24-2602-pr Caballero v. Miller
UNITED STATES COURT OF APPEALS FOR THE SECOND CIRCUIT
SUMMARY ORDER
RULINGS BY SUMMARY ORDER DO NOT HAVE PRECEDENTIAL EFFECT. CITATION TO A SUMMARY ORDER FILED ON OR AFTER JANUARY 1, 2007, IS PERMITTED AND IS GOVERNED BY FEDERAL RULE OF APPELLATE PROCEDURE 32.1 AND THIS COURT’S LOCAL RULE 32.1.1. WHEN CITING A SUMMARY ORDER IN A DOCUMENT FILED WITH THIS COURT, A PARTY MUST CITE EITHER THE FEDERAL APPENDIX OR AN ELECTRONIC DATABASE (WITH THE NOTATION “SUMMARY ORDER”). A PARTY CITING TO A SUMMARY ORDER MUST SERVE A COPY OF IT ON ANY PARTY NOT REPRESENTED BY COUNSEL.
At a stated term of The United States Court of Appeals for the Second Circuit, held at the Thurgood Marshall United States Courthouse, 40 Foley Square, in the City of New York, on the 2nd day of April, two thousand twenty-five.
PRESENT: GERARD E. LYNCH, BETH ROBINSON, ALISON J. NATHAN, Circuit Judges. _________________________________________
ANDREW CABALLERO,
Petitioner-Appellee,
v. No. 24-2602-pr
CHRISTOPHER MILLER, WARDEN/SUPERINTENDENT OF THE COMSTOCK CORRECTIONAL FACILITY,
Respondent-Appellant. *
* The Clerk’s office is respectfully directed to amend the caption as reflected above. _________________________________________
FOR PETITIONER-APPELLEE: SARAH KUNSTLER, Law Office of Sarah Kunstler, Brooklyn, NY.
FOR RESPONDENT-APPELLANT: DANIELLE M. O’BOYLE, Assistant District Attorney (John M. Castellano, Assistant District Attorney, on the brief), for Melinda Katz, District Attorney for Queens County, Kew Gardens, NY.
Appeal from a judgment of the United States District Court for the Eastern
District of New York (Korman, Judge).
UPON DUE CONSIDERATION WHEREOF, IT IS HEREBY ORDERED,
ADJUDGED, AND DECREED that the judgment entered on September 20, 2024,
is REVERSED.
The State of New York appeals from the district court’s judgment granting
Petitioner-Appellee Andrew Caballero’s petition for a writ of habeas corpus
pursuant to 28 U.S.C. § 2254. In granting Caballero’s petition, the district court
concluded that Caballero received ineffective assistance of trial and appellate
counsel and that the state court unreasonably applied clearly established federal
law in concluding otherwise. We disagree and REVERSE. We assume the parties’
familiarity with the underlying facts, procedural history, and arguments on
appeal, to which we refer only as necessary to explain our decision. I. BACKGROUND
On February 1, 1995, Jason Kollman was stabbed fifteen times, was pushed
off the rooftop of a building in New York City, and landed on a fifth-floor fire
escape, where he bled to death. The crime remained unsolved until 2012, when a
New York grand jury indicted Caballero after two witnesses, Oscar Balarezo, a
friend of Caballero, and Nadia Sierra, Caballero’s girlfriend at the time of the
murder, identified Caballero as the killer. In September 2012, Caballero retained
Attorney Robert DiDio to represent him.
In late May and early June 2014, several weeks before trial, the prosecution
disclosed Rosario 1 material, including police interview reports from multiple
witnesses. Those witnesses included Michael Lugo, Amanda Pabon, Angel
Hernandez, Javier Peralta, Kim Pietrovito, and Wayne Dewall.
In the interview reports, several of the witnesses described observing or
hearing about a fight between Kollman and Lugo shortly before Kollman’s death.
Specifically, Kollman’s close friend Peralta reportedly told police that he
witnessed a “fist fight” between Kollman and Lugo that evening. App’x at 2253.
1 In People v. Rosario, 9 N.Y.2d 286 (1961), the New York Court of Appeals held that the prior statement of a witness should be made available to the defense for possible use on cross- examination.
3 Lugo’s girlfriend Pietrovito reported that Lugo told her on the night of the murder
that he had gotten into a fight with Kollman earlier in the evening. Id. at 2255.
And when Lugo himself spoke to police, he admitted on two separate occasions
that he and Kollman had gotten into a fight—though he originally told police the
fight was over a girl, id. at 2240, and only years later admitted that it involved a
dispute over money related to a drug transaction, id. at 2243.
Some of the witnesses also indicated that they thought or had heard from
others that Lugo, not Caballero, was the murderer. For example, in October 1999,
Hernandez, a friend of Lugo, reportedly told police that “everybody knows” Lugo
killed Kollman, and that he had heard that information from “Amanda,”
Kollman’s former girlfriend. Id. at 2237. Amanda Pabon, Kollman’s ex-girlfriend,
told police in November 2011 that she “believe[d] that Mr. Lugo may have had
something to do with [Kollman’s] death” based on a tense interaction she had
witnessed between the two on the evening of the murder. Id. at 2252. Peralta
reportedly told police that “Chris,” a neighbor of Caballero, had seen Kollman,
Lugo, and someone named “Anthony” enter Caballero’s apartment building that
night, and, based on that information, he believed “that Anthony or Lugo, or both”
4 was responsible for Kollman’s death. 2 Id. at 2253–54. (Peralta himself was accused
of being the killer by Dewall, a friend of Kollman who reported to detectives in
1995 that Kollman was killed due to a relationship with Pabon, who was Peralta’s
girlfriend at the time. Id. at 2235.)
On July 21, 2014, the day before jury selection was to begin in Caballero’s
trial, the prosecution disclosed additional materials that included reports of
interviews with Nadia Sierra, Oscar Balarezo, Monique Moses, Christian
Cavalieri, and Samantha Newmark. The materials regarding Sierra and Balarezo
disclosed that they had told police, among other things, that Caballero was
Kollman’s killer. Moses’s report recounted a conversation with a distraught Sierra
in which Sierra reported that she had ended her relationship with Caballero, asked
Moses to watch her child because Sierra feared for the child’s safety, and said she
could not believe what Caballero had done. Meanwhile, Newmark’s report said
that she told police that she had heard from Kollman’s friend Chris that Kollman
“got into a fight on the street with Mike Lugo” before going into a building with
“Andy.” Id. at 2246. And Cavalieri’s report noted his observation that when he
2 Caballero correctly notes in his briefing that his first name is Andrew, not Anthony. However, it is perhaps worth noting that according to a police report documenting Christian Cavalieri’s interview, in which Cavalieri is referred to as “Chris,” Cavalieri identified Caballero as one of the two men who entered the apartment building with Kollman. App’x at 2232.
5 ran into Kollman that evening, Kollman’s “lip and face looked swollen” and he
thought Kollman may have had a broken jaw when Kollman went into the
apartment building with “two males, both [H]ispanic.” Id. at 2231.
The next day, after receiving these disclosures, DiDio sought an
adjournment so he could contact the witnesses from both the July disclosures and
the earlier May and June disclosures to explore avenues for impeaching Balarezo
and Sierra. At a hearing on the motion, DiDio explained that, because Balarezo
and Sierra had previously refused to speak to a defense investigator, the potential
import of the other witnesses’ statements about them had become clear only the
day before. Specifically, he pointed to (1) Lugo’s statement placing Balarezo at the
scene of the fight between Lugo and Kollman (which Balarezo claimed only to
have heard about second hand); (2) Peralta’s report that he had heard that Kollman
had gone to the apartment rooftop with Lugo, not Balarezo, as Balarezo claimed;
and (3) Moses’s description of Sierra’s statements, which DiDio argued
contradicted Sierra’s statements to police that she had told Moses “everything that
happened that night,” id. at 38.
The trial court denied the request, emphasizing that Caballero knew both
Balarezo and Sierra and his counsel could have attempted to investigate them—
6 and interview several other witnesses about them—well before receiving their
Rosario materials. Mindful of Sierra’s claims that Caballero had threatened her
following the murder and the fact that her name and expected incriminating
testimony had now been disclosed, the court ordered the trial to go forward.
DiDio enlisted an associate, Danielle Muscatello, who rushed to investigate
the additional witnesses as the trial began. She enlisted an investigator, Edward
Dowd, to run background checks on Sierra and Balarezo and locate Lugo, Moses,
and Peralta. Although Dowd was able to contact Moses by phone, he did not make
contact with Lugo or Peralta or run background checks on Sierra or Balarezo.
At trial, the prosecution presented no forensic or physical evidence tying
Caballero to the crime and relied heavily on the testimony of Balarezo and Sierra.
DiDio’s trial strategy focused primarily on impeaching them by highlighting
inconsistencies in their stories. DiDio also elicited testimony about the fight
between Kollman and Lugo in the hour before Kollman’s death, and in his
summation DiDio suggested that the jury should consider the “friction” between
Lugo and Kollman when deciding its verdict. App’x at 861-62. He did not
otherwise present Lugo as a potential alternative perpetrator at trial.
7 After deliberating for three days, the jury convicted Caballero, and the trial
court sentenced him to an indeterminate prison term of twenty-five years to life.
Muscatello, who represented Caballero on direct appeal, did not argue in that
appeal that DiDio had provided ineffective assistance of counsel. In 2016, the
Appellate Division affirmed Caballero’s conviction.
In June 2017, Caballero filed a pro se application for a writ of error coram
nobis, claiming ineffective assistance of appellate counsel for Muscatello’s failure
to challenge on direct appeal DiDio’s effectiveness. Specifically, Caballero argued
that Muscatello labored under a conflict of interest that prevented her from
bringing a meritorious claim of ineffective assistance. The Appellate Division
summarily denied that petition.
In September 2018, Caballero filed a pro se habeas corpus petition in the
Eastern District of New York. Court-appointed counsel then filed a supplemental
brief asserting ineffective assistance of both trial and appellate counsel. In 2021,
the district court stayed the petition so that Caballero could litigate the ineffective-
assistance-of-trial-counsel claim in state court.
At a hearing in state court under New York Criminal Procedural Law
§ 440.10, Caballero called Muscatello; Irwin Blye, an investigator who worked
8 with DiDio shortly after he was retained by Caballero in September 2012; Dowd,
the investigator enlisted by Muscatello to investigate potential witnesses on the
eve of trial; and Nelson LaSalle, a private investigator retained by Caballero’s
habeas counsel in February 2023. Because DiDio died before the hearing, the court
allowed the state to introduce his written contemporaneous notes.
Notably, LaSalle testified that he had been able to locate and speak to almost
all the potential witnesses identified in the Rosario material that DiDio received
before trial. Over the course of approximately 40 hours of work, he had
successfully found and spoken with Christian Cavalieri, Monique Moses,
Samantha Newmark, Amanda Pabon, Javier Peralta, Kimberly Pietrovito, and
Nadia Sierra. Though he had also located Lugo and Balarezo, he explained that
Lugo had not been willing to speak to him and Balarezo had not returned his call.
Wayne Dewall, the final witness he attempted to locate, had died. 3
During the hearing, the § 440.10 court did not allow LaSalle to provide
hearsay testimony about the contents of his conversations with the witnesses, and
none of the witnesses testified or provided written affidavits in support of
3 LaSalle had not been asked to locate witness Angel Hernandez.
9 Caballero’s case. 4 As a result, no new information was provided about the
witnesses’ recollections of the night of the murder or their potential testimony if
they had been called at trial.
After the hearing, the Queens County Supreme Court denied Caballero’s
motion, concluding that Caballero established neither deficient representation by
DiDio nor prejudice from any deficiency. The court found that DiDio’s notes
indicated that he had conducted some investigation of Caballero’s case and noted
that Caballero “has failed to indicate what, if any, relevant information [the]
individuals [whom DiDio failed to contact] would have been able to provide.”
App’x at 2347. The court concluded that Caballero had also failed to show that
DiDio’s limited investigation compromised the fairness of the litigation.
In 2024, the Eastern District of New York granted Caballero’s writ of habeas
corpus under 28 U.S.C. § 2254. The court concluded that DiDio provided deficient
representation by failing to investigate leads suggesting Lugo as a potential
alternative suspect, reasoning that there was no conceivable strategic basis for not
investigating this possible defense. The court concluded that the cumulative effect
4Caballero’s counsel did not make an offer of proof as to what LaSalle would have said had hearsay testimony been permitted.
10 of this failure to investigate, along with certain other trial errors, established
prejudice.
II. STANDARD OF REVIEW
“The grant or denial of habeas corpus is reviewed de novo, and the
underlying findings of fact are reviewed for clear error.” Clark v. Perez, 510 F.3d
382, 389 (2d Cir. 2008). 5 Pursuant to the Antiterrorism and Effective Death Penalty
Act of 1996 (“AEDPA”), a habeas petition “adjudicated on the merits in State
court” can be granted only if the state court adjudication “resulted in a decision
that was contrary to, or involved an unreasonable application of, clearly
established Federal law, as determined by the Supreme Court of the United
States.” 28 U.S.C. § 2254(d).
For ineffective assistance of counsel claims, “the merits of such claims are
squarely governed by the Supreme Court’s holding in Strickland v. Washington.”
Lindstadt v. Keane, 239 F.3d 191, 198 (2d Cir. 2001). To establish a claim for
ineffective assistance of counsel, a defendant must show that (1) “counsel’s
representation fell below an objective standard of reasonableness” and (2) “there
is a reasonable probability that, but for counsel’s unprofessional errors, the result
5 In quotations from caselaw and the parties’ briefing, this summary order omits all internal quotation marks, footnotes, and citations, and accepts all alterations, unless otherwise noted.
11 of the proceeding would have been different.” Strickland v. Washington, 466 U.S.
668, 687–88, 694 (1984).
Because “[t]he standards created by Strickland and § 2254(d) are both highly
deferential, . . . when the two apply in tandem, review is doubly so.” Harrington
v. Richter, 562 U.S. 86, 105 (2011). A state court thus has significant “latitude to
reasonably determine that a defendant has not satisfied [the] standard.” Sexton v.
Beaudreaux, 585 U.S. 961, 968 (2018). In accordance with that deferential standard
of review, the question faced by a federal court considering whether to grant a
habeas petition “is not whether [it] believes the state court’s determination under
the Strickland standard was incorrect but whether that determination was
unreasonable—a substantially higher threshold.” Knowles v. Mirzayance, 556 U.S.
111, 123 (2009) (emphasis added).
III. TRIAL COUNSEL
Caballero’s ineffective assistance of trial counsel claim hinged primarily on
two errors: (1) DiDio’s failure to investigate and attempt to interview Balarezo and
Sierra and (2) DiDio’s failure to investigate Lugo as an alternative suspect.
Caballero also points to DiDio’s apparent confusion about the layout of the
apartment building where Kollman was killed, which led DiDio to ask some
12 questions on cross-examination that he eventually had to concede before the jury
were based on a mistaken premise.
The State contends that the district court was insufficiently deferential to the
state court’s ruling on Caballero’s ineffective assistance claim. In particular, the
State emphasizes that the lack of testimony or affidavits from the witnesses
Caballero says DiDio should have interviewed “goes to the heart of the prejudice
analysis.” Appellant’s Br. at 59. Without such evidence, the State argues,
Caballero failed to demonstrate what relevant information those witnesses would
have provided, or even that they would have been willing to testify at trial. We
agree. Even if any reasonable jurist would necessarily conclude that DiDio’s
failure to investigate these critical leads was objectively unreasonable and
constituted deficient performance, we cannot conclude the state court
unreasonably applied Strickland when it held that Caballero was not prejudiced by
any deficiency in DiDio’s representation.
Caballero bears the burden of affirmatively proving prejudice and must
demonstrate “a reasonable probability that, absent the errors, the factfinder would
have had a reasonable doubt respecting guilt.” Strickland, 466 U.S. at 695. He
argues, and the district court agreed, that the relative weakness of the case against
13 him combined with DiDio’s failure to conduct a proper investigation necessarily
prejudiced him, such that he was not required to proffer evidence about what a
proper investigation would have revealed.
But the cases cited for that proposition do not support such a broad-brush
assessment. In Lindstadt, for example, the prejudice entailed by trial counsel’s
failure to investigate was clear: the petitioner identified witnesses who were aware
of and, had they been investigated and called to testify, could have testified to
information that would have deeply undermined the prosecution’s key witness.
239 F.3d at 205 (“The probation officers, as outside, impartial observers of the
Lindstadt household, could have provided essential corroboration for Lindstadt’s
testimony that his wife was vindictive; and their testimony would have put the
case in the context of her previous attempts to get her husband jailed.”).
The same is true of Schulz v. Marshal, 345 F. App’x 627 (2d Cir. 2009), on
which the district court relied for the proposition that failure to investigate the
potential witnesses alone might amount to prejudicial ineffective assistance.
There, we noted that trial counsel failed to interview a key eyewitness, who, when
questioned by new counsel after the trial, identified a perpetrator other than the
defendant. Schulz, 345 F. App’x at 629. Given other weaknesses in the state’s case,
14 we determined that the petitioner had shown a reasonable probability that
investigating this witness pretrial would have altered the outcome of the case –
and the state court had unreasonably applied Strickland when it held otherwise.
Id.
Here, in contrast, Caballero presented no evidence about what information
a more thorough investigation would have garnered. Caballero’s investigator,
LaSalle, contacted and spoke to virtually all the individuals identified in the
Rosario material. Despite those efforts, nothing in the record of the § 440.10 hearing
suggests that any of the witnesses would have provided additional testimony or
information that DiDio could have used to more effectively impeach Balarezo or
Sierra or to implicate Lugo as an alternative perpetrator if he had interviewed
them before trial. None of the witnesses testified (or provided affidavits stating)
that they would have provided information helpful to Caballero’s defense.
And Caballero has not argued—before the district court or before us—that
the sole reason he has failed to present such additional information is the state
court’s evidentiary ruling prohibiting LaSalle from providing hearsay testimony
about his conversations with the witnesses. Nor has Caballero argued that that
15 ruling itself was constitutionally infirm. Indeed, he has not even proffered what
LaSalle would have testified, absent that ruling.
That dearth of evidence is particularly significant given the substantial
amount of hearsay, speculation, and information otherwise unhelpful to the
defense that is apparent in the police reports upon which Caballero’s ineffective
assistance claims rely. Hernandez, Peralta, and Newmark all told police that they
had merely heard information from others—namely “Amanda” and “Chris”—that
made them suspect Lugo might have been Kollman’s murderer. And the direct
witnesses inculpated Caballero more than they inculpated Lugo. Even though
Pabon herself saw an interaction between Kollman and Lugo that raised her
suspicions, her interview reports also indicate that Caballero was the last person
with whom she had seen Kollman alive on the night of the murder, that Caballero
had been involved in drug-dealing with Kollman and Lugo, and that she had seen
Caballero and Kollman walking into the courtyard of the apartment building that
night—but only Caballero and Balarezo walking out approximately 30 minutes
later. Likewise, Cavalieri identified Caballero as one of the two Hispanic men
whom he saw entering the building with Kollman. Thus, nothing in either the
Rosario material or the post-trial record gives any reason to believe that any of
16 those witnesses, had they been interviewed by DiDio and called to testify at the
trial, could or would have provided admissible evidence helpful to Caballero’s
defense. 6
In short, Caballero has failed to show “a substantial, not just conceivable,
likelihood of a different result” had DiDio investigated the individuals Caballero
identifies in his habeas petition. Cullen v. Pinholster, 563 U.S. 170, 189 (2011). As a
result, the state court did not unreasonably apply Strickland when it concluded
Caballero was not prejudiced by any of DiDio’s asserted deficiencies. We
therefore reverse the district court’s grant of Caballero’s § 2254 petition on that
basis.
IV. APPELLATE COUNSEL
The district court concluded that the state court unreasonably applied
Strickland when it concluded Muscatello was not ineffective because DiDio’s
6 Nor does the record suggest that further investigation would have developed additional impeachment evidence about trial witnesses Balarezo and Sierra. Sierra agreed to speak to LaSalle in 2023; yet just like the witnesses from the Rosario materials, Caballero has not offered any indication that she provided information that would have been useful to his defense. Balarezo, in turn, declined to return LaSalle’s call, just as he had declined to be interviewed by defense counsel and investigators before trial. Both were effectively cross examined by DiDio; indeed, Caballero’s and the district court’s argument about the putative weaknesses of the case at trial is based entirely on points developed on cross-examination and arguments (such as the absence of physical evidence) presented by DiDio to the trial jury.
17 ineffectiveness was obvious on the face of the record, rendering Muscatello’s
failure to raise the argument on direct appeal constitutionally deficient. We
disagree.
“New York courts uniformly hold that where . . . an ineffective assistance of
counsel claim turns on facts that are outside of the trial-court record, the claim
must be brought in collateral proceedings, not on direct appeal.” Pierotti v. Walsh,
834 F.3d 171, 178 (2d Cir. 2016). A robust evaluation of whether DiDio was
ineffective for failing to better investigate this case would require additional
evidence beyond the trial record, including the Rosario materials, testimony from
DiDio about his potential strategy, and testimony about what evidence could have
been introduced at trial. Thus, the state court did not unreasonably apply
Strickland when it concluded Muscatello was not ineffective for failing to raise this
issue on direct appeal.
Moreover, even if New York law did not clearly require ineffective
assistance claims like Caballero’s to be raised in collateral proceedings, rather than
on direct appeal, his claim of ineffective assistance of appellate counsel would still
fail. Ineffective assistance claims addressing conduct by an appellate attorney are
governed by the same Strickland performance and prejudice standard as claims
18 addressing conduct by trial counsel. See Lynch v. Dolce, 789 F.3d 303, 311 (2d Cir.
2015). “To establish prejudice in the appellate context, a petitioner must show that,
had his claim been raised on appeal, there is a reasonable probability that it would
have succeeded before the state’s highest court.” Id. But, as noted above,
Caballero has failed to demonstrate a reasonable probability that he was
prejudiced by DiDio’s potential deficiencies. As a result, he has also failed to
demonstrate a reasonable probability that Muscatello would have succeeded in
showing that same required prejudice on appeal.
* * *
Accordingly, for the reasons stated above, the District Court’s judgment is
REVERSED.
FOR THE COURT: Catherine O’Hagan Wolfe, Clerk of Court