JOHNSON V. CALIFORNIA (03-636) 543 U.S. 499 (2005)
321 F.3d 791, reversed and remanded.
[ O’Connor ]
[ Ginsburg ]
[ Stevens ]
[ Thomas ]
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version

Stevens, J., dissenting


No. 03—636



[February 23, 2005]

Justice Stevens, dissenting.

In my judgment a state policy of segregating prisoners by race during the first 60 days of their incarceration, as well as the first 60 days after their transfer from one facility to another, violates the Equal Protection Clause of the Fourteenth Amendment. The California Department of Corrections (CDC) has had an ample opportunity to justify its policy during the course of this litigation, but has utterly failed to do so whether judged under strict scrutiny or the more deferential standard set out in Turner v. Safley, 482 U.S. 78 (1987). The CDC had no incentive in the proceedings below to withhold evidence supporting its policy; nor has the CDC made any offer of proof to suggest that a remand for further factual development would serve any purpose other than to postpone the inevitable. I therefore agree with the submission of the United States as amicus curiae that the Court should hold the policy unconstitutional on the current record.

The CDC’s segregation policy1 is based on a conclusive presumption that housing inmates of different races together creates an unacceptable risk of racial violence. Under the policy’s logic, an inmate’s race is a proxy for gang membership, and gang membership is a proxy for violence. The CDC, however, has offered scant empirical evidence or expert opinion to justify this use of race under even a minimal level of constitutional scrutiny. The presumption underlying the policy is undoubtedly overbroad. The CDC has made no effort to prove what fraction of new or transferred inmates are members of race-based gangs, nor has it shown more generally that interracial violence is disproportionately greater than intraracial violence in its prisons. Proclivity toward racial violence unquestionably varies from inmate to inmate, yet the CDC applies its blunderbuss policy to all new and transferred inmates housed in double cells regardless of their criminal histories or records of previous incarceration. Under the CDC’s policy, for example, two car thieves of different races–neither of whom has any history of gang involvement, or of violence, for that matter–would be barred from being housed together during their first two months of prison. This result derives from the CDC’s inflexible judgment that such integrated living conditions are simply too dangerous. This Court has never countenanced such racial prophylaxis.

To establish a link between integrated cells and violence, the CDC relies on the views of two state corrections officials. They attested to their belief that double-celling members of different races would lead to violence and that this violence would spill out into the prison yards. One of these officials, an associate warden, testified as follows:

“[W]ith the Asian population, the control sergeants have to be more careful than they do with Blacks, Whites, and Hispanics because, for example, you cannot house a Japanese inmate with a Chinese inmate. You cannot. They will kill each other. They won’t even tell you about it. They will just do it. The same with Laotians, Vietnamese, Cambodians, Filipinos. You have to be very careful about housing other Asians with other Asians. It’s very culturally heavy.” App. 189a.

Such musings inspire little confidence. Indeed, this comment supports the suspicion that the policy is based on racial stereotypes and outmoded fears about the dangers of racial integration. This Court should give no credence to such cynical, reflexive conclusions about race. See, e.g., Palmore v. Sidoti, 466 U.S. 429, 432 (1984) (“Classifying persons according to their race is more likely to reflect racial prejudice than legitimate public concerns; the race, not the person, dictates the category”); Watson v. Memphis, 373 U.S. 526, 536 (1963) (rejecting the city’s plea for delay in desegregating public facilities when “neither the asserted fears of violence and tumult nor the asserted inability to preserve the peace was demonstrated at trial to be anything more than personal speculations or vague disquietudes of city officials”).

The very real risk that prejudice (whether conscious or not) partly underlies the CDC’s policy counsels in favor of relaxing the usual deference we pay to corrections officials in these matters. We should instead insist on hard evidence, especially given that California’s policy is an outlier when compared to nationwide practice. The Federal Bureau of Prisons administers 104 institutions; no similar policy is applied in any of them. Countless state penal institutions are operated without such a policy. An amici brief filed by six former state corrections officials with an aggregate of over 120 years of experience managing prison systems in Wisconsin, Georgia, Oklahoma, Kansas, Alaska, and Washington makes clear that a blanket policy of even temporary segregation runs counter to the great weight of professional opinion on sound prison management. See Brief for Former State Corrections Officials as Amici Curiae 19. Tellingly, the CDC can only point to two other States, Texas and Oklahoma, that use racial status in assigning inmates in prison reception areas. It is doubtful from the record that these States’ policies have the same broad and inflexible sweep as California’s, and this is ultimately beside the point. What is important is that the Federal Government and the vast majority of States address the threat of interracial violence in prisons without resorting to the expedient of segregation.

In support of its policy, the CDC offers poignant evidence that its prisons are infested with violent race-based gangs. The most striking of this evidence involves a series of riots that took place between 1998 and 2001 at Pelican Bay State Prison. That prison houses some of the State’s most violent criminal offenders, including “validated” gang members who have been transferred from other prisons. The riots involved both interracial and intraracial violence. In the most serious incident, involving 250—300 inmates, “Southern Hispanic” gang members, joined by some white inmates, attacked a number of black inmates.

Our judicial role, however, requires that we scratch below the surface of this evidence, lest the sheer gravity of a threat be allowed to authorize any policy justified in its name. Upon inspection, the CDC’s post hoc, generalized evidence of gang violence is only tenuously related to its segregation policy. Significantly, the CDC has not cited a single specific incident of interracial violence between cellmates–much less a pattern of such violence–that prompted the adoption of its unique policy years ago. Nor is there any indication that antagonism between cellmates played any role in the more recent riots the CDC mentions. And despite the CDC’s focus on prison gangs and its suggestion that such gangs will recruit new inmates into committing racial violence during their 60-day stays in the reception centers, the CDC has cited no evidence of such recruitment, nor has it identified any instances in which new inmates committed racial violence against other new inmates in the common areas, such as the yard or the cafeteria. Perhaps the CDC’s evidence might provide a basis for arguing that at Pelican Bay and other facilities that have experienced similar riots, some race-conscious measures are justified if properly tailored. See Lee v. Washington, 390 U.S. 333, 334 (1968) (Black, J., concurring). But even if the incidents cited by the CDC, which occurred in the general prison population, were relevant to the conditions in the reception centers, they provide no support for the CDC’s decision to apply its segregation policy to all of its reception centers, without regard for each center’s security level or history of racial violence. Nor do the incidents provide any support for a policy applicable only to cellmates, while the common areas of the prison in which the disturbances occurred remain fully integrated.

Given the inherent indignity of segregation and its shameful historical connotations, one might assume that the CDC came to its policy only as a last resort. Distressingly, this is not so: There is no evidence that the CDC has ever experimented with, or even carefully considered, race-neutral methods of achieving its goals. That the policy is unwritten reflects, I think, the evident lack of deliberation that preceded its creation.

Specifically, the CDC has failed to explain why it could not, as an alternative to automatic segregation, rely on an individualized assessment of each inmate’s risk of violence when assigning him to a cell in a reception center. The Federal Bureau of Prisons and other state systems do so without any apparent difficulty. For inmates who are being transferred from one facility to another–who represent approximately 85% of those subject to the segregation policy–the CDC can simply examine their prison records to determine if they have any known gang affiliations or if they have ever engaged in or threatened racial violence. For example, the CDC has had an opportunity to observe the petitioner for almost 20 years; surely the CDC could have determined his placement without subjecting him to a period of segregation.2 For new inmates, assignments can be based on their presentence reports, which contain information about offense conduct, criminal record, and personal history–including any available information about gang affiliations. In fact, state law requires the county probation officer to transmit a presentence report to the CDC along with an inmate’s commitment papers. See Cal. Penal Code Ann. §1203c (West 2004); Cal. Rule of Court 4.411(d) (Criminal Cases) (West Supp. 2004).

Despite the rich information available in these records, the CDC considers these records only rarely in assigning inmates to cells in the reception centers. The CDC’s primary explanation for this is administrative inefficiency–the records, it says, simply do not arrive in time. The CDC’s counsel conceded at oral argument that presentence reports “have a fair amount of information,” but she stated that, “in California, the presentence report does not always accompany the inmate and frequently does not. It follows some period of time later from the county.” Tr. of Oral Arg. 33. Despite the state-law requirement to the contrary, counsel informed the Court that the counties are not preparing the presentence reports “in a timely fashion.” Ibid. Similarly, with regard to transferees, counsel stated that their prison records do not arrive at the reception centers in time to make cell assignments. Id., at 28. Even if such inefficiencies might explain a temporary expedient in some cases, they surely do not justify a system-wide policy. When the State’s interest in administrative convenience is pitted against the Fourteenth Amendment’s ban on racial segregation, the latter must prevail. When there has been no “serious, good faith consideration of workable race-neutral alternatives that will achieve the [desired goal],” Grutter v. Bollinger, 539 U.S. 306, 339 (2003), and when “obvious, easy alternatives” are available, Turner, 482 U.S., at 90, the conclusion that CDC’s policy is unconstitutional is inescapable regardless of the standard of review that the Court chooses to apply.3

In fact, the CDC’s failure to demand timely presentence reports and prison records undercuts the sincerity of its concern for inmate security during the reception process. Race is an unreliable and necessarily underinclusive predictor of violence. Without the inmate-specific information found in the records, there is a risk that corrections officials will, for example, house together inmates of the same race who are nevertheless members of rival gangs, such as the Bloods and Crips.4

Accordingly, while I agree that a remand is appropriate for a resolution of the issue of qualified immunity, I respectfully dissent from the Court’s refusal to decide, on the basis of the record before us, that the CDC’s policy is unconstitutional.


1. The CDC operates 32 prisons, 7 of which house reception centers. All new inmates and all inmates transferring between prisons are funneled through one of these reception centers before they are permanently placed. At the centers, inmates are housed either in dormitories, double cells, or single cells (of which there are few). Under the CDC’s segregation policy, race is a determinative factor in placing inmates in double cells, regardless of the other factors considered in such decisions. While a corrections official with 24 years of experience testified that an exception to this policy was once granted to a Hispanic inmate who had been “raised with Crips,” App. 184a, the CDC’s suggestion that its policy is therefore flexible, see Brief for Respondents 9, strains credulity. There is no evidence that the CDC routinely allows inmates to opt-out of segregation, much less evidence that the CDC informs inmates of their supposed right to do so.

2. In explaining why it cannot prescreen new inmates, the CDC’s brief all but concedes that segregating transferred inmates is unnecessary. See Brief for Respondents 42 (“If the officials had all of the necessary information to assess the inmates’ violence potential when the inmates arrived, perhaps a different practice could be used. But unlike the federal system, where the inmates are generally in federal custody from the moment they are arrested, state inmates are in county custody until they are convicted and later transferred to the custody of the CDC”).

3. Because the Turner factors boil down to a tailoring test, and I conclude that the CDC’s policy is, at best, an “exaggerated response” to its asserted security concerns, see Turner v. Safley, 482 U.S. 78, 90 (1987), I find it unnecessary to address specifically the other factors, such as whether new and transferred inmates have “alternative means” of exercising their right to equal protection during their period of housing segregation, id., at 89. Indeed, this case demonstrates once again that “[h]ow a court describes its standard of review when a prison regulation infringes fundamental constitutional rights often has far less consequences for the inmates than the actual showing that the court demands of the State in order to uphold the regulation.” Id., at 100 (Stevens, J., concurring in part and dissenting in part)

4. The CDC’s policy may be counterproductive in other ways. For example, an official policy of segregation may initiate new arrivals into a corrosive culture of prison racial segregation, lending credence to the view that members of other races are to be feared and that racial alliances are necessary. While integrated cells encourage inmates to gain valuable cross-racial experiences, segregated cells may well facilitate the formation of race-based gangs. See Brief for Former State Corrections Officials as Amici Curiae 19 (citing evidence and experience suggesting that the racial integration of cells on balance decreases interracial violence).