¶1           We granted certiorari to consider whether the trial court violated Michael Johnson’s statutory and constitutional rights by increasing his sentence on the remaining conviction on remand following his successful appeal on his other convictions. The trial court initially accepted, but at sentencing rejected, a plea agreement in which Johnson would plead guilty to an added count of possession of a controlled substance and the original counts would be dismissed. Due to the trial court’s rejection of the plea agreement, the case proceeded to trial where the jury found Johnson guilty of first-degree kidnapping, sexual assault, and possession of a controlled substance. The trial court subsequently sentenced him to an aggregate sentence of twenty years to life. Johnson appealed the sexual assault and kidnapping convictions on double jeopardy grounds. The court of appeals reversed and vacated those two convictions and sentences, remanding the case with instructions that the trial court reinstate the plea agreement and resentence Johnson on the remaining possession of a controlled substance conviction. On remand, the trial court imposed a twelve-year sentence for the possession of a controlled substance conviction—twice the length of the original six-year sentence for that conviction. Johnson appealed a second time, arguing that the increased sentence on the possession of a controlled substance conviction violated section 18-1-409, C.R.S. (2015), and his due process rights. The court of appeals rejected the statutory claim but then applied the presumption of vindictiveness to the trial court’s resentencing and held that the sentence violated Johnson’s due process rights. People v. Johnson, No. 09CA2203, slip op. at 10–11, 17 (Colo. App. May 31, 2012) (not selected for official publication) (modified opinion on denial of rehearing).

¶3           In 2003, the victim’s brother stole roughly one thousand dollars from Johnson. After the victim’s brother fled, Johnson kidnapped the victim. He told her that she would have to prostitute herself in order to earn back the money that her brother stole. Over the course of the next two days, Johnson took the victim to various locations and forced her to have sex with him and numerous other people. During that same time period, Johnson smoked crack cocaine and coerced the victim into smoking crack cocaine as well. While Johnson was transporting the victim to a hotel room, the police pulled over Johnson’s van and discovered the license plates were stolen. Johnson fled the scene, but the victim reported what he had done. The police found and arrested him nine days later. The People charged him with first-degree kidnapping, second-degree kidnapping, and sexual assault.

¶4           Before trial, the parties reached a plea agreement whereby Johnson would plead guilty to an added charge of possession of a controlled substance, and the People would dismiss the other three charges. The trial court initially accepted the plea agreement, but at sentencing the parties disputed whether the plea agreement included allowing the court to consider the facts underlying the dismissed charges when sentencing Johnson on the possession of a controlled substance count. As a result of this disagreement, the trial court rejected the plea agreement and set the case for trial on all counts, including the possession of a controlled substance charge. At trial, the jury found Johnson guilty of first-degree kidnapping, sexual assault, and possession of a controlled substance. The trial court thereafter sentenced Johnson to twenty years for first-degree kidnapping, twenty years to life for sexual assault, and six years for possession of a controlled substance, all running concurrently. As such, the aggregate sentence was twenty years to life.

¶5           Johnson appealed, and the court of appeals reversed the trial court, holding that the trial court violated Johnson’s double jeopardy rights when it first accepted and then at sentencing rejected the plea agreement, causing Johnson to face trial on the originally dismissed kidnapping and sexual assault counts.¹ Accordingly, the court of appeals remanded the case to the trial court with instructions to reinstate the plea agreement and sentence Johnson on only the possession of a controlled substance count.

¶6           At the sentencing hearing on remand, the trial court considered the factual basis for the dismissed sexual assault charge in its evaluation, found discretionary aggravation, and sentenced Johnson to twelve years for the possession of a controlled substance conviction—twice the six-year sentence originally imposed for that conviction and equal to the maximum aggravated sentence for a class 4 felony. § 18-1.3-401(1)(a)(V)(A), (6), C.R.S. (2015). In altering the original sentence, the trial court noted that when it originally sentenced Johnson on the possession of a controlled substance count, it did not take into account the factual basis for the sexual assault charge because, at that time, the court was separately sentencing Johnson on that charge. The court continued that it was “looking at a totally different situation than [it] was at the time when [Johnson] was facing a possibility of life in prison” during the initial sentencing. In addition, the trial court stated that in considering the factual basis for the sexual assault charge, it was relying not upon a bare conviction or statements from counsel about Johnson’s acts, but rather on the actual evidence presented in the original trial.

¶7           Johnson appealed a second time, arguing that the new sentence imposed on remand violated section 18-1-409(3), which prohibits increases of sentences successfully appealed under that statute except where “matters of aggravation” come to light after the original sentencing. Johnson, slip op. at 6. In addition, he claimed the increased sentence violated his due process rights. Id. at 14. The court of appeals concluded that section 18-1-409(3) did not apply to Johnson’s appeal since he did not originally appeal the propriety of his possession of a controlled substance sentence. Id. at 10–11. The court of appeals went on to hold, however, that the new sentence violated Johnson’s due process rights because the trial court increased his sentence for the possession of a controlled substance count, and there was a reasonable likelihood that actual vindictiveness influenced that increase. Id. at 16–17. The court applied a presumption of vindictiveness to the increased sentence for the possession of a controlled substance conviction, reasoning that no new information justifying the increase had come to light between the first and second sentencings. Id. at 14, 17. Accordingly, the court of appeals vacated Johnson’s sentence and remanded the case to the trial court for it to impose a sentence of no more than six years on the possession of a controlled substance conviction. Id. at 20.

¶8           We granted certiorari to consider whether the court of appeals erred in holding that section 18-1-409 does not apply to Johnson’s case and that the presumption of vindictiveness does apply to the trial court’s resentencing of Johnson’s possession of a controlled substance conviction.

II. Standard of Review

¶9           This court reviews issues of statutory construction, as well as constitutional challenges to sentencing determinations, de novo. Shaw v. 17 W. Mill St., LLC, 2013 CO 37, ¶ 13, 307 P.3d 1046, 1049; Lopez v. People, 113 P.3d 713, 720 (Colo. 2005).

III. Analysis

¶10           We first address whether the court of appeals correctly rejected Johnson’s argument that the trial court violated section 18-1-409(3) when it sentenced him to twelve years on the possession of a controlled substance count following his first appeal. We then consider whether the presumption of vindictiveness applies to the trial court’s imposition of the twelve-year sentence on the possession of a controlled substance count.

A. Section 18-1-409

¶11           Johnson argues that, when the trial court resentenced him on the possession of a controlled substance conviction to twelve years as opposed to the original six-year sentence for that count, it violated section 18-1-409(3), which he contends precludes the trial court from increasing his possession of a controlled substance sentence after his first appeal unless it learns of new matters of aggravation after imposing the original sentence. Whether section 18-1-409 applies turns on whether Johnson’s current appeal of his possession of a controlled substance sentence falls within the statute’s purview. When interpreting a statute, we seek “to ascertain and give effect to the purpose and intent of the General Assembly in enacting it.” Shaw, ¶ 13, 307 P.3d at 1049. To do so, we first look to the plain meaning of the statutory language. Id. When ascertaining the plain meaning of a statute, “we read words and phrases in context and construe them according to their common usage.” Bostelman v. People, 162 P.3d 686, 690 (Colo. 2007).

¶12           Section 18-1-409 creates a statutory right for defendants to appeal their sentence on certain grounds, and is part of a statutory scheme to afford defendants “minimum standards of criminal justice.” § 18-1-401. Section 18-1-409(1) states that defendants who receive a felony conviction have the right to “one appellate review of the propriety of the sentence, having regard to the nature of the offense, the character of the offender, and the public interest, and the manner in which the sentence was imposed, including the sufficiency and accuracy of the information on which it was based . . . .” See also  People v. Malacara, 606 P.2d 1300, 1302–03 (Colo. 1980) (summarizing the “propriety” of a sentence as the “intrinsic fairness or appropriateness of the sentence”). When a defendant appeals a sentence under the statute, the appellate reviewing court may “affirm the sentence under review, substitute for the sentence under review any penalty that was open to the sentencing court other than granting probation or other conditional release, or remand the case for any further proceedings . . . and for resentencing on the basis of such further proceedings.” § 18-1-409(3). If the appellate court conducting propriety review of a sentence under section 18-1-409 remands the case for resentencing, then the sentencing court may not impose a new sentence in excess of the original sentence “unless matters of aggravation in addition to those known to the court at the time of the original sentence are brought to the attention of the court during the hearing conducted under this section.” Id.

¶14           Hence, section 18-1-409(3) does not apply to Johnson’s case. Johnson originally appealed his convictions for kidnapping and sexual assault, arguing that the trial court violated his constitutional double jeopardy rights when it accepted and then later rejected the plea agreement that would have dismissed those charges. The court of appeals agreed with Johnson and directed the trial court on remand to dismiss those two charges and resentence Johnson on the possession of a controlled substance count pursuant to the plea agreement. Significantly, Johnson did not appeal the possession of a controlled substance conviction or sentence at all. We thus conclude that, because Johnson originally appealed only the kidnapping and sexual assault charges, both of which were vacated on double jeopardy grounds, section 18-1-409 does not apply here to his appeal of the possession of a controlled substance resentencing. Accordingly, we affirm the court of appeals on this issue.

¶15           Having resolved Johnson’s statutory argument, we next consider whether the trial court violated Johnson’s due process rights by vindictively sentencing him on the possession of a controlled substance count. We do so by determining first whether the presumption of vindictiveness applies, and second whether Johnson proved that the trial court acted out of actual vindictiveness.

B. The Presumption of Vindictiveness

¶16           To determine whether the presumption of vindictiveness applies in the context of Johnson’s case, we first examine a defendant’s right to appeal a criminal conviction and the protections courts provide to safeguard that right, including the presumption of vindictiveness. We also acknowledge a trial court’s need for broad discretion in sentencing to ensure a sentence is appropriate under the circumstances of a given case. We then consider the presumption of vindictiveness directly and examine how courts limit the applicability of the presumption to situations where the subsequent sentence is actually more severe than the original sentence, thereby striking a balance between protecting a defendant’s right to appeal and preserving a trial court’s sentencing discretion. We then assess three approaches other jurisdictions employ when determining what constitutes a more severe sentence following a defendant’s successful appeal—the “aggregate” approach, the “remainder aggregate” approach, and the “count-by-count” approach—and evaluate which approach best effectuates the presumption’s goals. Ultimately, we determine that the “aggregate” approach best meets those goals. After adopting the aggregate approach, we use it in Johnson’s case and conclude that the presumption of vindictiveness does not apply to the trial court’s decision to resentence him to twelve years for his possession of a controlled substance conviction, where originally he received six years for that count as part of an aggregate twenty-years-to-life sentence. We also conclude that Johnson failed to prove that the trial court acted out of actual vindictiveness.

1. The Right to Appeal

¶17           Within the judicial system, the right to appeal a criminal conviction is a fundamental and indispensable mechanism to protect a defendant’s rights and correct reversible trial court error. Wend v. People, 235 P.3d 1089, 1094 (Colo. 2010); see also § 16-12-101, C.R.S. (2015) (granting convicted persons the statutory right to appeal the proceedings resulting in conviction). To prevent infringement of a defendant’s right to appeal, the United States Supreme Court has consistently held that “[a] defendant’s exercise of a right of appeal must be free and unfettered” and that courts may not “put a price on an appeal.” North Carolina v. Pearce, 395 U.S. 711, 724 (1969) (quoting Worcester v. Comm’r of Internal Revenue, 370 F.2d 713, 718 (1st Cir. 1966)) (internal quotation marks omitted), overruled in part by Alabama v. Smith, 490 U.S. 794 (1989).

¶18           Accordingly, “[d]ue process of law . . . requires that vindictiveness against a defendant for having successfully attacked his first conviction must play no part in the sentence he receives after a new trial.” Id. at 725; see also U.S. Const. amend XIV, § 1. If a defendant wins on appeal and the trial court subsequently imposes a more severe sentence for a conviction on remand, the increased sentence raises concerns that the trial judge may have sought to punish the defendant for appealing the original decision. To deter this possibility, the Supreme Court has stated that, where there is a “reasonable likelihood” that an increase of a sentence is the product of actual vindictiveness on the part of the sentencing judge, a rebuttable presumption that the increase did in fact result from actual vindictiveness arises. Smith, 490 U.S. at 798–800 (quoting United States v. Goodwin, 457 U.S. 368, 373 (1982)); People v. Montgomery, 737 P.2d 413, 416 (Colo. 1987).

¶19           As the Supreme Court has indicated, the reason courts use the presumption where there is a reasonable likelihood of actual vindictiveness in the sentencing decision is to deter actual vindictiveness and prevent a chilling effect on defendants exercising their right of appeal out of a fear that the reversed trial court will have a retaliatory motive to impose a more severe sentence on remand. Pearce, 395 U.S. at 724–25. However, even if the “presumption of vindictiveness” arises, the prosecution may rebut it by identifying “objective information in the record justifying the increased sentence.” Goodwin, 457 U.S. at 374. If the prosecution successfully rebuts the presumption, the burden shifts back to the defendant to prove that the sentencing judge violated his due process rights by imposing a harsher sentence out of actual vindictiveness. Smith, 490 U.S. at 799–800 (citing Wasman v. United States, 468 U.S. 559, 569 (1984)). If the prosecution is unable to rebut the presumption of vindictiveness, the new sentence is invalid. Id. at 797–800.

2. Limitations on the Presumption of Vindictiveness

¶21           As a general matter, trial court judges possess significant discretion in determining an appropriate sentence for a specific defendant and offense, so long as the judgment is not influenced by actual vindictiveness against the defendant. Wasman, 468 U.S. at 563–64; Lopez, 113 P.3d at 720. To protect this trial court discretion during a resentencing on remand, the Court has explained that the presumption of vindictiveness is a narrow one, because reasons other than a judge’s vindictiveness may justify a modified or increased sentence. Smith, 490 U.S. at 799–800 (“[T]he Pearce  presumption [of vindictiveness] ‘may operate in the absence of any proof of an improper motive and thus . . . block a legitimate response to criminal conduct’ . . . .” (quoting Goodwin, 457 U.S. at 373)).

¶22           To balance a defendant’s right to appeal with a trial court’s broad discretion in sentencing, the Supreme Court has emphasized that the presumption of vindictiveness should apply only where its objectives—to deter actual vindictive sentencing and to prevent a chilling effect on defendants exercising their right to appeal—are “most efficaciously served.” Id. at 799 (quoting Texas v. McCullough, 475 U.S. 134, 138 (1986)). Accordingly, the presumption arises only if (1) the new sentence on remand is more severe than the original sentence, and (2) there is a reasonable likelihood that actual vindictiveness played a part in the resentencing. Id. at 799–800; Montgomery, 737 P.2d at 416.

3. Applicability of the Presumption of Vindictiveness

¶24           Neither the Supreme Court nor this court has dictated a particular method for determining whether a defendant’s subsequent sentence is objectively more severe than the original sentence for purposes of applying the presumption of vindictiveness. Other jurisdictions have taken three different approaches to determining this threshold issue. First, the “aggregate approach” compares the new aggregate sentence to the original aggregate sentence to determine whether there was an increase. Second, the “remainder aggregate approach” compares the new aggregate sentence only to the aggregate of the original sentences for the convictions that remained after appeal. Third, the “count-by-count approach” isolates each conviction and deems a defendant’s sentence increased if any of the individual sentences increased, regardless of the aggregate effect. Before adopting one of these approaches, it is useful to discuss their relative merits and their application in other jurisdictions.

¶25           First, a majority of courts that have considered the issue apply an aggregate approach to determine whether a new sentence is more or less severe than an original sentence. See, e.g., United States v. Campbell, 106 F.3d 64, 68 (5th Cir. 1997); United States v. Sullivan, 967 F.2d 370, 374–75 (10th Cir. 1992); People v. Savala, 195 Cal. Rptr. 193, 196–97 (Ct. App. 1983), overruled on other grounds by People v. Foley, 216 Cal. Rptr. 865 (Ct. App. 1985); State v. Hudson, 748 S.E.2d 910, 911–14 (Ga. 2013).² Under this approach, where changes to the defendant’s sentence on remand cause the aggregate total sentence to increase as compared to the original aggregate sentence, the presumption of vindictiveness applies.

¶26           The rationale that supports the aggregate approach is that judges imposing sentences in multi-count cases “typically craft sentences on the various counts as part of an overall sentencing scheme,” but when that scheme “unravels,” a judge should have discretion to reevaluate the underlying facts and the sentences on the remaining convictions to “reflect[] the gravity of the crimes for which the defendant is being resentenced.” Hudson, 748 S.E.2d at 913. The aggregate approach thus recognizes the fact-intensive analysis in which the trial court must engage when sentencing a defendant on multiple interrelated convictions, and acknowledges the need for the same level of discretion at resentencing as at the original sentencing proceeding; however, by presuming vindictiveness when the total aggregate sentence increases, the aggregate approach still protects defendants from the fear of actual vindictiveness. See Campbell, 106 F.3d at 68 (“When an appellate court subsequently reverses a conviction (or convictions) that was part of the original sentence, the district court’s job on remand is to reconsider the entirety of the (now-changed) circumstances and fashion a sentence that fits the crime and the criminal. . . . The aggregate approach’s inherent flexibility best comports with this important goal.” (footnote omitted) (citation omitted)).

¶27           The logic of the aggregate approach is apparent, as demonstrated in a recent court of appeals opinion, People v. Woellhaf, 199 P.3d 27, 31 (Colo. App. 2007). In Woellhaf, a jury originally found the defendant guilty of nine separate counts stemming from a sexual assault of a child. Id. at 28. The trial court sentenced the defendant to an aggregate of forty-eight years in prison. Id. After the court of appeals affirmed the trial court in pertinent part, this court granted certiorari and reversed, holding that certain counts violated double jeopardy and instructing the trial court to merge certain convictions and then resentence. Woellhaf v. People, 105 P.3d 209, 220 (Colo. 2005). On remand, the trial court merged the counts and then sentenced the defendant to twenty-four years on a single count of sexual assault on a child by one in a position of trust, and to a concurrent sentence of twelve years for a single count of sexual assault on a child. Woellhaf, 199 P.3d at 29.

¶28           On appeal a second time to the court of appeals, the defendant argued, among other things, that the new sentences violated his due process rights. Id. The court of appeals disagreed, holding that the presumption of vindictiveness did not arise because the aggregate period of incarceration did not increase from the original sentence to the new sentence on remand. Id. at 31–32. In so holding, the court of appeals noted that the aggregate approach deters actual vindictiveness and protects a defendant’s due process rights—the two goals of the presumption of vindictiveness—but also balances the sentencing court’s broad discretion and obligation to consider all relevant factors at resentencing.³Id. at 32 (“If . . . the trial court had been limited to imposing the length of incarceration originally imposed for one of [the four merged] convictions, it would have [sic] effectively have been prevented from considering all relevant and material facts relating to defendant’s conduct for purposes of resentencing.”). As the United States Court of Appeals for the Fifth Circuit has similarly observed, “the aggregate approach best reflects the realities faced by district court judges who sentence a defendant on related counts of an indictment. Sentencing is a fact-sensitive exercise that requires district judges to consider a wide array of factors when putting together a ‘sentencing package.’” Campbell, 106 F.3d at 68.

¶29           Turning to the second approach, two federal circuits have adopted a “remainder aggregate” approach, which compares the new aggregate sentence on remand with the original aggregate sentence only of the counts that remained following the defendant’s successful appeal. United States v. Monaco, 702 F.2d 860, 884–85, 884 n.47 (11th Cir. 1983); United States v. Markus, 603 F.2d 409, 413 (2d Cir. 1979); see also Campbell, 106 F.3d at 68 (describing the remainder aggregate approach). As the court in Monaco observed, the distinctive feature of the remainder aggregate approach is that it accounts for the fact that the defendant’s original appeal succeeded in getting certain convictions and sentences dismissed, and therefore disregards those sentences when determining whether the presumption of vindictiveness applies. 702 F.2d at 885. Thus, the remainder aggregate approach does not affect a trial judge’s discretion to impose a new sentence that is less than or equal to the “valid” aggregate sentence from the original sentencing proceeding, but it does give rise to the presumption of vindictiveness if the sentence exceeds that “valid” original sentence. See id.

¶30           Finally, a small minority of jurisdictions have adopted the “count-by-count” approach, which compares the trial court’s original sentence on an individual conviction against the trial court’s sentence on that same conviction after appeal and remand, applying the presumption of vindictiveness if there is an increase. See, e.g., People v. Moore, 686 N.E.2d 587, 595 (Ill. 1997); see also Anthony v. Hopper, 219 S.E.2d 413, 415 (Ga. 1975), overruled by Hudson, 748 S.E.2d at 914. The remainder aggregate approach and the count-by-count approach share certain characteristics and display similar benefits and drawbacks; therefore, we will discuss the merits of the two approaches in tandem in further detail below.

¶31           In comparison to the aggregate approach, the remainder aggregate and count-by-count approaches give rise to the presumption of vindictiveness in more factual scenarios. The remainder aggregate approach does so because it chooses a different reference point—the original aggregate sentence only of the remaining counts—and applies the presumption of vindictiveness if the new sentence exceeds that reference point, even if the new sentence does not exceed the aggregate sentence on all original counts. The count-by-count approach does so because it focuses the analysis on each individual count and gives rise to the presumption of vindictiveness for an increase on any one count, regardless of aggregate effect. Consequently, these two approaches provide substantial protection to defendants and recognize that the potential for vindictive sentencing presents a real danger of chilling a defendant’s right of appeal. Focusing on the principle that vindictiveness “must play no part in the sentence” a defendant receives after a successful appeal, Pearce, 395 U.S. at 725, the remainder aggregate approach and the count-by-count approach both opt to apply the presumption of vindictiveness more often than the aggregate approach in order to thoroughly deter vindictive sentencing and minimize any possibility, however remote, of creating a chilling effect on defendants’ right of appeal. These two approaches thus prefer to presume vindictiveness in a potentially over-inclusive set of factual circumstances in an attempt to ensure that all actually vindictive sentencing decisions fall within their presumption. If, however, vindictiveness is truly absent in a case in which the presumption arises under either of these approaches, the prosecution has the ability to rebut the presumption and prove that there is no reasonable likelihood of vindictiveness by identifying objective reasons in the record for an increased sentence. ¶32The protection these two approaches afford defendants deserves recognition, but allowing the application of the presumption of vindictiveness in a broader universe of cases could unduly restrain trial court sentencing discretion even where there may be a legitimate reason, under the sentencing statute and the facts of the case, to increase an individual count’s sentence on remand while preserving or decreasing the aggregate sentence. Restraining sentencing discretion in this way could result in harsher sentences for defendants because rather than sentencing each count separately and independently of one another to avoid any possibility of punishing a defendant double for a certain offense, judges might impose a harsher sentence on each count in order to protect against a claim of vindictiveness after a future resentencing. A recent Georgia Supreme Court case, State v. Hudson, helps to illustrate this effect and its possible repercussions. 748 S.E.2d at 911.

¶33           In Hudson, the defendant was originally sentenced to an aggregate of twenty-five years in prison and life on probation based upon concurrent sentences on one count of sexual battery and one count of child molestation. Id. Pertinent here, the child molestation component of that original sentence was ten years in prison and twenty years on probation. Id. After the defendant successfully appealed his convictions, on remand the trial court followed the court of appeals’ instructions and merged the two convictions. Id. It subsequently resentenced Hudson on the single remaining count, child molestation, to twenty-five years in prison and five years on probation. Id. Thus, the aggregate sentence decreased from twenty-five years in prison and life on probation to twenty-five years in prison and five years on probation. Id.  However, the child molestation count, when considered in isolation, increased from ten years in prison and twenty years on probation to twenty-five years in prison and five years on probation. Id. at 912. In light of the increased child molestation sentence and Georgia’s adoption in 1975 of the count-by-count approach, the Georgia Court of Appeals reversed the new sentence by presuming vindictiveness and concluding that the trial court violated the defendant’s due process rights. Id.; see also Anthony, 219 S.E.2d at 415. On appeal, the prosecution urged the Georgia Supreme Court to abandon the count-by-count approach in favor of the aggregate approach, relying in part on that court’s recent three-justice plurality recommending that course of action. Hudson, 748 S.E.2d at 912; see also Adams v. State, 696 S.E.2d 676, 679–81 (Ga. 2010). The court agreed that there was sufficient reason to part with its precedent and therefore overruled Anthony and adopted the aggregate approach. Hudson, 748 S.E.2d at 914.

¶34           In explaining its decision, the court provided several reasons to part with the count-by-count approach. Id. at 913. First, the court noted that the presumption of vindictiveness seeks to prevent vindictiveness, not any “mere imposition of an enhanced sentence on retrial or remand.” Id. The court therefore found it necessary to tailor the applicability of the presumption of vindictiveness and try to ensure that it applied only in cases in which that goal is “most efficaciously served.” Id. (quoting Smith, 490 U.S. at 799). After observing that the Supreme Court itself has “circumscribed” the applicability of the presumption of vindictiveness in numerous cases since Pearce, id. (citing Smith, 490 U.S. at 801; McCullough, 475 U.S. at 139–40; Chaffin v. Stynchcombe, 412 U.S. 17, 18 (1973)), the Georgia court concluded that the count-by-count approach was not satisfactorily precise in determining when the presumption should apply, and therefore adopted the aggregate approach in its place:

[T]he Pearce presumption [of vindictiveness] is a blunt instrument for ascertaining trial court vindictiveness. For the same reasons the Supreme Court has seen fit to limit the factual circumstances in which the Pearce  presumption is triggered, we see fit to adopt the aggregate approach, as the more precise of the two methods [the other being the count-by-count approach] in identifying actual trial court vindictiveness.

Id.

¶36           In sum, the Georgia Supreme Court concluded that the count-by-count approach does not strike as careful a balance as the aggregate approach, in large part because of “its relative imprecision in identifying actual trial court vindictiveness and failure to recognize the ‘big picture’ in a trial judge’s sentencing decisions.” Id. at 914. The same could be said of the remainder aggregate approach: A trial court may have legitimate reasons to increase the aggregate sentence on the remaining counts after an appellate court invalidates the conviction for some other count, but under the remainder aggregate approach the trial court’s discretion to impose such an increase is constrained by the specter of the presumption of vindictiveness, regardless of whether there was any “realistic motive for vindictive sentencing.” McCullough, 475 U.S. at 139. In part because of these concerns, both federal circuits that have adopted the remainder aggregate approach have strayed from that approach under certain circumstances, although neither has expressly overruled it. See United States v. Fowler, 749 F.3d 1010, 1022 (11th Cir. 2014) (limiting the application of Monaco’s remainder aggregate approach in light of changes to federal sentencing law); United States v. Mata, 133 F.3d 200, 201–02 (2d Cir. 1998) (holding that the presumption of vindictiveness did not arise, despite the fact that the remainder aggregate approach would have triggered it, because the facts did not show any “hint of vindictiveness”).

¶37           With all of these approaches in mind, we conclude that the aggregate approach taken by a majority of courts is the appropriate framework for applying the presumption of vindictiveness. The aggregate approach best effectuates the Supreme Court’s instruction to apply the presumption only where its objectives—to deter actual vindictiveness and avoid a chilling effect on defendants exercising their right to appeal—are “most efficaciously served.” Smith, 490 U.S. at 799 (quoting McCullough, 475 U.S. at 138). When a new sentence is less than the original aggregate sentence, the defendant necessarily benefits from his appeal and the presumption of vindictiveness need not apply in order to avoid a chilling effect on defendants exercising their constitutional rights. Moreover, under the aggregate approach, sentencing judges retain their broad discretion in sentencing a defendant both before and after remand, taking into account the panoply of factors relevant to those determinations. A reasonable likelihood of vindictiveness does not necessarily arise, and therefore vindictiveness should not be presumed, when a sentencing judge increases a sentence on one conviction but the aggregate time of incarceration does not exceed the aggregate original sentence; the increase on the individual count may simply reflect the fact that the trial court was faced with a different situation, in which the defendant was subject to a different combination of charges for the same criminal episode. In this way, the aggregate approach strikes the desired balance between protecting a defendant’s right to appeal, on the one hand, and ensuring that trial courts retain their broad discretion in sentencing, on the other.

¶38           Having adopted the aggregate approach, we now determine whether the presumption of vindictiveness applies to Johnson’s new possession of a controlled substance sentence.

4. Application

¶39           Applying the aggregate approach to Johnson’s case, we conclude that the presumption of vindictiveness does not arise. Johnson originally faced an aggregate sentence of twenty years to life based upon his concurrent sentences of twenty years for first-degree kidnapping, twenty years to life for sexual assault, and six years for possession of a controlled substance. After his successful appeal in which the court of appeals vacated the kidnapping and sexual assault convictions, the trial court resentenced Johnson to twelve years on the possession of a controlled substance count. Compared to the original aggregate of twenty years to life, twelve years is not an increased sentence and therefore does not give rise to a presumption of vindictiveness.

¶40           Notably, when originally sentencing Johnson the trial court was careful to sentence each count separately and independent of the other offenses in order to prevent any possibility of punishing Johnson double for a certain offense. In particular, although the court believed that the sexual assault was a factual basis for the possession of a controlled substance conviction, it chose not to consider that fact in sentencing Johnson on the possession of a controlled substance count because it was imposing a lengthy sentence on the sexual assault count. Subsequently, in increasing the possession of a controlled substance sentence on remand, the trial court emphasized that it was “looking at a totally different situation than [it] was at the time when [Johnson] was facing a possibility of life in prison” for the sexual assault charge. While the facts of the criminal episode known to the trial court were the same at the time of original sentencing and at resentencing, the convictions stemming from those facts had changed, potentially giving rise to a reasonable and objective justification for the increased sentence for the possession of a controlled substance conviction.⁴See Smith, 490 U.S. at 802 (“[A]ny unexplained change in the sentence is . . . subject to a presumption of vindictiveness. In cases like the present one, however, we think there are enough justifications for a heavier second sentence . . . .”). In short, and as the result of the aggregate approach shows, it is not reasonably likely that the trial court acted vindictively in this case because it possessed an objective reason to reevaluate Johnson’s possession of a controlled substance sentence and increase it without exceeding the original aggregate sentence imposed.

¶41           Given that the presumption of vindictiveness does not apply, Johnson must “affirmatively prove actual vindictiveness” in order to prevail on his due process claim. Wasman,468 U.S. at 569. In arguing that the new sentence for possession of a controlled substance was the product of actual vindictiveness, Johnson points to the fact that the trial court did not learn of new facts relating to his conduct after the original sentencing that would justify doubling the possession of a controlled substance sentence. Because the only intervening event was the court of appeals’ reversal of the kidnapping and sexual assault convictions, Johnson argues, the trial court’s decision to increase the possession of a controlled substance sentence on remand could only have been motivated by vindictiveness against Johnson for pursuing that reversal.

¶42           In essence, Johnson asks us to presume actual vindictiveness from the simple fact that the possession of a controlled substance sentence increased after the reversal and remand, even though we have already rejected the count-by-count approach and found the presumption of vindictiveness inapplicable to his case. Under the facts here, Johnson failed to present evidence sufficient to show that the trial court actually possessed a vindictive motive to sentence him to twelve years on the possession of a controlled substance count. At the original sentencing hearing, the trial court accounted for the fact that Johnson was subject to sentencing on three different counts and therefore discretely sentenced Johnson on each count in order to avoid punishing him double for any one crime. The court of appeals’ reversal of the kidnapping and sexual assault convictions, however, caused that sentencing scheme to unravel, thus requiring the trial court to reassess Johnson’s sentence and exercise its sentencing discretion under new and different circumstances. The trial court’s statements explaining why it did not initially consider the facts pertaining to the sexual assault when sentencing Johnson on the possession of a controlled substance conviction, and why it felt it necessary to consider those facts after the court of appeals vacated the kidnapping and sexual assault convictions, evidence a non-vindictive reason to modify Johnson’s sentence and increase the possession of a controlled substance sentence from six years to twelve years. Because there is no evidence to contradict the trial court’s statements and show the presence of actual vindictiveness, Johnson’s due process claim fails.

¶43           Because Johnson’s aggregate sentence decreased from twenty years to life to a determinate term of twelve years, the presumption of vindictiveness does not apply. Accordingly, we conclude that the court of appeals erred in applying the presumption of vindictiveness and holding that the trial court violated Johnson’s due process rights by sentencing Johnson to twelve years on the possession of a controlled substance count. We also conclude that Johnson failed to prove that the trial court violated his due process rights by increasing his possession of a controlled substance sentence out of actual vindictiveness.

IV. Conclusion

¶44           We hold that the court of appeals correctly rejected Johnson’s statutory argument that section 18-1-409 applies to his case, but that it erred in applying a presumption of vindictiveness to his sentence on the possession of a controlled substance count because his aggregate sentence following appeal was less than his original aggregate sentence. Accordingly, we affirm the judgment of the court of appeals rejecting Johnson’s statutory argument under section 18-1-409, but we reverse its holding that the new sentence violated Johnson’s due process rights under the presumption of vindictiveness. We also conclude that, based upon the record before us, the trial court did not act out of actual vindictiveness, and therefore did not violate Johnson’s due process rights, when it resentenced him to twelve years on the possession of a controlled substance count. We thus remand the case to the court of appeals for proceedings consistent with this opinion and for the court of appeals to consider the remaining issues raised on appeal.