.F. 2003) (agreement not to raise, during the accused’s unsworn statement, potential unconstitutional conduct by investigating law enforcement officers did not violate public policy); United States v. Burnell, 40 M.J. 175 (C.M.A. 1994) (agreement to waive members did not violate public policy); United States v. Libecap, 57 M.J. 611 (C.G. Ct. Crim. App. 2002) (agreement to request a bad conduct discharge was void as against public policy); United States v. Thomas, 60 M.J. 521 (N-M. Ct. Crim. App. 2004) (provision preventing the accused from accepting clemency violated public policy); United States v. Sunzeri, 59 M.J. 758 (N-M. Ct. Crim. App. 2004) (provision preventing testimony of witnesses violated public policy); United States v. Cassity, 36 M.J. 759 (N-M. Ct. Crim. App. 1992) (provision encouraging counterintuitive sentencing argument violated public policy); United States v. Forrester, 48 M.J. 1 (C.A.A.F. 1998) (agreement not to waive defenses did not violate public policy where it was not overly broad and where there were no defenses raised during providence or sentencing).

141. 2019 MCM, supra note 33, R.C.M. 910(f)(7).

142. Another common example would be an agreement to cooperate in the prosecution of a second accused.

143. 2019 MCM, supra note 33, R.C.M. 1104.