State v. Mendoza

455 P.3d 705, 248 Ariz. 6

Opinions

Opinion

IN THE

ARIZONA COURT OF APPEALS

DIVISION ONE

STATE OF ARIZONA, Appellee,

v.

VINCENT MENDOZA, Appellant.

No. 1 CA-CR 18‑0512

FILED 11-21-2019

Appeal from the Superior Court in Maricopa County No. 2017-002264-001 DT

The Honorable Annielaurie Van Wie, Judge Pro Tempore

VACATED IN PART AND REMANDED

COUNSEL

Arizona Attorney General’s Office, Phoenix By Eric Knobloch Counsel for Appellee

Maricopa County Public Defender’s Office, Phoenix By Paul J. Prato Counsel for Appellant

OPINION

Presiding Judge Paul J. McMurdie delivered the opinion of the Court, in which Judge Randall M. Howe and Judge Jennifer B. Campbell joined.

Opinion of the Court

M c M U R D I E, Judge:

¶1 Vincent Mendoza appeals his conviction and sentence for one count of aggravated driving or actual physical control of a vehicle while under the influence of intoxicating liquor or drugs, a class 4 felony. We hold: (1) a superior court judge who, in violation of Arizona Rule of Criminal Procedure (“Rule”) 17.4(a)(2), participates in settlement discussions between a defendant and the State without the parties’ consent, errs by thereafter presiding over that defendant’s trial and sentencing; (2) such error is fundamental if the totality of the circumstances raises a presumption of judicial vindictiveness; and (3) if the presumption is unrebutted by the State, it requires the defendant to be either resentenced or retried before a different judge. Because we find an unrebutted presumption of judicial vindictiveness exists regarding Mendoza’s sentence, we affirm his conviction but vacate his sentence and remand for resentencing before a different superior court judge.

FACTS AND PROCEDURAL BACKGROUND 1

¶2 On a night in October 2016, Officer Jaime Cole, a patrol officer with the Goodyear Police Department, noticed Mendoza’s vehicle traveling northbound at speed slower than the posted speed limit. Cole decided to follow the car and soon saw that Mendoza was having trouble staying in his lane. Mendoza nearly struck a curb while executing a lane change and crossed over a solid white fog line. Based on these observations, Cole initiated a traffic stop. Mendoza was seated in the driver’s seat. As Cole spoke with Mendoza, she noticed Mendoza’s eyes were red, bloodshot, and watery; he had difficulty multi-tasking; and the odor of alcohol was coming from inside the vehicle. Mendoza admitted to Cole that he drank six beers that night and had an ignition interlock device installed in the vehicle “to prevent this,” and added that a friend had blown into the machine to allow him to drive the car. Eventually, Cole arrested Mendoza for driving under the influence.

¶3 At the police station, Mendoza underwent blood and breath testing. The blood testing returned a blood alcohol concentration of 0.128. The breath testing returned results of 0.119 and 0.117. After the screening,

1 We view the facts in the light most favorable to upholding the verdict and resolve all reasonable inferences against Mendoza. State v. Harm, 236 Ariz. 402, 404, ¶ 2, n.2 (App. 2015).

2

Mendoza waived his Miranda 2 rights, and Officer Cole interviewed him. During the interview, Mendoza again admitted drinking at least six beers earlier that night but denied that he felt the effects of the alcohol or that he was impaired to the slightest degree.

¶4 The State charged Mendoza with one count of driving or actual physical control while under the influence of intoxicating liquor and under a court order to equip a certified ignition interlock device (“Count 1”) and one count of driving or actual physical control while he had an alcohol concentration of 0.08 or more in his body within two hours of the time of driving and while under a court order to equip a certified ignition interlock device (“Count 2”), both class 4 felonies. In March 2018, during a status conference before the assigned trial judge, the parties indicated to the court that the State had offered a plea agreement with that day as the deadline for Mendoza to accept or reject it. Mendoza’s counsel stated that Mendoza “wanted to discuss the case with the Court.” The court agreed and proceeded to inform Mendoza of the charges, their elements, and that the plea offer was for nine years’ imprisonment.

¶5 Upon further questioning, the State asserted that because Mendoza had two prior historical felony convictions for aggravated driving under the influence and aggravated assault involving a vehicle, he would be sentenced as a category three repetitive offender if convicted, Ariz. Rev. Stat. (“A.R.S.”) § 13-703(C), and that under A.R.S. § 13-703(J), he would face a sentencing range of 6 to 15 years. The court then made the following statements to Mendoza:

Okay. So, Mr. Mendoza, after trial you would face 6 to 15. That six is completely, 100 percent unreasonable. There is not a judge here that will give you six, period. The fact that you’ve already done seven and a half on a vehicular aggravated assault . . . means someone’s going to give you more. Okay? And seven and a half on a vehicular aggravated assault and then coming back with essentially what looks like a third felony DUI type activity, when you already assaulted somebody, like—no.

* * *

Your chances of getting nine after trial are literally about zero. Your chances of getting 10 are extremely low.

2 Miranda v. Arizona, 384 U.S. 436 (1966).

3

Your chances of getting above 10 like the 12, 13, 14, 15, that’s —it’s [a] better chance you get that than you get ten or nine for sure.

I can’t imagine anyone giving you less than the presumptive when you’ve already been to prison for seven and a half years related to a vehicular assault and you've got another Agg DUI and you’re here for another Agg DUI, and you go to trial because when you go to trial, you also lose the mitigation of acceptance of responsibility, remorse, saving taxpayer time, court time, saving money for the courts. You lose all of that that goes in the good pile.

Okay. When you go to trial, that just doesn’t exist.

Okay? So when you go to trial, you have less in the good pile. When we’re weighing the good and the bad, and if you have less in the good pile, the stuff in the bad pile weighs heavier. Does that make sense?

The court continued to discuss Mendoza’s case with him, including the sentence he would receive if he went to trial and was convicted, with little intervention from either Mendoza’s counsel or the State. The following colloquy occurred during this conversation:

THE COURT: If you are convicted at trial, you will get more than nine years. So your plea offer may save you some time. It might not be what you want, but they’re not going to give you something better.

[MENDOZA]: I’m by myself, Your Honor, so if I get 10 years, I get 10 years. . . . I have no family. When my mom died my family just—just disowned me and not supported me.

THE COURT: And your time is not worth anything? You’d rather just spend your years in prison?

[MENDOZA]: I was an electrician for—I was an electrician for the prison so I worked the whole time I was there.

THE COURT: Oh, so you like prison?

[MENDOZA]: I don’t like it, but I don’t like—just I’m putting pretty much a gun to my head. Either do or die.

4

THE COURT: No, you’re right. You’re pretty much in the same spot as everybody else in here when they have to decide on a plea.

[MENDOZA]: I made a mistake, Your Honor. I can’t change the past. But unfortunately—

THE COURT: Well, you can’t but you can try to minimize the damage. But if you want to go to prison for longer, you let me know at sentencing because I’ll be happy—

[MENDOZA]: I don’t.

THE COURT: —to send you for 15 and have no problem with it.

[MENDOZA]: I don’t want to, that’s why I (indiscernible) and maybe do some understanding. I’m not saying that I’m not guilty. I’m just understanding—

THE COURT: I’m telling you after trial, your chances of getting below 10 are about zero. And not just from me, from about anyone.

At the end of the hearing, Mendoza rejected the State’s plea offer, and the case went to trial. The same superior court judge who participated in the March 2018 ad hoc settlement conference presided over the trial.

¶6 After a four-day trial, the jury found Mendoza guilty of Count 1 but acquitted him of Count 2. Before Mendoza’s sentencing, the Maricopa County Adult Probation Department prepared a presentence report that recommended Mendoza be sentenced to the presumptive term of imprisonment, 10 years. The court found that Mendoza had four prior felony convictions—two of which were historical—and proceeded to sentencing. In considering the mitigators at issue in the case, the court acknowledged Mendoza had sought treatment while awaiting trial but explained that:

There is no acceptance of responsibility. There is no remorse. There is none of the mitigators that you get when you take a plea. I explain to people that . . . means that just doesn’t exist in that pile any longer. There is no punishment for taking a case to trial. However, I hear a lot more in the trial than I ever hear when somebody does a plea.

5

The court found as aggravating circumstances that Mendoza committed the offense a short time after being released from prison, he circumvented the ignition interlock device to drive his vehicle, and he was driving with a passenger in his car. The court also found that Mendoza had previously been convicted of aggravated driving under the influence within the ten years preceding the date of the offense, a statutory aggravator. A.R.S. § 13-701(D)(11). The court concluded the maximum sentence was appropriate and sentenced Mendoza to 12 years’ imprisonment with 315 days’ presentence incarceration credit. After rendering its sentence, the court stated:

The aggravators are just exceptional and my concern for the safety of the community given what I heard during the trial is —I’m very concerned. So based on that I don’t think I can give any less than [12 years].

Mendoza appealed, and we have jurisdiction under A.R.S. §§ 12-120.21(A)(1), 13-4031, and -4033(A)(1).

¶7 Mendoza’s appellate counsel filed a brief in accordance with Anders v. California, 386 U.S. 738 (1967), and State v. Leon, 104 Ariz. 297 (1969), certifying that, after a diligent search of the record, he found no arguable question of law that was not frivolous. Counsel asked this court to search the record for arguable issues. See Penson v. Ohio, 488 U.S. 75 (1988); State v. Clark, 196 Ariz. 530, 537, ¶ 30 (App. 1999). Mendoza was allowed to file a pro se supplemental brief and did so. After reviewing the record and Mendoza’s brief, we requested the parties brief whether the superior court judge who ultimately presided over Mendoza’s trial violated Rule 17.4(a)(2) by participating in the settlement discussions at the March 2018 hearing

6

without obtaining the consent of the parties and, if so, what remedy exists for a violation of the rule. 3

DISCUSSION

¶8 Mendoza argues the superior court judge who presided over his trial and sentencing committed fundamental error by violating Rule 17.4(a)(2)’s prohibition of an assigned trial judge’s participation in settlement discussions without the parties’ consent. Mendoza contends the judge’s statements at sentencing, when interpreted considering her erroneous participation in settlement discussions, reflected “a sentencing bias towards [him] because he did not take the plea offer.” In response, the State asserts: (1) Mendoza invited any theoretical error caused by the judge’s involvement in settlement discussions during the hearing; (2) no error occurred because Rule 17.4(a)(2) only applies to settlement conferences ordered by the superior court, not to ad hoc settlement discussions; and (3) even if the court violated Rule 17.4(a)(2), Mendoza has not carried his burden of showing that such error was fundamental and prejudicial.

¶9 Because Mendoza failed to object either to the assigned trial judge’s participation in settlement discussions or to the judge thereafter presiding over the trial, we agree with the parties that we must apply fundamental error review. In State v. Escalante, 245 Ariz. 135 (2018), our supreme court described the proper procedure for fundamental error review as follows:

3 In his supplemental brief, Mendoza also raised the following arguments: (1) if he had testified, he would have been found innocent; (2) he did not receive an adequately speedy trial; (3) the State should not have been allowed to introduce a recording of the traffic stop; and (4) his trial counsel was ineffective because she was inexperienced, lacked knowledge of his case, failed to object to evidence introduced by the State, and failed to introduce relevant evidence. “[I]neffective assistance of counsel claims are to be brought in [Arizona Rule of Criminal Procedure] 32 proceedings . . . [and] [a]ny such claims . . . will not be addressed by appellate courts regardless of merit.” State v. Spreitz, 202 Ariz. 1, 3, ¶ 9 (2002). As for the other arguments, we have reviewed the record for arguable error as to each claim and, finding none, will not address them further.

7

[T]he first step in fundamental error review is determining whether trial error exists. If it does, an appellate court must decide whether the error is fundamental. In doing so, the court should consider the totality of the circumstances. A defendant establishes fundamental error by showing that (1) the error went to the foundation of the case, (2) the error took from the defendant a right essential to his defense, or (3) the error was so egregious that he could not possibly have received a fair trial. If the defendant establishes fundamental error under prongs one or two, he must make a separate showing of prejudice, which also “involves a fact-intensive inquiry.” If the defendant establishes the third prong, he has shown both fundamental error and prejudice, and a new trial must be granted. The defendant bears the burden of persuasion at each step.

Id. at 142, ¶ 21 (citations omitted) (quoting State v. Henderson, 210 Ariz. 561, 568, ¶ 26 (2005)). We address each element in turn.

A. Rule 17.4(a)(2) Applies, and the Assigned Trial Judge Committed Error by Presiding Over Mendoza’s Trial and Sentencing After Participating in Settlement Discussions Without the Parties’ Consent.

¶10 Rule 17.4(a)(2) provides:

At either party’s request or on its own, a court may order counsel with settlement authority to participate in good faith discussions to resolve the case in a manner that serves the interests of justice. The assigned trial judge may participate in this discussion only if the parties consent. In all other cases, the discussion must be before another judge. If settlement discussions do not result in an agreement, the case must be returned to the trial judge.

(Emphasis added.)

¶11 The State argues the assigned trial judge’s participation in settlement discussions at the March 2018 hearing was not error. Referencing the statement in Rule 17.4(a)(2)’s first sentence that the court may “order counsel . . . to participate in good faith discussions,” the State argues the phrase “this discussion” in the rule’s second sentence refers only to settlement discussions ordered explicitly by the court. Because no order to participate in “good faith discussions” preceded the March 2018 hearing,

8

the State concludes the requirement that an assigned trial judge only participate with the parties’ consent was never triggered, Rule 17.4(a)(2) was not violated, and the court’s involvement in settlement discussions at that hearing was not error.

¶12 We review the interpretation of court rules de novo, State v. Fitzgerald, 232 Ariz. 208, 210, ¶ 10 (2013), and “according to the principles of statutory construction,” Potter v. Vanderpool, 225 Ariz. 495, 498, ¶ 8 (App. 2010) (quoting Bolding v. Hantman, 214 Ariz. 96, 100, ¶ 16 (App. 2006)). “If a rule’s language is plain and unambiguous, we apply it as written without further analysis.” State v. Salazar-Mercado, 234 Ariz. 590, 592, ¶ 4 (2014). In determining the plain meaning of a specific provision, we read its words in context and “look to the [rule] as a whole.” Stambaugh v. Killian, 242 Ariz. 508, 509, ¶ 7 (2017). However, when a rule contains ambiguous language, “we apply secondary principles of construction, such as examining the rule’s context, its relationship to related rules, its background, and the consequences of differing interpretations.” Roberto F. v. DCS, 237 Ariz. 440, 441, ¶ 6 (2015). We also apply secondary principles of construction if “application of the plain meaning would lead to impossible or absurd results.” Bilke v. State, 206 Ariz. 462, 464, ¶ 11 (2003).

¶13 Although Rule 17.4(a)(2)’s consent requirement does not explicitly apply to settlement discussions other than those ordered by the court, construing the provision’s requirements to apply to all judicial participation in settlement discussions is the only reasonable interpretation of the rule. Rule 17.4 outlines the process by which parties to a criminal proceeding may negotiate, craft, and submit plea agreements. Rule 17.4(a) sets out three categories of individuals—the parties (i.e., the defendant and the State), judges, and victims—and defines the rights of each group to participate in settlement discussions. Subsection 17.4(a)(1), for example, recognizes the right of the defendant and the State to “negotiate and reach agreement on any aspect of a case,” while subsection 17.4(a)(3) provides victims “the opportunity to be present and to be heard regarding settlement” if the defendant attends settlement discussions. This structure strongly suggests that Rule 17.4(a)(2) was intended to provide the exclusive avenue through which a court may participate in settlement discussions.

¶14 Review of the history and purpose of Rule 17.4(a) confirms that the rule’s drafters intended it to be the exclusive exception to a general rule barring judicial participation in settlement discussions. See State v. Jordan, 137 Ariz. 504, 508–09 (1983) (holding judges may not participate in settlement discussions). When the modern Arizona Rules of Criminal Procedure were first promulgated in 1973, Rule 17.4(a) stated:

9

The parties may negotiate concerning, and reach an agreement on, any aspect of the disposition of the case. The court shall not participate in any such negotiation.

(Emphasis added.) Ten years later, in State v. Jordan, our supreme court described the purpose of the original rule’s prohibition on judicial participation in settlement discussions. 137 Ariz. at 508. First, it found that Rule 17.4(a) prevented the trial judge’s presence and involvement in settlement discussions from naturally exerting a “subtle but powerful influence” on the defendant’s decision to accept or reject a plea offer. Id. at 509 (quoting United States ex rel. Elksnis v. Gilligan, 256 F. Supp. 244, 254 (S.D.N.Y. 1966)). Second, the court observed that the rule preserved “public confidence in the impartial and objective administration of criminal justice” by maintaining the trial judge’s image as “a neutral arbiter.” Id. (quoting United States v. Werker, 535 F.2d 198, 203 (2d Cir. 1976)). Finally, the court stated that “when a sentence is imposed after judicial participation in plea bargaining, it is unclear whether the sentence imposed was based on reason or based on the fulfillment of the previous threat.” Id. (citing People v. Clark, 515 P.2d 1242, 1242–43 (Colo. 1973)).

¶15 Rule 17.4(a) remained substantively unchanged until 1999, when, after a two-year experimental period, see Ariz. R. Crim. P. 17.4 cmt. to 1997 amend., the Arizona Supreme Court permanently adopted an amendment to the rule permitting judges to participate in settlement discussions. Like the current version of the rule, however, the 1999 amendment did not grant judges the discretion to participate in whatever manner they deemed appropriate. Instead, the 1999 amendment crafted a specific procedure for judges to follow before participating in settlement discussions:

At the request of either party, or sua sponte, the court may, in its sole discretion, participate in settlement discussions by directing counsel having the authority to settle to participate in a good faith discussion with the court regarding a non-trial or non-jury trial resolution which conforms to the interests of justice. . . . The trial judge shall only participate in settlement discussions with the consent of the parties. In all other cases, the discussions shall be before another judge or a settlement division. If settlement discussions do not result in an agreement, the case shall be returned to the trial judge.

Ariz. R. Crim. P. 17.4(a) (1999) (emphasis added). By limiting the participation of the trial judge, unless the parties consented to that judge’s

10

involvement, the 1999 amendment afforded defendants the benefits of judicial presence in settlement discussions—including the ability to obtain a neutral, candid, and informed perspective on the strengths and weaknesses of the case and any plea offer—while avoiding the constitutional hazards that prompted the original rule.

¶16 Rule 17.4(a) was abrogated and replaced along with the rest of the Arizona Rules of Criminal Procedure in 2017. However, the drafters of the current version of Rule 17.4 noted that any changes to the provisions of the rule governing judicial participation in settlement conferences were intended only to be stylistic, not substantive. See Petition to Amend the Arizona Rules of Criminal Procedure, R-17-0002, app. B, at 31 (proposed Jan. 8, 2017) (listing substantive changes to Rule 17.4 and noting that all other changes are intended to be stylistic), https://www.azcourts.gov/Rules-Forum/aft/661; see also Ariz. R. Crim. P. prefatory cmt. to 2018 amends. (“The intent of these differences is to make the [Arizona Rules of Criminal Procedure] more functional, and easier to understand and use.”). Therefore, we conclude Rule 17.4(a)(2) retains the careful balance struck by the 1999 version of the rule, including the requirement that the assigned trial judge “only participate in settlement discussions with the consent of the parties.” Ariz. R. Crim. P. 17.4(a) (1999).

¶17 Having analyzed Rule 17.4(a)(2)’s meaning and purpose, we now turn to the application of the rule in this case. The State argues Mendoza implicitly consented to the assigned trial judge’s participation by defense counsel’s statement that he “wanted to discuss the case with the Court.” We decline to draw the consent Rule 17.4(a)(2) requires from such an ambiguous statement. Moreover, the assigned trial judge, not the parties, crossed the line into settlement discussions at the March 2018 hearing by going beyond a simple explanation of the charges and the potential sentences and expressing her opinion on the sentence Mendoza would receive were he convicted after trial. However, the judge’s violation of Rule 17.4(a)(2) at the March 2018 hearing did not, in and of itself, amount to legally cognizable error. Thus, if the assigned trial judge had acquired the parties’ consent to remain the trial judge, or recused herself at any time before the trial, the due process concerns potentially raised by a violation of the rule would have dissipated and Mendoza could claim no error, fundamental or otherwise. The error that occurred here was that the assigned trial judge presided over Mendoza’s trial and sentencing without getting the parties’ consent after conducting the settlement conference.

¶18 Before we address whether this error was fundamental, we note a critical exception to our conclusion that Rule 17.4(a)(2) applies to all

11

judicial involvement in settlement discussions: a Donald 4 hearing. What transpired was not a Donald hearing. A Donald hearing is a pre-trial hearing where a defendant is informed of any outstanding plea offer and the consequences of conviction so that a record of the defendant’s rejection of the plea offer can be made to guard against any “late, frivolous, or fabricated claims” of ineffective assistance of counsel “after a trial leading to conviction with resulting harsh consequences.” Missouri v. Frye, 566 U.S. 134, 146 (2012). Although a Donald hearing may have some similarities to settlement discussions contemplated by Rule 17.4(a)(2), the rule’s requirements are not triggered because a Donald hearing is merely informational; the court only informs the defendant of the plea offer and the sentencing range faced if convicted. A court crosses the line from a Donald hearing to a settlement conference when the court discusses the pros and cons of accepting a plea, the likelihood of a conviction or sentence, and any attempt to negotiate a settlement. Neither party disputes that the court, in this case, crossed that line.

B. The Assigned Trial Judge’s Improper Participation in the Settlement Discussions Was Fundamental Error.

1. The Invited Error Doctrine Does Not Apply.

¶19 The State next argues that even if an error occurred as a result of the assigned trial judge’s participation in settlement discussions, we should not consider whether it was fundamental or prejudicial because Mendoza invited any error by asking the assigned trial judge to discuss the case with him. We disagree.

¶20 The invited error doctrine describes the well-established judicial principle that appellate courts “will not find reversible error when the party complaining of it invited the error.” State v. Logan, 200 Ariz. 564, 565–66, ¶ 9 (2001) (collecting cases). The purpose of this doctrine is to “prevent a party from ‘inject[ing] error in the record and then profit[ing] from it on appeal.’” Id. at 566, ¶ 11 (alterations in original) (quoting State v. Tassler, 159 Ariz. 183, 185 (App. 1988)). To determine whether the invited error doctrine should preclude a party from raising an allegation of error on appeal, courts “must look ‘to the source of the error, which must be the party urging the error.’” State v. Lucero, 223 Ariz. 129, 135, ¶ 17 (App. 2009) (quoting Logan, 200 Ariz. at 566, ¶ 11); see also Escalante, 245 Ariz. at 145,

4 State v. Donald, 198 Ariz. 406, 418, ¶ 46 (App. 2000).

12

¶ 38 (“‘Invited error’ occurs when the defendant is the source of that error.”).

¶21 We reject the State’s assertion that Mendoza’s request to “discuss the case with the Court” made him the source of the error. To say Mendoza invited whatever error resulted from his general request to speak with the court would cause an unprecedented expansion of a doctrine that should be approached with “extreme caution.” Lucero, 223 Ariz. at 135, ¶ 18 (quoting State v. Smith, 101 Ariz. 407, 409 (1966)). Accordingly, we conclude the invited error doctrine does not apply.

2. The Assigned Trial Judge’s Participation in Settlement Discussions Raised a Presumption of Judicial Vindictiveness and, Therefore, Constituted Fundamental Error.

¶22 We must next determine whether the trial judge’s participation in settlement discussions without consent of the parties was fundamental error. Mendoza argues the trial judge’s statements during the settlement conference, including her assertion that he would not receive less than the presumptive sentence if he exercised his right to a jury trial, reflects “a sentencing bias . . . because he did not take the plea offer” that ultimately influenced her decision to impose the maximum sentence. In response, the State argues the assigned trial judge’s participation was a mere “technical violation of a prophylactic rule” and, therefore, “not the rare case that involves fundamental error.” To meet his burden at this stage, Mendoza must show the error “was fundamental in light of the facts and circumstances of this case, recognizing that ‘the same error may be fundamental in one case but not in another.’” State v. James, 231 Ariz. 490, 493, ¶ 13 (App. 2013) (quoting State v. Bible, 175 Ariz. 549, 572 (1993)).

¶23 Mendoza has not argued that the trial judge’s erroneous participation in settlement discussions had any effect on his trial, and rightly so. Our review of the record reveals no indication that Mendoza received anything but a fair trial before an impartial judge and jury. Instead, we agree with the State that Mendoza’s claim of fundamental error is best characterized as an allegation of “judicial vindictiveness,” or a claim that the assigned trial judge violated his due-process rights by imposing the maximum sentence, at least in part, because he rejected the State’s plea offer and exercised his Sixth Amendment right to a jury trial.

¶24 The constitutional prohibition upon judicial vindictiveness in sentencing was first recognized by the Supreme Court in North Carolina v.

13

Pearce, 395 U.S. 711 (1969). Pearce concerned two consolidated actions brought by defendants who, after successfully appealing and reversing their original convictions due to constitutional error, received harsher sentences following a retrial. Id. at 713–15. The court held that “penalizing those who choose to exercise” a constitutional or statutory right to appeal their original conviction “would be patently unconstitutional” under the Due Process Clause of the Fourteenth Amendment. Id. at 724 (quoting United States v. Jackson, 390 U.S. 570, 581 (1968)). Thus, the Pierce Court concluded due process required a presumption of judicial vindictiveness “whenever a judge imposes a more severe sentence upon a defendant after a new trial.” Id. at 726. To rebut that presumption, the Court held that the reasons for the more severe sentence must “affirmatively appear” in the record and be “based upon objective information concerning identifiable conduct on the part of the defendant occurring after the time of the original sentencing proceeding.” Id.

¶25 Twenty years later, in Alabama v. Smith, 490 U.S. 794, 798 (1989), the Court reconsidered the Pearce presumption in a situation where, after a defendant successfully moved to set aside a plea agreement, the same judge imposed a harsher sentence after trial than the judge had imposed pursuant to the plea agreement. Acknowledging Pearce’s progeny had limited its application to instances “where its objectives are thought most efficaciously served,” Smith, 490 U.S. at 799 (quoting Texas v. McCullough, 475 U.S. 134, 138 (1986)), the Court held that a presumption of judicial vindictiveness would only arise where there was a “‘reasonable likelihood’ that the increase in sentence [was] the product of actual vindictiveness on the part of the sentencing authority,” id. (citation omitted) (quoting United States v. Goodwin, 457 U.S. 368, 373 (1982)). The Court then applied the “reasonable likelihood” test and concluded that “when a greater penalty is imposed after trial than was imposed after a prior guilty plea, the increase in sentence is not more likely than not attributable to the vindictiveness on the part of the sentencing judge.” Id. at 801. In doing so, the court recognized crucial differences between the factors and information at play at sentencing following a guilty plea and sentencing after trial, including that: (1) “the judge may gather a fuller appreciation of the nature and extent of the crimes charged”; (2) “[t]he defendant’s conduct during trial may give the judge insights into his moral character and suitability for rehabilitation”; and (3) “the factors that may have indicated leniency as consideration for the guilty plea are no longer present.” Id.

¶26 However, neither Smith nor any decision of the Court since has applied a judicial vindictiveness analysis to a situation analogous to the one presented in this case. In State v. Gutierrez, 240 Ariz. 460, 468, ¶ 30 (App.

14

2016), this court addressed as a matter of first impression whether a presumption of judicial vindictiveness arises “whenever a judge imposes a greater sentence after trial than the judge offered during pretrial plea negotiations.” Applying Smith, we held that the “mere imposition of a greater sentence after trial than offered in exchange for a pretrial plea” was not enough to establish a presumption of judicial vindictiveness. Id. But unlike Gutierrez, this case involves not only a disparity between the sentence offered during settlement discussions and the sentence imposed after trial. It also concerns the assigned trial judge’s comments during the improperly conducted settlement discussions advocating for Mendoza to accept the plea offer, and addressing the sentence she would impose if Mendoza rejected the offer and went to trial. Thus, to assess whether a presumption of judicial vindictiveness should arise under these circumstances, we must look beyond Smith and Gutierrez for guidance.

¶27 We find the approach the Florida Supreme Court adopted in Wilson v. State, 845 So.2d 142, 156–57 (Fla. 2003), to be persuasive. Like this case, Wilson concerned whether a presumption of judicial vindictiveness arises when a trial judge participating in settlement discussions spontaneously comments on the sentence a defendant would receive if he or she rejects a plea offer. Id. at 145–46. After examining Pearce and Smith, the Florida Supreme Court adopted a totality of the circumstances analysis “to determine whether a defendant’s constitutional right to due process of law was violated by the imposition of an increased sentence after unsuccessful plea discussions in which the trial judge participated” and thereby raise a presumption of judicial vindictiveness. Id. at 156. The court described several nonexclusive factors courts should consider, including:

(1) whether the trial judge initiated the plea discussions with the defendant . . .; (2) whether the trial judge, through his or her comments on the record, appears to have departed from his or her role as an impartial arbiter by either urging the defendant to accept a plea, or by implying or stating that the sentence imposed would hinge on future procedural choices, such as exercising the right to trial; (3) the disparity between the plea offer and the ultimate sentence imposed; and (4) the lack of any facts on the record that explain the reason for the increased sentence other than that the defendant exercised his or her right to a trial or hearing.

Id. (footnote omitted). In doing so, the court recognized that a flexible, factor-based test was necessary to properly balance “the realistic expectation recognized in Smith that ‘a guilty plea may justify leniency,’

15

and the constitutional imperative that a defendant should not be penalized by the judicial system for not accepting a plea and exercising his or her constitutional right to proceed to trial.” Id. (quoting Smith, 490 U.S. at 802).

¶28 We believe that Wilson provides an appropriate method to review claims of judicial vindictiveness arising out of a trial judge’s participation in settlement discussions outside of the procedures in Rule 17.4. First, reviewing claims of judicial vindictiveness based on the totality of the circumstances is consistent with the analysis employed by the Supreme Court following Pearce and explicitly endorsed by Smith. See McCullough, 475 U.S. at 138 (“Accordingly, in each case, we look to the need, under the circumstances, to ‘guard against vindictiveness . . . .’” (quoting Chaffin v. Stynchcombe, 412 U.S. 17, 25 (1973))); Smith, 490 U.S. at 799 (citing McCullough to support its description of the “reasonable likelihood” test). This approach is also consistent with the test our state has adopted for assessing claims of prosecutorial vindictiveness, a related concept to judicial vindictiveness. State v. Mieg, 225 Ariz. 445, 448–49, ¶ 15 (App. 2010).

¶29 Second, Wilson readily aligns with the approaches adopted by many state and federal courts that have assessed the potential harm caused by judicial participation in settlement discussions. See State v. D’Antonio, 877 A.2d 696, 712–18 (Conn. 2005) (adopting a totality of the circumstances test based in part on Wilson and collecting state and federal cases taking similar approaches); see also, e.g., United States v. Stockwell, 472 F.2d 1186, 1187–88 (9th Cir. 1973) (“[O]nce it appears in the record that the court has taken a hand in plea bargaining, that a tentative sentence has been discussed, and that a harsher sentence has followed a breakdown in negotiations, the record must show that no improper weight was given the failure to plead guilty.”); State v. Baldwin, 629 P.2d 222, 226 (Mont. 1981) (finding “no assurance” in the record that trial court did not impose a harsher sentence “in retaliation for defendant’s insistence on a trial by jury”); State v. Moore, 547 N.W.2d 159, 171–72 (Neb. Ct. App. 1996) (concluding, based on the record before it, that the trial court sentenced defendant more severely because he exercised his right to a jury trial).

¶30 Third, we find the factors enumerated in Wilson particularly useful in assessing such claims within Arizona because Florida, like Arizona, recognizes the value and propriety of limited judicial participation in settlement discussions. See State v. Warner, 762 So.2d 507, 514 (Fla. 2000) (permitting judicial participation in settlement discussions provided specific procedural requirements are met); Ariz. R. Crim. P. 17.4(a)(2).

16

¶31 For these reasons, the Wilson test correctly recognizes and accounts for the careful balance that must be maintained between the leniency a judge may afford defendants who plead guilty and the patently unconstitutional act of penalizing those who choose to exercise their Sixth Amendment right to a jury trial. See Jackson, 390 U.S. at 581; see also Bordenkircher v. Hayes, 434 U.S. 357, 363 (1978) (“To punish a person because he has done what the law plainly allows him to do is a due process violation of the most basic sort.”); Stockwell, 472 F.2d at 1187. Therefore, we conclude that a totality of the circumstances analysis incorporating the considerations described in Wilson should be used to determine whether a trial judge’s improper participation in settlement discussions created a “reasonable likelihood” that a defendant’s sentence was the product of actual vindictiveness. When such a reasonable likelihood exists, a presumption of judicial vindictiveness arises; when it does not, “the burden remains upon the defendant to prove actual vindictiveness.” Smith, 490 U.S. at 799.

¶32 We now apply the analysis to this case. During the settlement discussions, the trial judge repeatedly asserted that she intended to impose a greater sentence than the State’s 9-year plea offer if Mendoza was convicted after a trial. In an attempt to further convince Mendoza to accept the plea offer, the assigned trial judge informed him that by going to trial, he would lose the “mitigation of acceptance of responsibility, remorse, saving taxpayer time, court time, [and] saving money for the courts.” When Mendoza continued to hesitate, the judge promised that he would “not get nine years at trial,” and expressed her belief that Mendoza was “trying to roll the dice” and see if that promise was “true or not.” Despite the judge’s repeated urging, Mendoza rejected the State’s plea offer and proceeded to trial, after which he was sentenced to 12 years’ imprisonment.

¶33 These comments alone, coupled with the 12-year sentence Mendoza ultimately received, are more than enough to raise a presumption of judicial vindictiveness concerning the sentence the trial judge imposed. We do not mean to imply that the trial judge’s sentence was the result of personal animus or hostility towards Mendoza; we agree with the State of Florida that the expression “vindictiveness” in this context “is simply a term of art which expresses the legal effect of a given objective course of action, and does not imply any personal or subjective animosity between the court . . . and the defendant.” Charles v. State, 816 So.2d 731, 734, n.1 (Fla. Dist. Ct. App. 2002). We do not presume the trial judge’s intentions were improper, but the fact remains that she clearly and repeatedly departed from her critical role as a neutral arbiter by: (1) urging Mendoza to accept the plea offer to save him some time in his sentence and avoid a trial she

17

indicated he had no chance of winning; and (2) promising to impose a sentence no lower than the presumptive sentence, 10 years, if Mendoza chose to go to trial. See Wilson, 845 So.2d at 156 (comments that appear to show judge “departed from the role of a neutral arbiter . . . alone may give rise to a presumption of judicial vindictiveness”); id. at 157 (presumption of judicial vindictive arose, in part, because trial judge urged defendant to accept plea and stated he “certainly” would not get a sentence lower than the plea offer); see also Stockwell, 472 F.2d at 1187 (unrebutted presumption of judicial vindictiveness arose by trial court’s statement that defendant would receive two to four more years’ imprisonment than plea offer if he chose to stand trial and was convicted).

¶34 We are also concerned by the assigned trial judge’s statement that Mendoza would lose the “mitigation of acceptance of responsibility, remorse, saving taxpayer time, court time, [and] saving money for the courts” if he elected to exercise his Sixth Amendment right to a jury trial. Arizona’s statutory sentencing scheme permits the superior court to consider “any . . . factor that is relevant to the defendant’s character or background or to the nature or circumstances of the crime and that the court finds to be mitigating.” A.R.S. § 13-701(E)(6). However, the court’s broad discretion in this sphere cannot override the constitutional imperative that a defendant cannot be punished “because he has done what the law plainly allows him to do.” Bordenkircher, 434 U.S. at 363; accord State v. Carriger, 143 Ariz. 142, 162 (1984) (“A defendant is guilty when convicted and if he chooses not to publicly admit his guilt, that is irrelevant to a sentencing determination.”); State v. Trujillo, 227 Ariz. 314, 318, ¶ 14 (App. 2011) (“[I]n sentencing a defendant, a court may not consider . . . lack of remorse or failure to admit guilt.”). Some of the mitigating factors the trial judge listed are unique to a defendant’s decision to plead guilty, and thus likely fall under the umbrella of circumstances justifying leniency as consideration for a guilty plea. See Smith, 490 U.S. at 802. But “acceptance of responsibility” and “remorse” are not mitigating factors unique to that decision.

¶35 In sum, after reviewing the totality of the circumstances and applying the factors outlined in Wilson, we hold Mendoza has shown a reasonable likelihood that his sentence was more likely than not the product of actual vindictiveness, thus raising a presumption of judicial vindictiveness. And because we presume the court violated Mendoza’s due-process rights by imposing an increased sentence to penalize him for exercising his Sixth Amendment right to a jury trial, we also hold that he has established that the trial judge’s erroneous participation in settlement discussions, in this case, constituted fundamental error. See State v. Thues,

18

203 Ariz. 339, 340, ¶ 4 (App. 2007) (“Imposition of an illegal sentence constitutes fundamental error.”).

¶36 In reaching these conclusions, we again emphasize that we would not have found error arising from the March 2018 settlement discussions, let alone fundamental error, if they had occurred in accordance with Rule 17.4(a)(2); that is, either before a judge other than the assigned trial judge or before the assigned trial judge with the parties’ consent or subsequent ratification. The constitutional issues raised by judicial participation in settlement discussions are typically “avoided altogether” when the prophylactic rules designed to safeguard defendants’ due-process rights in the plea-bargaining process are correctly utilized. Gutierrez, 240 Ariz. at 469, ¶ 34. When followed, Rule 17.4(a)(2) either provides a defendant the assurance that the judge involved in settlement discussions will play no role in the trial or sentence, or makes a record of the defendant’s knowing and voluntary waiver of that assurance.

¶37 Thus, within the process the rule establishes, judges may engage in a candid dialogue with the parties concerning any topic useful to facilitating settlement, including the strength of a prospective plea offer, the sentence likely to be imposed after trial, and any mitigating or aggravating factors at issue in the case, without fear of generating the perception that they have departed from their role as a neutral arbiter. When a trial judge becomes involved in settlement discussions outside of that framework, however, the judge risks being perceived as an advocate for a particular outcome. No matter how well-intentioned the judge may have been in engaging in those discussions, they can subject the judge to serious claims challenging his or her impartiality, including claims of judicial vindictiveness. Here, we believe the trial judge’s statements during the settlement discussions realized those concerns. Consequently, this case did not involve, as the State contends, a “mere technical violation of a prophylactic rule,” but an error going to the heart of the very due-process concerns the rule was originally designed to protect. See Jordan, 137 Ariz. at 509.

C. Because Objective Evidence in the Record Does Not Rebut the Presumption of Judicial Vindictiveness Raised by the Trial Judge’s Participation in Settlement Discussions, Mendoza has Demonstrated Prejudice.

¶38 Having found fundamental error, we must now examine whether Mendoza has established prejudice. “Establishing prejudice from fundamental error varies depending on the nature of the error and the

19

unique case facts.” Escalante, 245 Ariz. at 144, ¶ 29. Although Mendoza still “bears the burden of persuasion” at this step, id. at 142, ¶ 21, because a presumption of judicial vindictiveness has been raised concerning the sentence imposed by the assigned trial judge, the burden of production shifts to the State “to produce affirmative evidence on the record to dispel the presumption.” Wilson, 845 So.2d at 156. The State may do so by pointing to “objective information . . . justifying the increased sentence.” McCullough, 475 U.S. at 142 (quoting Goodwin, 457 U.S. at 374).

¶39 Based on the record before us, the State has not successfully justified the 12-year sentence the trial judge imposed to the extent necessary to rebut the presumption raised by her statements at the settlement conference. We do not dispute the State’s assertions that the aggravators considered by the court were proper and that Mendoza’s sentence fell within the range prescribed by law. But we cannot say the same for the judge’s statements at the sentencing hearing concerning the mitigators present in Mendoza’s case. During the hearing, the court made the following statements concerning the mitigators of acceptance of responsibility and remorse:

There is no acceptance of responsibility. There is no remorse. There is none of the mitigators you get when you take a plea. I explain to people that . . . means that just doesn’t exist in that pile any longer.

These are the exact sentiments the court improperly made at the March 2018 hearing, and we cannot ignore the possibility that by repeating them at the sentencing hearing, the assigned trial judge was making good on her promise to categorically refuse to consider mitigating evidence because Mendoza elected to go to trial. That the judge followed these statements by saying there was “no punishment for taking a case to trial” and that she “hear[s] a lot more in the trial” is of little comfort, especially when the judge thereafter failed to acknowledge acts or statements by Mendoza indicating an acceptance of responsibility or remorse such as starting a 12-step program to address his alcohol issues.

¶40 It remains unclear whether the increased sentence was imposed based on a proper balancing of the aggravators and mitigators present in Mendoza’s case or instead “based on the fulfillment of the previous threat[s]” made by the assigned trial judge regarding both the mitigating evidence she would consider and the sentence Mendoza would ultimately receive were he to reject the plea offer and go to trial. Jordan, 137 Ariz. at 509. And while not binding on the trial court, its rejection of the

20

presentence report’s recommendation of the presumptive sentence likewise supports Mendoza’s claim of prejudice. Such is the inherent danger of judicial involvement in settlement discussions outside the confines of Rule 17.4(a)(2) and the significance of the assigned trial judge’s departure from her role as a neutral arbiter at the hearing. Accordingly, we must conclude the presumption of judicial vindictiveness has not been rebutted and that Mendoza has established prejudice justifying relief here.

¶41 We turn briefly to address the question of the appropriate remedy for cases involving an unrebutted presumption of judicial vindictiveness in sentencing. In Wilson, the court reasoned that resentencing before a different judge was necessary to “place[] the defendant in the same position he or she would have been in if the plea discussions had never occurred—before a neutral arbiter to receive a lawful sentence.” 845 So.2d at 159. We agree with that rationale and hold the proper remedy, in this case, is a resentencing before a different judge.

CONCLUSION

¶42 We affirm Mendoza’s conviction but vacate his sentence and remand for resentencing before a different judge.

AMY M. WOOD • Clerk of the Court FILED: AA

21