By Matthew B. Keller
Last week, HCB’s own Alan Lewis received a COA grant from the Second Circuit in Umali v. Heath, No. 12-3243 (2d Cir. COA granted Dec. 18, 2012). Not one to toot his own horn, I will trumpet for him – a Circuit COA grant is rare enough these days that it’s important to keep track of all of them. This one is particularly noteworthy because it gives the Second Circuit a chance to consider one important, but largely ignored aspect of the Supreme Court’s Lafler v. Cooper opinion from last term – the boundary between the contrary to and unreasonable application clauses of AEDPA.
The case began as a New York State murder prosecution against Isaias Umali for killing Dana Blake during a 2003 fight at an East Village club. Blake, a bouncer, was in the process of violently removing Umali’s friend, Jonathan Chan, from the club when Umali stabbed Blake in the thigh/groin, severing Blake’s femoral artery. Surgery to repair the artery failed and Blake died the following day.
Following a trial, Umali was convicted of First Degree Manslaughter and sentenced to 17 years in prison. In his state court appeals he raised several issues relating to his right to present a defense, the trial court’s exclusion of lay and expert witness testimony, as well as the issue on which the Second Circuit granted a COA – the court’s burden shifting instruction on justification.
The Trial Court’s Justification Instruction
Umali ’s defense was that, at the time of the stabbing, he believed that Blake was using deadly physical force against Chan. Both defense and prosecution witnesses testified that Blake (a much larger man than Chan) held Chan by the neck or throat and that Chan unsuccessfully tried to free himself from Blake’s grasp. The defense argued that Chan could not breathe and was therefore in grave danger.
In New York, deadly force can be justified where someone reasonably believes that another person is using or is about to use deadly physical force. For those who haven’t cracked a NY Penal Law since the Bar Exam, you’ll recall that this defense has both “subjective” and “objective” components: the defendant must subjectively believe at the time that he is acting to stop deadly physical force upon himself or another, and that belief must be objectively reasonable under the circumstances.
Where this gets tricky is in the burden of proof. Even though justification is raised by the defendant, in New York it remains the prosecutor’s burden to prove that a defendant was NOT justified in using deadly physical force. And indeed, in Umali’s case, the court began by correctly instructing the jury that it was, “the burden of the prosecution to disprove [justification] beyond a reasonable doubt.” (The court also mentioned that, in general, “the defendant has no burden of proof”). However, the court ran into trouble when it broke down the burden of proof as to each prong (subjective and objective) of justification, stating:
If the evidence convinces you beyond a reasonable doubt that deadly force was necessary to prevent the imminent use – that the defendant believed that deadly physical force was necessary to prevent the imminent use of deadly physical force you must still find the second test, which is the objective test, were defendant’s beliefs reasonable under an objective standard.
If you didn’t catch it, the problem with this explanation is that, by suggesting that the jury had to be “convince[d] beyond a reasonable doubt that . . . the defendant believed that deadly physical force was necessary,” the court shifted the burden of proof on the subjective prong to the defendant. Umali’s trial attorney (HCB’s Alan Lewis) caught it, and asked the judge to re-charge the jury and explain its previous error. The court refused, maintaining that it “correctly said” the justification charge.
A burden-shifting charging error is a big deal. The criminal justice system relies on juries to give effect to major constitutional rights, including the one at issue here: the requirement that the State prove guilt beyond a reasonable doubt. Rights don’t get much more fundamental to American justice than that. Therefore, when a court gives a constitutionally defective instruction, the Supreme Court has held that it is not enough that other parts of the jury charge may have instructed correctly on the same subject. Instead, the court is required to “explain” the defective instruction: “Language that merely contradicts and does not explain a constitutionally infirm instruction will not suffice to absolve the infirmity.” Francis v. Franklin, 471 U.S. 307, 322 (1985).
Appellate Court and District Court Habeas Review
Umali took this argument to New York’s appellate courts, which rejected it. The NY Court of Appeals agreed with Umali that the instruction on the subjective element of justification had impermissibly shifted the burden of proof onto him. However, it also held that the trial court corrected its mistake by repeatedly referring to the correct burden of proof in reference to the overall justification defense and to the case in general. “In light of these repeated references to the correct legal standard,” the Court of Appeals concluded, “the instructions as a whole could not have misled the jury regarding the applicable burden of proof. Thus, a supplemental charge explicitly disavowing the court’s single misstatement of law was not required in this case, although it certainly would have been preferable had the court realized that it misspoke.”
If you’re reading closely you might be asking yourself, “what about that Franklin case he just mentioned? Doesn’t that require an explanation of the error?” It sure seems to me that it does. But you wouldn’t know that from reading the Court of Appeals’ decision. Even though Umali emphasized the Franklin holding – it is literally the first case cited in Umali’s Court of Appeals’ brief – the opinion refers to Franklin ZERO times. And neither of the two cases the Court of Appeals relied upon for its expression of the legal standard refers to Franklin. See People v Drake, 7 N.Y.3d 28 (2006); People v Fields, 87 N.Y.2d 821 (1995).
Umali petitioned the SDNY for habeas review. The petition was referred to a federal Magistrate Judge, who recommended that District Judge Rakoff deny it. See Umali v. Heath, No. 10 Civ. 1787 (S.D.N.Y. Report & Recommendation filed Oct. 13, 2011) (Gorenstein, M.J.) (“R & R”). Judge Rakoff adopted the R & R and also denied a COA.
To answer the jury instruction question, the R & R applied a modified form of the three-part test of Jackson v. Edwards, 404 F.3d 612, 621 (2d Cir. 2005), asking the following three questions: (1) was the charge incorrect under state law?; (2) if so, was the giving of the faulty charge a denial of due process?; and (3) if so, did the state court’s contrary conclusion constitute an unreasonable application of clear Supreme Court law?
Before getting to the answers, you have to wonder whether the Magistrate asked the right questions – particularly the third, applying AEDPA’s deferential standard of review under the unreasonable application prong. Last term in Cooper, the Supreme Court held that state court decisions that fail to apply controlling Supreme Court precedent to a constitutional claim do not get AEDPA deference because they are “contrary to” clearly established federal law as opposed to an unreasonable application of that law. 132 S. Ct. 1376, 1390. The state court decision at issue in Cooper failed to apply Strickland v. Washington to assess an ineffective assistance of counsel claim. Here, there is no doubt that the New York Court of Appeals failed to apply Franklin to Umali’s jury instruction claim, because the court never mentioned the case. So, whence AEDPA deference? By asking the third question above, the district court simply assumed that AEDPA deference would apply. A threshold question for the Circuit will be how to deal with the lower court’s unconsidered application of AEDPA deference in light of Cooper.
As for the first two questions, the Magistrate Judge answered yes and no, respectively. The subjective prong language was incorrect under state law because it incorrectly shifted the burden on to the defendant. However, the faulty charge did not result in a denial of due process because in the context of the entire charge it did not deny Umali of a fair trial.
Unlike the NY Court of Appeals, the Magistrate Judge applied Franklin; but then he distinguished it from Umali’s case. Sure, the court said, under Franklin you cannot correct an erroneous burden-shifting instruction merely by using more “general [correct] instructions” on burden of proof to “contradict” the bad instruction on the subjective prong of justification. But according to the district court, that’s not what happened here: the measures taken by Umali’s judge were much more thorough and “direct” than those taken in Francis. After parsing the language of the jury charge and the trial judge’s responses to jury questions, the Magistrate Judge found that the trial court, “twice made clear that it was the prosecution’s responsibility to prove the absence of [a justification] defense, which necessarily included both the subjective and objective prongs.” Therefore, the Magistrate concluded, “[t]hese instructions thus constituted a far more direct correction of the incorrectly-given instruction than had occurred in Francis.”
I’m not sure what Judge Rakoff – an excellent and thoughtful judge – was thinking when he denied COA on this. It is a fair reading of Franklin that a trial court must correct an unconstitutional instruction by, at a minimum, referencing the bad instruction and then explaining why it is incorrect. Clearly that did not happen here. In addition, none of the judge’s correct instructions were as specific as the burden-shifting one. That is, the burden shifting occurred in the only instruction in which the trial judge dealt specifically with the subjective element of justification.
The reasoning behind the R & R seems to be that Umali’s case is different from Franklin because here the correct instructions on burden of proof related to a specific claim or defense (i.e. justification) and therefore it’s okay to expect a lay jury to reason that the correct instruction should override an erroneous instruction applying to an element of that claim or defense. In Franklin, the correct instructions were given only in the abstract, and therefore it was not fair to expect a jury to apply them to override a contradictory instruction as to an element of a claim or defense.
I’m dubious. As a legal matter, Franklin says that abstract or global instructions cannot override contradictory specific instructions. Yet both the NY Court of Appeals and the Magistrate Judge specifically relied on such global instructions (as well as the “more specific” ones relating to justification) to defend the legitimacy of the overall charge. That reliance was misplaced under any reading of Franklin.
But even on a more basic level, the Magistrate’s reasoning makes an unwarranted assumption about the legal reasoning skills of lay jurors. Trial judges should charge each jury as if it knows nothing about the law. If a jury were made up entirely of lawyers (or even law students) it might be reasonable to assume that such a jury could easily reason from a defense to an element of that defense. These are precisely the sort of reasoning skills that those in (or training for) the legal profession regularly employ. But I’m not sure why this level of deductive reasoning is any easier for a lay jury than that deemed too difficult for a jury by the Supreme Court's Franklin opinion.
The Second Circuit did the right thing by granting COA. That will enable it to consider the merits as well as the application of AEDPA deference in light of Cooper. I’m hoping that Cooper begins a shift away from the sometimes unexamined assumption that every state court decision receives deference under AEDPA. State court decisions that fail to apply the correct federal standard should not receive that (or any) level of deference.