Az v. Shinseki

United States Court of Appeals for the Federal Circuit ______________________ AZ, Claimant-Appellant, v. ERIC K. SHINSEKI, Secretary of Veterans Affairs, Respondent-Appellee. ______________________ 2012-7046 ______________________ Appeal from the United States Court of Appeals for Veterans Claims in No. 10-2393, Judge Lawrence B. Hagel. ---------------------- AY, Claimant-Appellant, v. ERIC K. SHINSEKI, Secretary of Veterans Affairs, Respondent-Appellee. ______________________ 2012-7048 ______________________ 2 AZ v. SHINSEKI Appeal from the United States Court of Appeals for Veterans Claims in No. 10-2390, Chief Judge Bruce E. Kasold. ______________________ Decided: September 30, 2013 ______________________ ZACHARY M. STOLZ, Chisholm, Chisholm & Kilpatrick, of Providence, Rhode Island, argued for claimant- appellant in appeal no. 2012-7046. On the brief was SEAN A. RAVIN, of Washington, DC. Of counsel was KENNETH M. CARPENTER, Carpenter, Chartered, of Topeka, Kansas. STEVEN M. MAGER, Trial Attorney, Commercial Liti- gation Branch, Civil Division, United States Department of Justice, of Washington, DC, argued for respondent- appellee. With him on the brief were STUART F. DELERY, Acting Assistant Attorney General, JEANNE E. DAVIDSON, Director, and KIRK T. MANHARDT, Assistant Director. Of counsel was ELIZABETH M. HOSFORD, Senior Trial Coun- sel. Of counsel on the brief were DAVID J. BARRANS, Depu- ty Assistant General Counsel, and TRACEY P. WARREN, Attorney, United States Department of Veterans Affairs, of Washington, DC. ______________________ KENNETH M. CARPENTER, Carpenter, Chartered, of Topeka, Kansas, argued for claimant-appellant in appeal no. 2012-7048. On the brief was SEAN A. RAVIN, of Wash- ington, DC. STEVEN M. MAGER, Trial Attorney, Commercial Liti- gation Branch, Civil Division, United States Department of Justice, of Washington, DC, argued for respondent- appellee. With him on the brief were STUART F. DELERY, Acting Assistant Attorney General, JEANNE E. DAVIDSON, AZ v. SHINSEKI 3 Director, and MARTIN F. HOCKEY, JR., Assistant Director. Of counsel was ELIZABETH M. HOSFORD, Senior Trial Counsel. Of counsel on the brief were MICHAEL J. TIMINSKI, Deputy Assistant General Counsel, and TRACEY P. WARREN, Attorney, United States Department of Vet- erans Affairs, of Washington, DC. ______________________ Before DYK, CLEVENGER, and MOORE, Circuit Judges. Opinion for the Court filed by Circuit Judge DYK. Dissenting opinion filed by Circuit Judge MOORE. DYK, Circuit Judge. Veterans AZ and AY filed claims with the Department of Veterans Affairs (“VA”) seeking disability compensation for post-traumatic stress disorder (“PTSD”) alleged to have resulted from sexual assaults that occurred during service. The veterans’ service records do not reflect any reports of the alleged sexual assaults. The VA Regional Office (“RO”), Board of Veterans’ Claims (“Board”), and the Court of Appeals for Veterans Claims (“Veterans Court”) rejected the claims in part on the ground that the veterans’ service records did not include reports of the alleged assaults, and because the veterans stated that the assaults were never reported to military authorities. The veterans argue that the Board and Veterans Court erred by treating the absence of reports of the alleged sexual assaults as pertinent evidence that the assaults did not occur. We agree with the veterans that the absence of a service record documenting an unreport- ed sexual assault is not pertinent evidence that the sexual assault did not occur. We further hold that the Board and Veterans Court may not rely on a veteran’s failure to report an in-service sexual assault to military authorities as pertinent evidence that the sexual assault did not occur. We vacate and remand. 4 AZ v. SHINSEKI BACKGROUND I. AZ Appellant AZ served honorably on active duty from March 1973 to July 1974. She was pregnant when she left service, and gave birth to a daughter in October of 1974. In 2004, she was diagnosed with psychiatric prob- lems including PTSD. She attributes her PTSD to sexual and physical abuse by the father of her child, a non- commissioned officer of superior rank: While in the service, I was sexually assaulted and beaten by Sgt. [J.H., a superior non-commissioned officer]. I became pregnant after one of the sexual assaults on about January or February 1974. On October 21, 1974, I had a daughter as a result of one of the sexual assaults. AZ J.A. 242 (Feb. 2004 Statement in Support of Claim). She allegedly began having nightmares while still in service. AZ did not report the alleged sexual assaults to military authorities, and her service records do not reflect any report of an assault. However, as discussed below, there is evidence that the alleged assaults were disclosed contemporaneously to AZ’s family members. In February 2004, AZ filed a claim for service connec- tion for PTSD. The VA RO denied her claim, noting that her service records did not document that a sexual assault had occurred. J.A. 239–40 (June 2004 Rating Decision). AZ requested reconsideration, and submitted lay statements from three siblings, who reported that she was outgoing prior to service but became less communicative after meeting her alleged abuser. The siblings stated that in “about her fourth or fifth month of pregnancy [AZ] told us she had been sexually assaulted, verbally abused and beaten by Sgt. [J.H.],” and that AZ did not report the assaults to military authorities because she was afraid and did not think she would be believed. AZ J.A. 233; see AZ v. SHINSEKI 5 also AZ J.A. 234, 235. AZ’s request for reconsideration explained that “she did not report these incidents to the military legal authorities” because she “was a young girl, sexually assaulted, verbally abused and beaten by a superior [officer] and she was in fear of her life.” AZ J.A. 232 (July 19, 2004 Request for Reconsideration). The RO again denied service connection. The RO found that although AZ’s siblings had reported that she told them about the assault during her pregnancy, [y]our service medical records and military per- sonnel file were review [sic] for verification of the assaults or at least some indication you were as- saulted. The service medical records are negative for any comments made by you or the physicians regarding episodes of beatings or sexual trauma. AZ J.A. 228 (Feb. 2006 Statement of the Case). AZ appealed to the Board, which remanded for further evidentiary development. In re AZ, No. 06-08 672 (Bd. Vet. App. Mar. 10, 2008). Service connection was again denied, and the Board affirmed that denial in January, 2010. In re AZ, No. 06-08 672 (Bd. Vet. App. Jan. 22, 2010). The Board determined that AZ’s service records “do not show any complaints, treatment or diagnosis for any psychiatric disorder or any reports of injuries from a personal assault during service,” id., slip op. at 6–7, and that there is no documentation in the service records to indicate that the Veteran reported having been personally assaulted, or that she instigated pro- ceedings against her alleged attacker. Moreover, the Veteran’s service treatment records contain no evidence that the Veteran sought treatment for the alleged sexual or physical assault itself. Id. at 15. The Board stated that under the applicable regulations, “[s]ervice department records must support, 6 AZ v. SHINSEKI and not contradict, the veteran’s testimony regarding non-combat stressors.” Id. at 10 (emphasis added). According to the Board, [t]he crux of the issue . . . is whether there is com- petent evidence of record corroborating the Veter- an’s allegation that she was sexually assaulted in service. . . . Here, the evidence of record does not corrobo- rate the Veteran’s account . . . . [T]here is no doc- umentation in the service records to indicate that the Veteran reported having been personally as- saulted, or that she instigated proceedings against her alleged attacker. Moreover, the Veteran’s service treatment records contain no evidence that the Veteran sought treatment for the alleged sex- ual or physical assault itself. . . . . . . The Veteran has stated that she did not report sexual assault to military or civilian au- thorities. It is noted that a positive pregnancy test was reflected in the service treatment rec- ords; however, there is no notation that the preg- nancy was a result of sexual abuse. Id. at 15 (emphasis added). The Board acknowledged that the three lay state- ments submitted in support of AZ’s claim “reflect[ed] the Veteran’s reports of sexual and physical abuse.” Id. However, the Board found that “service treatment records and the report of examination prior to separation show no complaints or findings indicative of a psychiatric prob- lem,” and that “[r]ecords from service do not document any in service assault.” Id. at 13. The lay statements were insufficient to overcome this deficit, because “none of these individuals . . . claimed to witness any personal assault take place.” Id. at 15. Therefore, the statements were “not as probative as the contemporaneous service . . . AZ v. SHINSEKI 7 records that do not reflect that the Veteran was assaulted while on active duty.” Id. at 16. Relying on the absence of service records of the assault, records of a disciplinary problem predating the alleged assault, a service medical record indicating AZ “plann[ed] on getting married,” documentation of possible post-service stressors such as unemployment, and other evidence, the Board concluded that “the evidence of record is insufficient to confirm that the [alleged assault] occurred.” Id. at 15–18. AZ appealed to the Veterans Court. She argued, inter alia, that the Board improperly rejected her siblings’ statements solely because there was no documentation of the sexual assault in her service records: Given that the [VA] Secretary’s own procedures manual acknowledges that very few in-service as- saults are documented, it is perplexing that the Board would use the lack of documentation of an assault as a basis for diminishing the probative weight of the statements. AZ J.A. 294 (Appellant’s Brief to the Veterans Court). The Veterans Court affirmed. AZ v. Shinseki, No. 10- 2393 (Vet. App. Nov. 28, 2011). It found that the Board did not err by weighing “these lay statements against the other evidence of record and f[i]nd[ing] them less proba- tive,” because “the Board is permitted to weigh the ab- sence of corroborating records and documents against the lay evidence of record.” Id., slip op. at 5. Likewise, the Veterans Court found that the Board gave due considera- tion to the evidence pertaining to AZ’s pregnancy, includ- ing the lack of medical records indicating that the pregnancy was a result of sexual assault, again stating that “[t]he Board is permitted to weigh the absence of corroborating records and documents against the lay evidence of record.” Id. at 7. AZ timely appealed the Veterans Court’s decision to this court. We have jurisdic- tion pursuant to 38 U.S.C. § 7292. 8 AZ v. SHINSEKI II. AY Appellant AY served honorably on active duty from July 1980 to July 1983. She was diagnosed with PTSD in 2002. She attributes her PTSD to a sexual assault com- mitted by another soldier during her military training. Her service records contain no record of a report of sexual assault, treatment for sexual assault, or psychiatric problems. She stated that she did not report the alleged assault to military authorities. As discussed below, there is evidence the assault was contemporaneously reported to other individuals. In 2004, AY filed a claim for service connection for a psychiatric disorder, including PTSD. AY’s ex-husband submitted a statement in support of her claim, stating that she told him about the alleged assault while they were in service together. The RO denied entitlement to service connection. In August 2005, AY requested that her claim be reo- pened and submitted three more lay statements from additional individuals who knew her during service. ES, a fellow soldier stationed with AY during training, report- ed that AY told her about the sexual assault the day after it occurred, and that AY subsequently became despondent and discussed suicide. AH, AY’s roommate at her subse- quent duty assignment, reported that AY attempted suicide and received treatment for the incident at a base hospital. AY’s sister stated that prior to entering the military, AY was outgoing, but that afterward, she was “crazy.” AY J.A. 121. The RO again denied service connection. The RO acknowledged that AY had “provided statements from [four individuals] who support that they knew you while in service and that you told them about the rape.” AY J.A. 105 (Mar. 2006 Rating Decision). But the RO found there was “no evidence” to corroborate the alleged assault: AZ v. SHINSEKI 9 Service Medical Records . . . failed to provide any evidence to support that you were raped or at- tempted suicide while in service. The records pro- vided no information which would be indicative of a personal assault. . . . Military Personnel Records provided no evi- dence of any personnel problems or reported rape while in service. . . . Service connection has been denied since there is insufficient evidence to support that you were raped while in service. The statements you provided from your friends and family were insuf- ficient to substantiate your claimed stressor, since none of them witnessed the incident, and only knew of the incident due to your statements alone. The service records failed to show that you had any difficulties while in service. Your records show that you had commendable military service. There is no evidence that can substantiate that you were raped while in service. Therefore service connection for PTSD is denied . . . . Id. (emphases added). AY appealed to the Board. The Board also acknowl- edged the lay statements, but found that they were “di- rectly contradicted by other evidence,” and that “in this case, other evidence strongly diminishes the[ir] probative value.” In re AY, No. 07-16 960, slip op. at 14 (Bd. Vet. App. Oct. 26, 2009). The Board observed that AY “did not report the [alleged assault] to police at the time that it occurred,” or “to anyone immediately”; that her service records contained no records of psychiatric treatment; and that there was “no record that the Veteran reported that she was sexually assaulted to medical personnel.” Id. at 11–12. The Board found that these omissions contradict- ed the lay statements, and moreover, that AY’s “service records appear[ed] to be inconsistent with her reports and 10 AZ v. SHINSEKI the lay statements indicating that the Veteran experi- enced [severe] depression during service.” Id. at 14. Thus, “the lay statements submitted by the Veteran [we]re insufficient to corroborate her claimed stressor.” Id. at 15. Citing inconsistencies in AY’s accounts of the alleged assault, service records commending AY’s duty performance and “cheerful demeanor,” and other evi- dence, the Board determined that “the evidence is insuffi- cient to confirm the occurrence of [the alleged sexual] assault.” Id. at 15–16. AY appealed, contending “that the Board improperly found the lay evidence not credible merely because it was unaccompanied by contemporaneous medical evidence.” See AY v. Shinseki, No. 10-2390, slip op. at 2 (Vet. App. Aug. 17, 2011). The Veterans Court disagreed, pointing out that the lay statements were not merely uncorrobo- rated, but contradicted by the absence of records docu- menting treatment for the suicide attempt as described by AH. See id. The Court also observed that the records commending AY’s duty performance and demeanor “con- tradicted” the lay statements reporting that AY suffered from depression and other psychiatric problems after the sexual assault. Id. Thus, [a]lthough AY submitted statements regarding the [alleged assault] and lay witnesses submitted statements regarding her demeanor after the [al- leged assault], the Board found them inconsistent with other evidence in the record and therefore not credible. Id. Concluding that “the Board rejected the lay evidence on permissible grounds: internal inconsistencies and contradictions with other evidence in the record,” the Veterans Court affirmed the Board’s finding that there was “no credible supporting evidence corroborating the in- service assault.” Id. AY timely appealed to this court. We have jurisdiction pursuant to 38 U.S.C. § 7292. AZ v. SHINSEKI 11 * * * Following oral argument in both cases, we requested and received supplemental briefing from the parties on the question of “whether, in considering a disability claim based on in-service sexual assault, the [VA] may properly rely on the absence of contemporaneous service records reporting a sexual assault.” See AZ v. Shinseki, 2012- 7046 (Fed. Cir. Mar. 12, 2013); AY v. Shinseki, 2012-7048 (Fed. Cir. Mar. 12, 2013). DISCUSSION These cases involve two simple but important ques- tions. The first is whether, when adjudicating a PTSD claim based on an alleged in-service sexual assault, the Board may treat the absence of contemporaneous service records reporting the sexual assault as pertinent evidence that the sexual assault did not occur. The second ques- tion, related to the first, is whether the fact that no report of the alleged sexual assault was made to military author- ities should be considered evidence that the alleged sexual assault did not occur. This court’s review (apart from constitutional issues) is limited to questions of law, 38 U.S.C. § 7292(d), and our review of legal issues is without deference. Cook v. Prin- cipi, 353 F.3d 937, 938 (Fed. Cir. 2003). We have jurisdic- tion to review the Veterans Court’s legal determinations with respect to “the types of evidence which may support a claim for benefits.” Buchanan v. Nicholson, 451 F.3d 1331, 1335 (Fed. Cir. 2006); see id. (holding that “lay evidence is one type of evidence that must be considered, if submitted, when a veteran’s claim seeks disability benefits”); see also Fagan v. Shinseki, 573 F.3d 1282, 1289 (Fed. Cir. 2009) (holding that the Veterans Court did not err by disregarding an inconclusive, nonprobative medical opinion because it was “not pertinent evidence, one way or the other” to service connection). Accordingly, we have 12 AZ v. SHINSEKI jurisdiction to hear this appeal. See Fagan, 573 F.3d at 1286; Buchanan, 451 F.3d at 1335. I A By statute, the VA is required to “consider all infor- mation and lay and medical evidence of record” in deter- mining service connection, 38 U.S.C. § 5107(b), but the statute also requires that medical and lay evidence be “pertinent,” see 38 U.S.C. § 1154(a)(1). In particular, the statute requires the VA Secretary to adopt regulations requiring that in each case where a veteran is seeking service-connection for any disability due consideration shall be given to the places, types, and circumstances of such veteran’s service as shown by such veteran’s service record, the official history of each organization in which such veteran served, such veteran’s medical records, and all pertinent medical and lay evidence. Id. (emphasis added); see also 38 C.F.R. § 3.303(a) (simi- lar). 1 Establishing service connection for a PTSD claim re- quires (1) a medical diagnosis of PTSD; (2) “a link, estab- lished by medical evidence, between [the] current symptoms and an in-service stressor”; and (3) “credible 1 The dissent cites provisions of the statute and VA regulations that are supposedly contrary to the majority’s holding. Dissent at 7. But as discussed below, these provisions are not addressed to the absence of service records. Rather, they are addressed to existing service records of “the places, types, and circumstances of such veteran’s service.” See 38 U.S.C. § 1154(a)(1). We agree that such service records must be given “due considera- tion,” where they exist. AZ v. SHINSEKI 13 supporting evidence that the claimed in-service stressor occurred.” 38 C.F.R. § 3.304(f). “[C]orroborating evidence of an in-service stressor may in some situations be provid- ed by lay evidence.” See Nat’l Org. of Veterans’ Advocates, Inc. v. Sec’y of Veterans Affairs, 330 F.3d 1345, 1352 (Fed. Cir. 2003) (upholding the validity of section 3.304(f)). When a PTSD claim is based on “in-service personal assault,” which includes sexual assault, the regulation provides that evidence from sources other than the veteran’s service records may corroborate the veteran’s ac- count of the stressor incident. Examples of such evidence include, but are not limited to: records from law enforcement authorities, rape crisis cen- ters, mental health counseling centers, hospitals, or physicians; pregnancy tests or tests for sexually transmitted diseases; and statements from family members, roommates, fellow service members, or clergy. § 3.304(f)(5). Supporting evidence found in such sources, if credible and pertinent, is positive evidence of the in- service stressor that the VA must consider. There is no dispute that a service record documenting an alleged sexual assault, if it existed, would be evidence supporting the conclusion that the alleged assault oc- curred. However, neither the statute, section 3.304(f)(5), nor any other VA regulation directly addresses the role that the absence of service records reporting the alleged assault should play in a disability determination. As we explained in Fagan, the VA must consider all evidence “pertinent” to service connection. See 573 F.3d at 1287–88. Pertinent evidence is evidence that is rele- vant, that is, it must tend to prove or disprove a material fact. See generally Fed. R. Evid. 401; 1 McCormick on Evidence § 185, at 994–1004 (7th ed. 2013). Evidence that is insufficiently probative, such as an inconclusive medi- 14 AZ v. SHINSEKI cal report, “provides neither positive nor negative support for service connection,” and “is not pertinent evidence, one way or the other, regarding service connection.” Fagan, 573 F.3d at 1289 (emphasis added); see also 2 Jack B. Weinstein & Margaret A. Berger, Weinstein’s Federal Evidence § 401.07 (2d ed. 2012) (“Weinstein’s Federal Evidence”) (“[Evidence that] simply does not tend to prove a fact that is of consequence to the action[] . . . is not relevant.”) (emphases omitted). The absence of certain evidence may be pertinent if it tends to disprove (or prove) a material fact. See Forshey v. Principi, 284 F.3d 1335, 1358 (Fed. Cir. 2002) (en banc) (“[T]he definition of evidence encompasses ‘negative evidence,’ which tends to disprove the existence of an alleged fact . . . .”). In Buchanan, we addressed a situa- tion in which confirmatory service records were absent. There, a veteran seeking service connection for schizo- phrenia challenged the Board’s refusal to credit several lay statements reporting that he had suffered from symp- toms of schizophrenia “while in service or soon thereaf- ter.” 451 F.3d at 1333. The Board found that these lay statements “lack[ed] credibility absent confirmatory clinical records to substantiate such recollections” of the veteran’s in-service symptoms. Id. (quotation marks omitted). We vacated and remanded. We held that it would be “legally untenable” for the Board to conclude “that absent confirmatory [record] evidence, lay evidence lacks credi- bility,” and explained that “the lack of such records does not, in and of itself, render lay evidence not credible.” Id. at 1336. We recognized, however, that as a general matter, “the lack of contemporaneous medical records may be a fact that the Board can consider and weigh against a veteran’s lay evidence.” Id. The appellants do not dispute that the absence of con- temporaneous service records of an event or condition will AZ v. SHINSEKI 15 often be pertinent. However, they argue that the absence of contemporaneous service records of unreported in- service sexual assaults is not pertinent, because it is not reasonable to expect that such assaults would have been reported to superior officers, or that records of unreported assaults would exist. 2 B The VA does not dispute that, in the great majority of cases, such incidents are not reported to military authori- ties, and therefore such records do not exist. The regula- tory history of section 3.304(f)(5) reveals that it was enacted in part to address the fact that “[m]any incidents of in-service personal assault are not officially reported.” See Post-Traumatic Stress Disorder Claims Based on Personal Assault, 65 Fed. Reg. 61,132, 61,132 (Oct. 16, 2000) (proposed rule). This is especially true of in-service sexual assaults. The Department of Defense (“DoD”), which is required by statute to report to Congress annual- ly on “[t]he number of sexual assaults [involving] mem- bers of the Armed Force[s] that were reported to military officials,” see Ronald W. Reagan National Defense Author- ization Act for Fiscal Year 2005, Pub. L. No. 108-375, § 577(f), 118 Stat. 1811, 1927 (2004), has estimated that a significant majority of in-service sexual assaults are unreported. In 2012, the “[DoD] estimate[d] that about 11 percent of the sexual assaults [involving Service mem- bers] that occur each year are reported to a DoD authori- ty.” 1 DoD Sexual Assault Prevention & Response Office (“SAPRO”), Department of Defense Annual Report on 2 To the extent the appellants argue here that the Board or Veterans Court, contrary to Buchanan, erred by requiring a record of the alleged assault to corroborate the lay statements, we do not think they imposed any such requirement. 16 AZ v. SHINSEKI Sexual Assault in the Military: Fiscal Year 2012, at 18 (2013) (“2012 SAPRO Report”). DoD estimated that [i]n 2010, reports by victims accounted for about 14 percent of the sexual assaults estimated to have occurred. . . . The majority of sexual assaults against Service members each year remain unre- ported. DoD SAPRO, Department of Defense Annual Report on Sexual Assault in the Military: Fiscal Year 2010, at 22 (2011) (“2010 SAPRO Report”). 3 And “[i]n 2006, reports to DoD authorities accounted for about 7 percent of the [in- service] sexual assaults estimated to have occurred that year.” 2010 SAPRO Report, supra, at 22; see also id. at 98 (stating that the “[e]stimated [n]umber of [s]exual [a]ssaults [g]oing [u]nreported to DoD” in 2006 was 93%). The DoD estimates that between 2006 and 2012, fewer 3 For incidents of sexual assault occurring after the 2007 revisions to the Uniform Code of Military Justice (“UCMJ”) and prior to the 2012 UCMJ revisions, “sexual assault” include[s] rape, aggravated sexual assault, non- consensual sodomy, aggravated sexual contact, abusive sexual contact, wrongful sexual contact, and attempts to commit these offenses. 2010 SAPRO Report, supra, at 64; see also UCMJ art. 120(m) (2007) (defining “wrongful sexual contact” as “without legal justification or lawful authorization, en- gag[ing] in sexual contact with another person without that other person’s permission”), codified at 10 U.S.C. § 920(m) (2007); UCMJ art. 120(t)(2) (2007) (defining “sexual contact” as “intentional touching” of certain body parts “with an intent to abuse, humiliate, or degrade any person or to arouse or gratify the sexual desire of any person”), codified at 10 U.S.C. § 920(t)(2) (2007). AZ v. SHINSEKI 17 than 15 percent of military sexual assault victims report- ed the assault to a military authority. 2012 SAPRO Report, supra, at 53. 4 VA has long recognized that the underreporting of in- service sexual assaults is a problem for claimants. As early as 1996, the VA claims manual state[d], with respect to claims based upon a per- sonal assault: “The service record may be devoid of evidence because many victims of personal as- sault, especially sexual assault and domestic vio- lence, do not file official reports either with military or civilian authorities.” YR v. West, 11 Vet. App. 393, 398 (1998) (quoting VA Adjudication Procedure Manual M21-1 (Feb. 20, 1996)). Servicemen and servicewomen who experience in- service sexual assaults face “unique” disincentives to report. See DoD Care for Victims of Sexual Assault Task Force, Report on Care for Victims of Sexual Assault 28 (2004) (“2004 Task Force Report”) (“Finding 12: There are barriers to reporting incidents of sexual assault. Some 4 Recognizing the disincentives and reluctance to report, in 2005 the DoD adopted a policy allowing ser- vicemembers to make a “restricted” report of a sexual assault, which “allows victims to confidentially access medical care and advocacy services without initiating an official investigation or command notification.” 2012 SAPRO Report, supra, at 17; see also id. app. B at 99–101 (describing the restricted reporting option). These confi- dential reports apparently are not included in the service member’s records. The DoD figures cited in the text for reports of sexual assault in the years 2006–2012 include both restricted and unrestricted reports. Thus, it seems probable that the percentage of reports that would appear in a service member’s records is even lower. 18 AZ v. SHINSEKI are consistent with those in the civilian community while others are unique in a military setting.”). DoD findings from 2010 indicate that more than half of the female service members who experienced, but did not report, an incident of unwanted sexual contact 5 during the previous year “cite[d] fear of retaliation or reprisals . . . as a reason for not reporting.” See 2010 SAPRO Report, supra, at 95. 6 5 DoD bases its estimates in part on the Workplace and Gender Relations Survey of Active Duty Members (“WGRA”), conducted every two to four years. See gener- ally Defense Manpower Data Center, DMDC Report No. 2010-025, 2010 WGRA: Overview Report on Sexual As- sault (2011). WGRA uses the term “unwanted sexual contact” for acts corresponding to sexual assault, see id. at iv; see also 2010 SAPRO Report, supra, at 92 n.132, and distinguishes these acts from “sexual harassment,” which comprises “crude/offensive behavior, unwanted sexual attention, and sexual coercion,” see Defense Manpower Data Center, DMDC Report No. 2010-023, 2010 Service Academy Gender Relations Survey 4 (2010). 6 According to VA clinical materials, sexual victimization that occurs in [the military] setting often means that victims are relying on their perpetrators (or associates of the perpetra- tor) to provide for basic needs including medical and psychological care. . . . Because organizational cohesion is so highly val- ued within the military environment, divulging any negative information about a fellow soldier is considered taboo. Accordingly, many victims are reluctant to report sexual trauma . . . . Amy Street & Jane Stafford, Military Sexual Trauma: Issues in Caring for Veterans in Eve B. Carlson et al., Iraq War Clinician Guide, 66, 66–67 (2d ed. 2004), available at AZ v. SHINSEKI 19 Of those who did make a report, “most female victims surveyed indicate[d] experiencing some kind of retaliation (either professional or social) or administrative action against them associated with their reporting the sexual assault.” Id. at 41; see also DoD, Calendar Year 2004 Report: Sexual Offenses Involving Members of the Armed Forces 5 (2005) (“2004 DoD Report”) (“A victim’s fear of punishment [for misconduct occurring at the same time as the assault] is a significant barrier to reporting sexual assault.”). Many victims also fear “the stigma associated with sexual assault reporting,” and may “mistakenly assume that being the victim of a sexual assault will make them appear weak or incapable of performing their mission.” 2010 SAPRO Report, supra, at 20; see also Gov’t Accountability Office, GAO-08-1013T, Military Personnel: Preliminary Observations on DoD’s and the Coast Guard’s Sexual Assault Prevention and Response Programs 14 (2008) (finding that “[c]ommonly cited reasons [for not reporting] at the installations we visited included: (1) the belief that nothing would be done; (2) fear of ostracism, harassment, or ridicule by peers; and (3) the belief that their peers would gossip about the incident”). In sum, due to numerous disincentives to reporting, “[o]ver the past 6 years, the [DoD] estimates that fewer than 15 percent of military sexual assault victims re- port[ed] the matter to a military authority.” 2012 SAPRO Report, supra, at 53. There is also no reason to believe that the most severe assaults are reported with greater frequency. First, the government has not suggested that this is the case. Second, there is no reason to believe that the factors which usually deter reporting, such as fear of stigma or fear of reprisal by the perpetrator, are lessened in those http://www.ptsd.va.gov/professional/manuals/manual- pdf/iwcg/iraq_clinician_guide_v2.pdf. 20 AZ v. SHINSEKI cases where a rape is completed. Third, the DoD has consistently stated that the underreporting of rape is a significant problem. See generally Office of the Inspector General of the DoD, Interim Report on the USAFA Sexual Assault Survey (2003) (discussing incidence and reporting of in-service rape); see also, e.g., DoD, 2006 DoD Annual Report on Military Services Sexual Assault for Calendar Year 2006, at 2 (2007). Fourth, the DoD figures from the pre-2007 period, when “sexual assault” was defined more narrowly (as rape, nonconsensual sodomy, indecent assault, and attempts to commit those offenses), indicate that the vast majority of offenses were not reported. See 2010 SAPRO Report, supra, at 22, 98 (comparing esti- mates for 2006 (93% unreported to the DoD) and 2010 (86% unreported to the DoD)). 7 Military academy surveys during that period also indicate that significantly more rapes occurred than were reported to authorities. For example, a 2006 survey of U.S. Air Force Academy (“USAFA”) cadets found that [o]f the 60 Women [reporting at least one event of unwanted sexual contact], 26 reported completed sex or other sex acts, 29 reported attempted sex or other sex acts, and 49 reported sexual touching. 7 “For incidents that occurred prior to the changes made to the UCMJ on October 1, 2007, sexual assault included rape, nonconsensual sodomy, indecent assault, and attempts to commit these acts.” Id. at 64; see also 2004 DoD Report, supra, at 3 (reflecting the same defini- tion). “Indecent assault” is an “attempt[] or offer[] with unlawful force or violence to do bodily harm to another person,” committed upon a person other than the defend- ant’s spouse, “with the intent to gratify the lust or sexual desires of the accused.” See UCMJ art. 128 (2005) (defin- ing assault), codified at 10 U.S.C. § 928 (2005); Manual for Courts-Martial of the United States, ¶ 63, at IV-98 to -99 (2005 ed.) AZ v. SHINSEKI 21 . . . Only 3 (5%) Women reported [the incident af- fecting them most] to an authority or organiza- tion. USAFA, Report on Sexual Harassment and Violence: USAFA Academy Program Year 2006 16 (2006). A 2003 survey similarly found that of 26 female cadets who reported experiencing a single incident of sexual assault which involved rape or attempted rape, six had reported the assault to military authorities, Office of the Inspector General of the DoD, supra, at 17–18, and further noted that three of these six cadets reported suffering reprisal by their peers or by military authorities. Id. Appellants argue that given the pervasive nonreport- ing of in-service sexual assaults, the Board and Veterans Court acted contrary to the statute and section 3.304(f)(5) by “accord[ing] evidentiary status to the absence of [ser- vice record] evidence” reporting the alleged assaults. See AZ Br. 20; AY Br. 21. 8 C At common law, the majority of courts held that where circumstances supported the conclusion that “an entry would naturally have been made if a transaction had occurred,” then evidence showing the absence of an entry “should ordinarily be equivalent to an assertion that no such transaction occurred, and therefore should be admissible in evidence for that purpose.” 5 John Henry Wigmore, Evidence in Trials at Common Law § 1531, at 463 (James H. Chadbourn rev., 1974) (“5 Wigmore on 8Appellants do not dispute that where a veteran claims to have reported a sexual assault to military authorities, but no report is found, the absence of the report may be pertinent to the evaluation of his or her claim. 22 AZ v. SHINSEKI Evidence”) (emphasis added); see also id. n.2 (collecting cases). 9 This rule has long been followed by the Supreme Court. For example, in Chesapeake & Delaware Canal Co. v. United States, the Supreme Court explained that evidence of the absence of a payment entry from Treasury Department records was admissible as evidence of non- payment, because [s]uch books so kept presumptively contained a record of all payments made and the absence of any entry of payment, where it naturally would have been found if it had been made, was evidence of nonpayment proper for the consideration of the jury. 250 U.S. 123, 129 (1919) (emphases added). Similarly, evidence of “[t]he absence of all traces of [land] grants, where evidence would usually be found, if it had existed,” is admissible to show a purported grant was never made. Hornsby v. United States, 77 U.S. 224, 241 (1870) (em- phasis added). The rule is logical because where a com- prehensive record of events is regularly kept, as in an archive of land grants, “[t]he absence of any record evi- dence is remarkable, if the title is genuine.” United States v. Teschmaker, 63 U.S. 392, 405 (1860). Following this general approach, lower federal courts applying common law evidentiary principles have gener- 9 A minority of courts appear to have held that evi- dence of the absence of an entry from regularly kept records was inadmissible. See Shreve v. United States, 77 F.2d 2, 7 (9th Cir. 1935) (collecting “cases which hold that the nonexistence of a debt or obligation cannot be estab- lished by proof that the books contain no such entry”); see also 5 Wigmore on Evidence, supra, § 1531, at 463 (criti- cizing the minority rule). AZ v. SHINSEKI 23 ally held that “[t]he absence of a record of an event which would ordinarily be recorded gives rise to a legitimate negative inference that the event did not occur.” See United States v. Robinson, 544 F.2d 110, 114 (2d Cir. 1976) (emphasis added); see also United States v. De Georgia, 420 F.2d 889, 893 (9th Cir. 1969) (“[I]f a business record designed to note every transaction of a particular kind contains no notation of such a transaction between specified dates, no such transaction occurred between those dates.”) (emphasis added); Keith v. United States, 250 F.2d 355, 356 (5th Cir. 1957) (holding that an agent’s testimony that “he checked with the police department, the sheriff’s office, public utilities, chamber of commerce, telephone and business directories and talked to local residents” but found no record of a person was admissible as proof that the person was fictitious); Nichols v. United States, 48 F.2d 46, 49 (5th Cir. 1931) (similar). 10 10See also, e.g., Commonwealth ex rel. Funk v. Clark, 225 S.W.2d 118, 119 (Ky. 1949) (“The absence of an entry in a public [school] record that would appear in it in the usual course may be generally accepted as evidence that an event did not take place or that something was not done.”) (emphasis added); Duren v. Ark. State Bd. of Optometry, 201 S.W.2d 578, 579 (Ark. 1947) (admitting testimony that the appellant’s name did not appear on “a roster of all men licensed to practice optometry in Arkan- sas” as proof that he was unlicensed); Eisminger v. Mitch- ell, 73 P.2d 862, 865 (Okla. 1937) (affirming the trial court’s admission of court records as evidence that no previous judgment was entered where, “because of the statute and general rule, [the records] should [have] contain[ed] such an entry if judgment had been entered”); Sharp v. Pawhuska Ice Co., 217 P. 214, 217–18 (Okla. 1923) (collecting cases refusing to admit the absence of records, and holding them inapplicable where business records were regularly kept in triplicate); Griffin v. Wise, 24 AZ v. SHINSEKI Correspondingly, courts have refused to admit evi- dence of the absence of a record to show that an event did not occur, where it was not reasonable to expect the event to have been recorded. For example, in Shreve v. United States, the Ninth Circuit held that, even assuming evi- dence of the absence of an entry from a company’s account books would generally be admissible to show a transac- tion had not occurred, the account books in question “did not show that the [missing] transaction was fictitious,” as it had not been established that they “contained an accu- rate record of all the business of the corporation, and particularly that under the method of bookkeeping adopt- ed by the corporation the books would disclose the exist- ence of all outstanding indebtedness.” 77 F.2d 2, 6–7 (9th Cir. 1935) (citation omitted) (emphases added). Similarly, in Bowman v. Kaufman, the Second Circuit held that police records were inadmissible to prove that no witness- es had reported a brake failure, because the records did not indicate that the police officer would have asked about or recorded any statements regarding the cause of the accident; thus, although the records “g[ave] no suggestion that [the police officer] was ever told of a brake failure, neither [wa]s there anything in [the records] inconsistent with this having happened.” 387 F.2d 582, 587–88 (2d Cir. 1967); see also id. at 587 n.5 (citing 5 Wigmore on Evidence, supra, § 1531). The common-law rule has been codified in Rule 803 of the Federal Rules of Evidence, which represents a “syn- 41 S.E. 1003, 1004 (Ga. 1902) (“A [tax] book which would contain an entry if such an entry existed is admissible . . . to show that such entry is not in existence.”) (emphasis added); Knapp v. Day, 34 P. 1008, 1009 (Colo. App. 1893) (affirming the trial court’s admission of testimony of “the postmaster, having charge of the records of the [post] office, . . . that the records did not show [registered] letters to have been received”). AZ v. SHINSEKI 25 thesis” of common law principles. See Fed. R. Evid. 803 advisory committee’s note (1972). While these rules are not mandatory in VA disability hearings, they are further evidence of the common law approach, and we agree with the Veterans Court that they offer useful guidance. See Gambill v. Shinseki, 576 F.3d 1307, 1330 (Fed. Cir. 2009) (Moore, J., concurring) (recognizing that the Veterans Court has looked to the Federal Rules as “‘guiding factors to be used by the Board in evaluating the probative value of medical opinion evidence’” (quoting Nieves-Rodriguez v. Peake, 22 Vet. App. 295, 302 (2008))); see also Stevenson v. Linens of the Week, 688 F.2d 93, 98–99 (D.C. Cir. 1982) (approving the Benefits Review Board’s reliance by analo- gy to Rule 803 “on the principle that the absence of an entry in a business record is probative of the non- occurrence of the event in question”). Rule 803(7) allows admission of “[e]vidence that a matter is not included in a record” of a regularly conduct- ed activity in order “to prove that the matter did not occur or exist,” so long as “a record was regularly kept for a matter of that kind” and there are no “circumstances indicat[ing] a lack of trustworthiness.” Fed. R. Evid. 803(7). 11 Rule 803(10) similarly authorizes admission of Some courts characterized testimony about the 11 absence of a record or statement as “hearsay” or “negative hearsay.” See, e.g., Menard v. Cashman, 55 A.2d 156, 160–61 (N.H. 1947) (holding that testimony that no patrons ever complained about allegedly defective stairs was correctly excluded because “[i]f the excluded testimo- ny was not hearsay, and therefore inadmissible, it was at most only evidence of inconclusive silence” (citations omitted)); see also Edmund M. Morgan, Hearsay Dangers and the Application of the Hearsay Concept, 62 Harv. L. Rev. 177, 213 (1948) (“The decisions are in conflict as to whether the silence is to be treated as hearsay.”). 26 AZ v. SHINSEKI evidence of the absence of a public record to prove that “a matter did not occur or exist, if a public office regularly kept a record or statement for a matter of that kind.” Fed. R. Evid. 803(10)(B). Notably for our purposes, both rules require for ad- missibility that “a record was regularly kept” for the type of event in question. See Fed. R. Evid. 803(7); see also Fed. R. Evid. 803(10)(B). To establish that the record tends to prove (or disprove) a matter, it is “crucial” to “[d]emonstrat[e] that the records were kept in such a way that the matter would have been recorded had it oc- curred.” 5 Weinstein’s Federal Evidence, supra, at § 803.09 (emphasis added) (discussing Rule 803(7)); see also United States v. Rich, 580 F.2d 929, 938 (9th Cir. 1979) (noting that Rules 803(7) and 803(10) “are grounded on the high probability of [the records’] accuracy”). 12 While the absence of a record is “probably not hearsay as defined in” the general hearsay rule, its inclusion under the rubric of hearsay exceptions establishes that regard- less of its hearsay status, the “[f]ailure of a record to mention a matter which would ordinarily be mentioned is satisfactory evidence of its nonexistence,” thereby “set[ting] the question at rest in favor of admissibility” of such evidence. See Fed. R. Evid. 803(7) & advisory com- mittee’s note (1972) (emphasis added); see also United States v. Rich, 580 F.2d 929, 937–38 (9th Cir. 1979) (stating that Rules 803(7) and 803(10) “resolved the issue” of admissibility by “treat[ing] evidence of the absence of entries . . . as an exception to the hearsay rule”). 12 Similar requirements have been adopted in state statutes. See, e.g., Cal. Evid. Code § 1272 (2013) (admit- ting evidence of the absence of a business record to prove the nonoccurrence of an event if “all such acts, conditions, or events” were recorded, and “the absence of a record . . . is a trustworthy indication that the act or event did not occur”). AZ v. SHINSEKI 27 Evidence that an entry is missing from a deficient record is inadmissible under the Rules. See, e.g., Robinson, 544 F.2d at 114–15; see also United States v. Rith, 164 F.3d 1323, 1336 (10th Cir. 1999) (collecting cases). There is no reason why the well-established common- law principle concerning the inadmissibility of unreliable record evidence should not be followed here. Indeed, the Veterans Court has held in similar circumstances that where a condition would not normally have been recorded, “the Board may not consider the absence of [administra- tive record] evidence as substantive negative evidence” of that condition. See Buczynski v. Shinseki, 24 Vet. App. 221, 224 (2011). In Buczynski, the Veterans Court ruled that the Board erred in basing its determination that a veteran’s skin condition was “not exceptionally repug- nant” on the absence of a medical record documenting such a condition: [T]here [was not] any medical reason why a doctor would be expected to comment on the repugnance of [the veteran’s] condition. Therefore, this is not a situation where the silence in regard to a condi- tion can be taken as proof that a doctor did not ob- serve the symptom. Id. at 224 (emphasis added); see also id. (citing Federal Rule of Evidence 803(7)). Similarly, in Horn v. Shinseki, the Veterans Court held that the Board erred in relying on the lack of service records documenting a medical condition because “there is no evidentiary foundation, or even a logical reason to suppose” that the condition would have been recorded during the veteran’s treatment. 25 Vet. App. 231, 239 n.7 (2012). The Veterans Court recent- ly applied the rule stated in Buczynski to a PTSD claim based on alleged in-service sexual assault. See Nesbit- Netcliff v. Shinseki, No. 11–0950, 2012 WL 3205518 (Vet. App. Aug. 8, 2012) (nonprecedential). The Veterans Court held that where service medical records “are silent as to 28 AZ v. SHINSEKI the appellant having being [sic] raped,” the records “do not contradict [the veteran’s] statement” that she was raped, because “[i]t is not surprising that a rape victim would be silent regarding the fact that she had been raped.” Id. at *4 (emphasis in the original) (reversing and remanding). These cases are consistent with our holding in Fagan, which addressed the Board’s treatment of evidence lack- ing probative value (an inconclusive medical report re- counting a doctor’s inability to reach a conclusion one way or the other). Fagan, 573 F.3d at 1289–90. We held that such evidence provides neither positive nor negative support for service connection. Therefore, it is not pertinent evidence, one way or the other, regarding service connection. Id. at 1289 (emphasis added) (citation omitted). Like the inconclusive medical report in Fagan, the ab- sence of a report of an unreported sexual assault is too ambiguous to have probative value. Because the alleged assaults were not reported to military authorities, no reasonable person could expect records documenting the assaults to exist, or infer that the absence of such records tends to prove the assaults did not occur. Thus, the absence of records “provides neither positive nor negative support for service connection,” see id. at 1289, and is “not pertinent evidence, one way or the other,” to that deter- mination, see id. In sum, basic evidentiary principles preclude treating the absence of a record of an unreported sexual assault as evidence of the nonoccurrence of the assault. Therefore, we agree with the appellants that where an alleged sexual assault, like most in-service sexual assaults, is not report- ed, the absence of service records documenting the alleged AZ v. SHINSEKI 29 assault is not pertinent evidence that the assault did not occur. See Fagan, 573 F.3d at 1289–90. 13 II The second question is whether the veterans’ admit- ted failure to report the sexual assaults to superiors in the chain of command is itself pertinent evidence that the assaults did not occur. In both AZ and AY, the Board noted, and may have given weight to, the failure of the veterans to report the rapes to military authorities. It is true that in the context of criminal rape trials, courts historically presumed that “it [wa]s so natural as to be almost inevitable” that a rape victim would “make immediate complaint [about the rape] to her mother or other confidential friend.” See Baccio v. People, 41 N.Y. 265, 268 (1869); see also 4 John Henry Wigmore, Evidence in Trials at Common Law § 1135, at 298–301 & nn.1–2 (James H. Chadbourn rev., 1972) (“4 Wigmore on Evi- dence”) (collecting cases); State v. Hill, 578 A.2d 370, 374– 77 (N.J. 1990) (discussing the history of this presump- tion). Therefore, it was thought that a victim’s failure to promptly report the rape to anyone was a “suspicious inconsistency”: 13We do not suggest that service records cannot otherwise be relied on by the VA, for example, to show AZ had a disciplinary problem prior to the alleged assault, or to establish that AY was never treated for attempted suicide at a military hospital. The dissent suggests “the absence of a report in [AZ and AY’s] records cannot possibly have any bearing on the outcome of the cases before us.” Dissent at 6. But if an absence of a documentary record is irrelevant, the VA should not have relied on that absence in reaching its decision. 30 AZ v. SHINSEKI It has already been seen that the fact of a failure to speak when it would have been natural to do so is in effect an inconsistent statement or self- contradiction . . . . [Where] the accused denies [that the rape oc- curred], its very commission thus coming into is- sue, the circumstance that at the time of the alleged rape the woman said nothing about it to anybody constitutes in effect a self-contradiction of the above sort. It was entirely natural, after becoming the victim of an assault against her will, that she should have spoken out. That she did not, that she went about as if nothing had hap- pened, was in effect an assertion that nothing vio- lent had been done. Thus the failure of the woman, at the time of an alleged rape, to make any complaint could be offered in evidence (as all concede) as a virtual self-contradiction discrediting her present testi- mony. 4 Wigmore on Evidence, supra, § 1135 at 298 (emphases in the original) (citation omitted); see also, e.g., Baccio, 41 N.Y. at 268. There are several reasons why this common law theo- ry of pertinence is inapposite here. First, the failure to report the rape was considered relevant only where the victim failed “to make any complaint” to anyone, not where the victim failed to make an official report to a supervisor, an employer, or the police. See 4 Wigmore on Evidence, supra, § 1135 at 298 (emphasis in the original). Indeed, the cases generally asked whether the victim had reported the rape to her family or friends. See, e.g., State v. Balles, 221 A.2d 1, 5 (N.J. 1966) (admitting “proof that the violated victim complained within a reasonable time to someone she would ordinarily turn to for sympathy, protection and advice”—specifically, her mother); Thomas AZ v. SHINSEKI 31 v. State, 87 S.E. 8, 9 (Ga. 1915) (admitting testimony that the victim “made complaint to those to whom complaint of such an occurrence would naturally be made”— specifically, her parents); Baccio, 41 N.Y. at 268 (stating that a victim generally “will make immediate complaint [of the rape] to her mother or other confidential friend”); State v. De Wolf, 8 Conn. 93, 100 (1830) (“If a female testifies, that [she has been raped], an enquiry is, at once, suggested, why it was not communicated to her female friends.”). In other words, under the common law, a rape victim was expected to complain to the very categories of individuals specified in section 3.304(f)(5) as potential sources of supporting evidence—the same kinds of indi- viduals who submitted supporting statements here. We are unaware of cases holding that the failure to report the alleged rape to the police or a workplace supervisor tends to show that the rape did not occur. Second, modern courts are skeptical that the lack of a prompt report has probative value because a rape victim will “natural[ly]” report the assault. Rather, they recog- nize that “[t]he overwhelming body of current empirical studies, data, and other information establishes that it is not inherently ‘natural’ for the victim to confide in some- one or to disclose, immediately following commission of the offense, that he or she was sexually assaulted.” See People v. Brown, 883 P.2d 949, 956 (Cal. 1994); see also, e.g., Commonwealth v. Licata, 591 N.E.2d 672, 674 (Mass. 1992) (“[L]ack of a fresh complaint in no way necessarily . . . implies lack of rape.”). It is now known that sexual assault is generally underreported. See, e.g., S. Rep. No. 102–197, at 44 (1991) (“Consider the typical rape victim. . . . Few even report the crime.”); id. at 38 (stating that rape is among “the most underreported crimes in Ameri- ca”). Courts recognize that many victims are afraid to report the crime because they fear stigma, humiliation, or 32 AZ v. SHINSEKI the ordeal of testifying about the attack. 14 To the extent that courts permit criminal defendants to cross-examine prosecution witnesses, including rape victims, about omissions that are arguably inconsistent with the prose- cution’s case, that practice rests on a unique consideration inapplicable to a civil benefits proceeding: the overriding need to safeguard a defendant’s right to a fair trial and Sixth Amendment right of confrontation. See Olden v. Kentucky, 488 U.S. 227, 230–32 (1988) (holding that the defendant in a rape trial had a constitutionally protected right to cross-examine the complainant on an issue the trial court excluded as too prejudicial). In the context of a non-adversarial civil benefits proceeding, such as a VA benefits proceeding, there is no criminal defendant, and these constitutional concerns are inapposite. Third, as a general matter, courts have held in both criminal and civil contexts that, similar to the rule on omissions from unreliable records, where there is reason 14 For example, in Commonwealth v. Morgan, the Supreme Court of Pennsylvania pointed out that not all rape victims report: Women who are violated as [the victim] claims she was may be divided into three classes: (1) those who furiously turn on their attacker with any weapon available or even with their bare hands; (2) those who because of a natural dread of having the indignity publicized, remain silent; and (3) those who, after counseling for a day or two with husband and friends, invoke the law. . . . The members of all three classes are entitled to respect. 58 A.2d 330, 334 (Pa. 1948); see also id. at 333 (stating that “[m]ost women are reluctant to make a complaint in rape cases” because they fear “notoriety” and the ordeal of testifying). AZ v. SHINSEKI 33 to suspect that no report or other statement would have been made, testimony about the failure to make a report or statement is inadmissible. As the Supreme Court has recognized, “[i]n most circumstances silence is so ambigu- ous that it is of little probative force.” United States v. Hale, 422 U.S. 171, 176 (1975). When evaluating whether the failure to make a statement may be probative of a material fact, “the underlying test is, would it have been natural for the person to make the assertion in question?” 3A John Henry Wigmore, Evidence in Trials at Common Law § 1042 at 1058 (James H. Chadbourn rev., 1970) (emphasis added). Thus, silence is commonly thought to lack probative val- ue on the question of whether a person has ex- pressed tacit agreement or disagreement with contemporaneous statements of others. . . . Fail- ure to contest an assertion . . . is considered evi- dence of acquiescence only if it would have been natural under the circumstances to object to the assertion in question. Hale, 422 U.S. at 176 (citation omitted); see also Jenkins v. Anderson, 447 U.S. 231, 239 (1980) (“Common law traditionally has allowed witnesses to be impeached by their previous failure to state a fact in circumstances in which that fact naturally would have been asserted.”) (emphasis added). Thus, where an individual would not be expected to make a statement, courts generally do not admit testimo- ny that no statement was made. 15 Similarly, courts 15For example, the Supreme Court of Georgia held in Sherling v. Continental Trust Co. that the trial court erred by allowing testimony that a decedent never said anything to [the witness] about a con- tract [he allegedly made with the plaintiff and a third party]. This evidence should have been ex- 34 AZ v. SHINSEKI generally refuse to allow a witness to testify that he was never told a fact, where one would not expect that witness to have been informed. 16 It is generally recognized that cluded . . . on the ground that it was irrelevant and immaterial, and no circumstances were shown requiring [the decedent] to make any statement about the contract referred to or to admit or deny the making of the same. 165 S.E. 560, 561 (Ga. 1932) (emphasis added) (headnote). And in Lake Drainage Commissioners v. Spencer, where the fact in question was whether or not a summons had been served on the defendants’ mother, the Supreme Court of North Carolina held that the trial court erred by admitting the defendants’ testimony that they had never heard their mother say she was served with summons as tending to prove that no service was made, because the mother’s failure to mention the summons “prove[d] noth- ing, and if it proved anything, would tend to show that she had been served.” 93 S.E. 435, 435 (N.C. 1917). 16 For example, in a suit by a stepdaughter alleging abuse by her stepmother, the Court of Appeals of Indiana held that testimony by the defendant’s son to the effect that his stepsister never mentioned the alleged abuse in his presence was inadmissible to show that the abuse did not occur, because “[t]here was nothing obligatory upon [the stepsister] to discuss the fact of the trouble between herself and her stepmother with the witness, or with other persons in his presence.” Treschman v. Treschman, 61 N.E. 961, 964 (Ind. Ct. App. 1901). Courts have been more accepting of the proposition that because close family members would be told of a loved one’s illness or injury, the fact that they were not told about an illness or injury tends to show that it never occurred. See, e.g., Fidelity Serv. Ins. Co. v. Jones, 191 So.2d 20, 29-30 (Ala. 1966) (allowing testimony that the deceased never told AZ v. SHINSEKI 35 [s]ilence may imply assent to [a third party’s statement], but on certain conditions only. The general principle of relevancy tells us that the in- ference of assent may safely be made only when no other explanation is equally consistent with si- lence . . . . 4 Wigmore on Evidence, supra, § 1071, at 102; see also, e.g., Vail v. Strong, 10 Vt. 457, 463 (1838) (“The mere silence of the party creates no evidence, one way or the other. There are, indeed, cases, where the silence of the party creates a presumption or inference against him; but this presumption derives all its force from the circum- stances, under which the statement is made . . .”). And in contract disputes, evidence that other customers failed to complain is usually inadmissible to show that the goods were not defective, because there are numerous reasons customers might not complain about receiving defective goods. 17 family members he was unwell prior to an alleged acci- dent); Fogg v. Or. Short Line R.R. Co., 1 P.2d 954, 956-57 (Utah 1931) (allowing testimony from the plaintiff’s wife that he did not complain of any injury from a prior acci- dent). 17See, e.g., S. J. Van Lill Co. v. Frederick City Packing Co., 141 A. 898, 903–04 (Md. 1928) (holding the absence of complaints from another purchaser inadmissi- ble because, given the many possible reasons for not complaining, the inference that the goods were acceptable from the later purchaser’s failure to complain was “too remote to warrant its recognition as judicial proof”); Siegel, King & Co. v. Penny & Baldwin, 2 S.W.2d 1082, 1084 (Ark. 1928) (observing that other purchasers may well have been “damaged without[] complaining”). In Vermont Food Industries, Inc. v. Ralston Purina Co., the Second Circuit held that the trial court had properly 36 AZ v. SHINSEKI Fourth, there is a unique deterrent to reporting in- service sexual assaults to military authorities that is not usually present in criminal cases: that is the fear of reprisals. A servicemember who is raped by another servicemember, and reports that rape within the chain of command, is making the report to a person professionally (and perhaps personally) associated with the rapist. Thus, many servicemembers fear that the act of reporting a rape to military authorities will subject them to person- al and professional reprisals. See, e.g., 2010 SAPRO Report, supra, at 19–21, 25, 41, 95 (discussing reprisals and the fear of reprisals); Office of the Inspector General of the DoD, supra, at 17–20 (similar); Gov’t Accountability Office, supra, at 14 (similar). This fear creates incentives not to report rapes to military authorities that do not exist in the typical criminal case in the civilian context. Finally, the veteran’s benefits system is based on “so- licitude for the claimant.” Walters v. Nat’l Ass’n of Radia- tion Survivors, 473 U.S. 305, 311 (1985); see also Hodge v. West, 155 F.3d 1356, 1362 (Fed. Cir. 1998) (“This court and the Supreme Court both have long recognized that the character of the veterans’ benefits statutes is strongly and uniquely pro-claimant.”). This solicitude argues against the use of evidence against a claimant when it has prohibited questions about customer complaints to a company’s home office because none of these questions was probative of the real point the appellant was trying to get before the jury, namely, that no complaints concerning the appellant’s product had in fact been made. Many complaints might have been made locally, but re- solved there. In other words, the absence of for- mal complaints at the home office might not be indicative of the situation in the field. 514 F.2d 456, 465 (2d Cir. 1975). AZ v. SHINSEKI 37 doubtful pertinence. To the extent that Congress has relaxed evidentiary requirements in the VA context, it did so to benefit, not penalize, claimants. See generally H.R. Rep. No. 100-963, at 13–14 (1988), reprinted in 1988 U.S.C.C.A.N. 5782, 5795–96. Here, the Secretary, relying on empirical evidence, has expressly recognized that the failure to report to authorities is typical of active duty victims of sexual assault. See 65 Fed. Reg. at 61,132. The Secretary expects that many assaults will not have been reported to authorities, for reasons unrelated to the merits of the claim. Penalizing assault victims for that failure would hardly comport with a system in which “the importance of systemic fairness and the appearance of fairness carr[y] great weight.” Hodge, 155 F.3d at 1363. For all these reasons, we conclude that the VA may not treat a claimant’s failure to report an alleged sexual assault to military authorities as pertinent evidence that the sexual assault did not occur. III In each of these cases, the Board appears to have treated the absence of service records documenting the alleged assaults, and the absence of a report to military authorities, as evidence that the assaults did not occur. In AZ, the Veterans Court expressly acknowledged that the Board had given the absence of service records docu- menting the alleged assault evidentiary weight, but held that it was not error to do so, because “the Board is per- mitted to weigh the absence of corroborating records and documents against the lay evidence.” AZ, No. 10-2393, slip op. at 5 (Vet. App. Nov. 28, 2011) (emphasis added). In AY, the RO relied heavily on the absence of service records of the assault, but the Board mentioned the absence of service records only briefly, and the Veterans Court did not clearly decide whether such reliance was 38 AZ v. SHINSEKI proper. 18 In both cases, the Board and Veterans Court also appeared to treat the veteran’s failure to report the assault to military authorities as evidence that the as- sault did not occur. We remand to the Veterans Court for further proceed- ings. In each case, the Veterans Court should consider to what extent the Board improperly relied on the absence of service records documenting the alleged sexual assaults, or on the veterans’ failure to contemporaneously report the alleged sexual assaults to military authorities, as pertinent evidence that the alleged assaults did not occur. Because such an approach is unsupported by the applica- ble statute and regulations, contradicted by the empirical evidence, and contrary to general evidence law, the Vet- erans Court should consider whether remand to the Board 18 We note that in both cases, the VA implied that the lay statements lacked probative value because they were not made by eyewitnesses to the alleged sexual assaults. See AZ, No. 06-08 672, slip op. at 15 (Bd. Vet. App. Jan. 22, 2010); AY J.A. 105 (Mar. 2006 Rating Deci- sion). However, we do not think the Veterans Court relied on any such interpretation of section 3.304(f)(5), which would clearly be incorrect. The regulation express- ly authorizes the submission of lay statements from individuals unlikely to have any opportunity to witness an in-service sexual assault, including clergy and the veteran’s family. See 38 C.F.R. § 3.304(f)(5). As the VA’s own adjudication manual acknowledges, sexual assaults are unlikely to be witnessed by third parties. See VA Adjudicative Procedures Manual Rewrite M21-1MR, Pt. III, Subpart iv, Ch. 4, Sec. H(32)(a) (2011) (“When deter- mining the occurrence of stressors to establish service connection for PTSD, consider the following: . . . the trauma may be experienced alone, such as in cases of rape or assault . . . .”). AZ v. SHINSEKI 39 is required so that the cases can be re-adjudicated in the light of the correct standard. VACATED AND REMANDED COSTS Costs to appellants. United States Court of Appeals for the Federal Circuit ______________________ AZ, Claimant-Appellant, v. ERIC K. SHINSEKI, Secretary of Veterans Affairs, Respondent-Appellee. ______________________ 2012-7046 ______________________ Appeal from the United States Court of Appeals for Veterans Claims in No. 10-2393, Judge Lawrence B. Hagel. ---------------------- AY, Claimant-Appellant, v. ERIC K. SHINSEKI, Secretary of Veterans Affairs, Respondent-Appellee. ______________________ 2012-7048 ______________________ 2 AZ v. SHINSEKI Appeal from the United States Court of Appeals for Veterans Claims in No. 10-2390, Chief Judge Bruce E. Kasold. ______________________ MOORE, Circuit Judge, dissenting. The majority adopts a general, categorical rule of law that, at least in the military, when a person claims sexual assault years after the alleged incident, the absence of a contemporaneous report to officials able to act against the alleged perpetrator is irrelevant to assessing the credibility of the claim. According to the majority, failure to report a sexual assault is not relevant to whether or not the assault took place. The majority is wrong. AZ and AY are sympathetic claimants, but our jurisdiction prevents us from reviewing fact findings or even applications of law to fact. And as such, the majority was forced to adopt this new, categorical rule of law that is at odds with other courts, which have consistently found that non-reporting of sexual assault is relevant. I respectfully dissent. Under the legal standard applicable to veterans’ benefits cases, a veteran’s medical and service records and “all pertinent medical and lay evidence” must be considered. 38 C.F.R. § 3.303(a). In short, the VA must consider all relevant evidence. Evidence is relevant if “it has any tendency to make a fact more or less probable than it would be without the evidence.” Fed. R. Evid. 401(a) (emphasis added); see also id. R. 402. The majority never applies this universally accepted standard of relevance. The majority accepts that the existence of a report to officials is relevant and admissible: it has a tendency to make the fact of the rape more probable than it would be without the evidence. Inexplicably, it denies the converse: that the failure to report a rape has any tendency to make the occurrence of the rape less probable than it would be without the failure to report. It simply AZ v. SHINSEKI 3 defies credulity to conclude that a veteran’s failure to report a sexual assault has zero probative value and therefore fails the low relevancy threshold in every case. See Forshey v. Principi, 284 F.3d 1335, 1358 (Fed. Cir. 2002) (en banc) (“[T]he definition of evidence encompasses ‘negative evidence,’ which tends to disprove the existence of an alleged fact . . . .”). The majority’s decision hinges on a study finding that only 15% of sexual assaults in the military were reported between 2006 and 2012. There are several problems with this reliance. First, even if this number accurately reflects reporting of rape in the military in the 1970s and 1980s, when these incidents allegedly occurred, the fact of reporting or non-reporting still meets the low relevance threshold: namely, non-reporting has some tendency to make the fact that the rape occurred less probable. It may well be that a low incidence of reporting would result in little weight being given to the non-reporting, but it doesn’t render that evidence irrelevant. Second, as the majority acknowledges, its 15% figure applies to all sexual assaults. It may be, however, that the reporting rate for the particularly severe incidents of sexual assault of the sort at issue in these appeals (rape) is higher. We do not know, and the majority admits that it has no idea. Maj. Op. at 19–20. The larger problem is that none of the studies cited by the majority were a part of the record below, and the VA was not given an opportunity to explain their import to the cases before us. Even if we had the authority to create new rules to protect victims of sexual assault, a job for Congress not the courts, we would need to gather additional data and solicit further input from the affected parties in order to reach the result the majority writes into law today. The majority justifies its blanket rule by cobbling together a “common law” of evidence from a hodge-podge of sources presenting unrelated circumstances. The cases 4 AZ v. SHINSEKI cited by the majority address adoption by silence of the statements of others, exclusion of hearsay evidence, and even the lack of customer complaints in breach of contract cases. But none of these cases support the proposition that the failure to complain of rape to the authorities is not relevant to the question of whether the rape occurred. We do not need to resort to cases of questionable significance because there exist on-point authorities dealing with sexual assault. These authorities make clear that the fact of non-reporting is routinely admitted into evidence. See, e.g., State v. W.B., 17 A.3d 187, 206–07 (N.J. 2011) (approving the jury instruction that “you may consider the silence/delayed disclosure along with all of the other evidence including [complaining witness’s] explanation for his/her silence/delayed disclosure when you decide how much weight to afford to [complaining witness’s] testimony”); 18 Pa. C.S.A. § 3105 (1995) (“Prompt reporting to public authority is not required in a prosecution [of sexual offenses]: Provided, however, That nothing in this section shall be construed to prohibit a defendant from introducing evidence of the complainant’s failure to promptly report the crime.”). Indeed, a major issue in this area of law is the admissibility of expert or other testimony to explain why victims of sexual assault do not report the crime or delay reporting, which suggests that the admissibility of non-reporting itself is not in question. See, e.g., State v. Obeta, 796 N.W.2d 282 (Minn. 2011); State v. Hicks, 535 A.2d 776 (Vt. 1987); State v. Rizzo, 640 N.W.2d 93 (Wis. 2002). Cf. Jensen v. Lawrence, 162 P. 40, 42 (Wash. 1916) (holding, in a civil rape case, that “[a] delay or even failure to complain at all may be accounted for by proof of circumstances excusing such delay”). Contrary to the majority’s suggestion, I find no indication that any of these authorities are motivated by special solicitude for criminal defendants. Evidence of non-reporting is admitted because it is relevant. AZ v. SHINSEKI 5 While most jurisdictions appear to have rejected the presumption that a failure to make a contemporaneous report means that the assault did not occur, the majority cites no authority denying the relevance (and admissibility) of that evidence. For example, the majority cites Commonwealth v. Licata, which states that the “lack of a fresh complaint in no way necessarily implies lack of rape.” 591 N.E.2d 672, 674 (Mass. 1992) (emphasis added). Licata does not say that the lack of a complaint to the authorities is categorically irrelevant, and neither do any of the other cases cited by the majority. For example, People v. Brown, 883 P.2d 949 (Cal. 1994), cited by the majority, directly contradicts the majority’s holding: “when the victim of an alleged sexual offense did not make a prompt complaint but instead disclosed the alleged incident only some time later, evidence of the fact and circumstances surrounding the delayed complaint . . . may be relevant to the jury’s evaluation of the likelihood that the offense did or did not occur.” Id. at 958. Accord Commonwealth v. Lane, 555 A.2d 1246, 1250 (Pa. 1989) (“The lack of a prompt complaint by a victim of a crime, although not dispositive of the merits of the case, may justifiably produce a doubt as to whether the offense indeed occurred, or whether it was a recent fabrication by the complaining witness.”). The majority’s real issue is with the way that the VA weighed the fact of non-reporting, not that fact’s admissibility. I may well agree with the majority that the VA clearly erred by according undue probative value to the fact that AZ and AY did not report the assaults to the authorities. 1 But we have no jurisdiction to review the 1I note that the majority’s opinion makes it appear as if the VA’s decisions in these cases rested largely on the fact of non-reporting. This is unfair to the VA, and simply inaccurate. The VA clearly considered other 6 AZ v. SHINSEKI VA’s weighing of the evidence. See King v. Shinseki, 700 F.3d 1339, 1345–46 (Fed. Cir. 2012); Madden v. Gober, 125 F.3d 1477, 1481 (Fed. Cir. 1997); see also 38 U.S.C. § 7292(d)(2). The majority’s additional rule that the absence of a report of sexual assault in the veteran’s records cannot be considered by the VA is completely inapposite to the appeals before us. AZ and AY admitted in the proceedings below that they did not report the alleged assaults to the authorities until long after they had occurred. “The Veteran has stated that she did not report sexual assault to military or civilian authorities.” AZ J.A. 29 (emphasis added); see AZ Br. 9; see also AY J.A. 19 (noting that AY stated at a hearing that “[s]he did not report the incident to police at the time that it occurred”). Because AZ and AY positively admitted that they made no contemporaneous report of rape, the absence of a report in their records cannot possibly have any bearing on the outcome of the cases before us. Furthermore, there is not the slightest basis for thinking that reports of rape that are made to the authorities do not appear in the records. The majority’s discussion of the relevance of absence of reports of rape is pure dicta. Even if these veterans’ records mattered, the majority’s conclusion that the absence of a report is irrelevant is unsupportable for similar reasons as the fact evidence. See, e.g., AZ J.A. 28 (“[T]he reports of misconduct occurred relatively consistently throughout the Veteran’s period of service, and . . . did not suddenly begin at point in time concurrent with the beginning of alleged assaults.”); id. at 31 (noting recent stressors unrelated to service); see also AY J.A. 22–23 (noting that AY’s demeanor while in service contradicted her claim of being depressed); id. at 23 (noting inconsistencies in AY’s accounts of the incident). AZ v. SHINSEKI 7 of non-reporting: the relevancy threshold is low, and the absence of reports clears it. The VA in these cases did exactly what the applicable statute commands—it gave “due consideration” to the service and medical records of the veterans claiming service-connected disability. 38 U.S.C. § 1154(a)(1) (2012); see also 38 C.F.R. § 3.303(a). Not only are the statute and regulations clearly contrary to the majority’s holding, but in Buchanan v. Nicholson, we held that “the lack of contemporaneous medical records may be a fact that the Board can consider and weigh against a veteran’s lay evidence.” 451 F.3d 1331, 1336 (Fed. Cir. 2006); see also id. at 1337 (“Nor do we hold that the Board cannot weigh the absence of contemporaneous medical evidence against the lay evidence of record.”). In doing so, Buchanan simply restated well-established law—the VA may consider the lack of contemporaneous medical records in the veteran’s file because § 3.303(a) grants it the power to do so. We cannot ignore this binding precedent. The majority attempts to justify its departure from the governing statute and regulations by arguing that the absence of reporting in a veteran’s records ought not be admissible because it would not meet hearsay exceptions codified in Rules 803(7) and 803(10) of the Federal Rules of Evidence. But veterans’ benefits cases allow hearsay, and thus evidence does not need to meet any of the exceptions to the rule against hearsay to become admissible. *** When Congress sees the need to protect victims of sexual assault, it acts. For example, Congress decided that evidence of an alleged sexual assault victim’s prior sexual history is categorically inadmissible to prove that the assault did not occur. Fed. R. Evid. 412(a)(1). Congress also amended the FREs to allow admission in 8 AZ v. SHINSEKI criminal cases of evidence that a defendant previously committed sexual assault. Id. R. 413(a). Congress did not, however, enact a rule stating that evidence of non- reporting of sexual assault is categorically inadmissible— be it in criminal, civil, or VA settings. Today, the majority usurps Congress’s role with its broad proclamation on the admissibility of certain evidence in the VA system. AZ and AY are sympathetic claimants. And as a judge, a woman, and a human being, I am dubious about the weighing of the evidence and the fact findings of the VA in this case. But the applicable statutes and basic principles of evidence law leave us without power to help them. We are not allowed to assess the probative value of a veteran’s failure to report a sexual assault to the authorities and mandate to the VA what inferences it may draw from that failure. Nor do we have jurisdiction to weigh the absence of a report of sexual assault in the veteran’s records. Solicitude for veterans does not justify making up rules as we go along.