ORDER
The panel has voted to deny Appellees’ petition for rehearing. Judge Reinhardt and Judge Graber have voted to deny the petition for rehearing en banc and Judge Skopil has recommended denying the petition for rehearing en banc.
The opinion is amended as follows:
(1) At Slip Op. 4197, second paragraph, line 4:
“one-third” is changed to “one-quarter”
(2) At Slip Op. 4197, second paragraph, line 6:
“two days” is changed to “three days”
(3) At Slip Op. 4198, top paragraph; line 1:
“two-thirds” is changed to “three-quarters” and the citation to In re Williams is deleted
(4) At Slip Op. 4198, top paragraph, line 3:
“five days” is changed to “six days”
[1134]*1134(5) At Slip Op. 4198, top paragraph, lines 9-12:
last sentence is deleted (“Accordingly ... would be superfluous”)
(6) At Slip Op. 4199, top paragraph, line 2:
“2” is changed to “3”
(7) At Slip Op. 4199, top paragraph, line 16:
“approximately a third as much as” is changed to “less than half that which”
(8) At Slip Op. 4200, line 10:
“20” is changed to “30” and “51” is changed to “34”
(9) At Slip Op. 4200, lines 11-12:
“nearly triple the” is changed to “double the” and n. 7 is deleted
(10) At Slip Op. 4201, top paragraph, line 3:
“20 served” is changed to “30 served” and “72 days” is changed to “48 days”
(11) At Slip Op. 4201, top paragraph, line 5:
“nearly triple” is changed to “more than double”
(12) At Slip Op 4210, top paragraph, line 9:
“33 days” is changed to “16 days”
(13) At Slip Op. 4210, top paragraph, line 12:
“47 days” is changed to “23 days” .
(14) At Slip Op. 4212, second paragraph is amended to read as follows:
MacFarlane and Fogle also challenge the counties’ good-performance policies, which raise the same constitutional question as the good-conduct policies, but require a different answer.
(15) At Slip Op. 4213, line 3:
“it is their” is deleted
(16) At Slip Op. 4213, line 4:
“that” is deleted
(17) At Slip Op. 4213, line 12: ■
“thus” is deleted
(18) At Slip Op. 4213, lines 25-33:
sentence is deleted (“In addition ... given the statutory maximum”)
OPINION
REINHARDT, Circuit Judge:
Petitioners Donald MacFarlane (“MacFarlane”) and James Fogle (“Fogle”) appeal the district court’s grant of summary judgment in favor of prison officials Kay Walter and Kenneth Ducharme in petitioners’ consolidated 28 U.S.C. § 2254 habeas corpus petitions. MacFarlane’s and Fogle’s state habeas petitions alleged that the Pierce and Clark County Jails’ “early release,” or “good conduct” and “good performance,” policies, as applied to them, violate equal protection and due process. These early-release policies prevent county pre-sentence detainees, such as MacFarlane and Fogle, who are unable to afford bail and are ultimately sentenced to a state facility operated by the Department of Corrections, from earning the same early-release credit as prisoners who are financially able to post bail and thus serve their entire sentences in a state facility. We must determine whether the Washington courts’ denial of MacFarlane’s and Fogle’s petitions was “contrary to” or an “unreasonable application of’ clearly established federal law. See 28 U.S.C. § 2254(d). Because we conclude that the denial was contrary to clearly established federal law, we reverse in part, and affirm in part, the decision of the district court, and order that the petitioners’ writs of habeas corpus be granted.
I.
A. The Washington Sentencing and “Early Release” System
The State of Washington operates a determinate sentencing system, under which persons convicted of felonies receive a sentence of a specific number of months based upon a sentencing grid contained in the Sentencing Reform Act of 1981. Wash. Rev.Code Chapter 9.94A. When a sentence exceeds twelve months, the prisoner must serve the sentence in a state facility. RCW 9.94A.190(1). Prior to sentencing, however, many prisoners who are ultimately confined in a state facility serve time in a county jail because they have not [1135]*1135posted, or have been denied, bail. The time actually served in county jail as pretrial detainees is ultimately credited against their sentences.
In addition to being reduced for time served, a sentence imposed under the Sentencing Reform Act is reduced when a prisoner is credited with early-release time, either for good conduct or for good performance. In In re Mota, 114 Wash.2d 465, 472, 788 P.2d 538 (1990), the Washington Supreme Court held that equal protection requires that all prisoners be eligible for early-release credit for time served, both in pre-sentence detention and post-sentence incarceration. Mota did not, however, concern the issue presented here — whether all prisoners must be eligible for the same amount of early-release credit. The Washington legislature, in accordance with In re Mota, authorized correctional agencies, both state and county, to develop procedures for implementing a system of granting early-release credit. Wash. Rev.Code 9.94A.150(1). Pursuant to the statutory mandate, the Department of Corrections, whose policies apply only in state facilities, and the authorities who operate the county jails, have developed independent early-release credit policies. Ultimately, however, in the case of prisoners like MacFarlane and Fogle, who are considered to have served part of their sentences in a county jail and part in a state facility, it is the Department of Corrections which must compute the total amount of early-release credits to be applied to their sentences, combining the amount of early-release credit certified by the county jail with the amount of early-release credit accumulated at the state facility. See In re Williams, 121 Wash.2d 655, 658-59, 853 P.2d 444 (1993).
1. Department of Corrections’ Early-Release Policy
The Department of Corrections’ early-release credit system, applicable in state facilities, consists of two types of credit: good conduct and good performance.1 A prisoner in a state facility can earn good-conduct credit of up to one-quarter of his sentence — a prisoner, that is, can earn one day of good-conduct credit for every three days served. In other words, a prisoner who earns the maximum credit serves only three-quarters of the sentence imposed. In addition, a prisoner can earn good performance credit of up to one day for every six days served, if he participates in work, academic, or treatment programs. The Department of Corrections then combines the total good-conduct and good-performance credit to determine the final sentence reduction, but only up to the statutory maximum of one-third of the imposed sentence. Wash. Rev.Code 9.94A.150(1).
2. Clark County and Pierce County Early Release Policies
The Clark County Jail and Pierce County Detention and Corrections Center (the county jails at issue in this case) used tiered credit systems to award early release credits at the time the petitions were filed: All general population inmates (like MacFarlane and Fogle) were eligible for good-conduct credit at a rate of 15% of their sentences.2 This means, of course, that they were able to earn substantially [1136]*1136less good-time credit than prisoners confined in state facilities. Specifically, as construed by the Washington Supreme Court, the counties’ policies provided that detainees received 1 day of credit for every 5-2/3 days of good conduct, instead of 1 for every 3.3 Some inmates can also receive additional good-performance credit through work programs. When the credit from the good-conduct and performance programs is combined, Clark County Jail allows inmates to achieve a maximum early-release credit of 30% of the sentence, and Pierce County Jail allows up to either 30% or 20%, depending on the type of work or other program involved. However, because pre-sentence detainees are not eligible to participate in work or other good performance programs, see In re Fogle, 128 Wash.2d 56, 63, 904 P.2d 722 (1995), they can earn only good-conduct and not good-performance credit. Accordingly, the maximum early-release time that can be earned by pre-sentence detainees, such as Fogle and MacFarlane, in these county jails is less than half that which can be earned by defendants who are able to post bail and begin to serve their time in a state facility after sentencing.4
B. Petitioners MacFarlane and Fogle
Fogle and MacFarlane were unable to afford to post bail because of indigence and were, therefore, detained in county jails prior to trial and sentencing.5 In re Fogle, 128 Wash.2d at 63, 904 P.2d 722. Both Fogle and MacFarlane received the maximum 15% good-conduct credit as general population county jail prisoners. Neither was eligible to participate in any program that would have allowed him to accumulate good-performance credit in the county jails. Fogle spent 102 pre-sen-tence days in the Pierce County Jail on a 60-month sentence and received 17 days of early-release time upon transfer to the state facility.6 Had Fogle posted bail, not been held in custody until after sentencing, and then been committed to the state facility after sentencing, he would have earned 10 days for every 30 served, or 34 days of early-release credit for his good conduct during his initial period of custody, or double the time he was credited with by the county jail. MacFarlane spent 144 pre-sentence days in the Clark County Jail prior to receiving concurrent 72- and 8-month sentences; the jail certified 21 days of early release credit.7 Had [1137]*1137MacFarlane been able to post bail, not been held in custody until after sentencing and then been committed to the state facility, he would have earned 10 days for every 30 served, or 48 days of early-release credit for good conduct during his initial 144 days in custody, or more than double the time he was credited with by the county jail. Petitioners challenge the constitutionality of the lesser award of early-release credit for the time they were detained in jail and seek credit equal to that which they would have received had they been able to post bail and serve their full sentences in state facilities.
C. Prior Habeas Proceedings
MacFarlane and Fogle filed separate personal restraint petitions in the Washington Court of Appeals, challenging the county jails’ early-release policies as viola-tive of equal protection, due process, and double jeopardy.8 They contended that they were denied equal protection by not being awarded the amount of early-release time that state prisoners receive — one-third of their imposed sentences, Wash. Rev.Code 9.94A.150(1). The Court of Appeals rejected Fogle’s equal protection claim, holding that county jails have plenary authority to develop procedures by which their prisoners can earn early-release credits and that Pierce County’s system of awarding such credits was constitutional. The Washington Supreme Court granted Fogle’s petition for discretionary review, certified MacFarlane’s petition, which was then pending in the Court of Appeals, and consolidated the cases for review.
In a 5-to-4 decision, the Washington Supreme Court determined that the county jails’ policies granting pre-sentence detainees substantially less early release credit than that afforded under the state policy were statutorily and constitutionally permissible. In re Fogle. In resolving the equal protection claim, the court relied solely on earlier Washington cases, providing no citations to any federal cases. The court decided to apply the intermediate scrutiny standard, on the assumption that MacFarlane and Fogle were eligible for bail but financially unable to provide it, and that a semi-suspect classification, indi-gency, thus served as the basis for the disparate treatment. 128 Wash.2d at 62, 904 P.2d 722. The court concluded that the counties’ substantial interest in maintaining prisoner discipline sufficiently justified disparate treatment of pre-trial detainees to overcome the equal protection challenge, that disparate policies regarding the amount of good-conduct and good-performance credit to be awarded did not result in multiple punishment so as to trigger double jeopardy protection, and that the jails established and observed their written policies governing such credits in a manner consistent with due process requirements. Id. at 64-66, 904 P.2d 722.
The dissent analyzed the equal protection issue by examining whether the disparate treatment “may fairly be viewed as furthering a substantial interest of the State,” citing Plyler v. Doe, 457 U.S. 202, 217-18, 102 S.Ct. 2382, 72 L.Ed.2d 786 (1982). It found an equal protection violation, because it “fails to see how awarding a person who makes bail a more generous earned release credit than one who does not has any effect on jail discipline.” 128 Wash.2d at 72, 904 P.2d 722.
In August of 1996, MacFarlane and Fo-gle filed habeas corpus petitions in the U.S. District Court for the Western District of Washington; the petitions were then consolidated. The Magistrate Judge issued a Report and Recommendation recommending that respondents’ motion for summary judgment be granted, on the assumption that the intermediate level of scrutiny applied by the Washington Su[1138]*1138preme Court was consistent with Bearden v. Georgia, 461 U.S. 660, 103 S.Ct. 2064, 76 L.Ed.2d 221 (1983), the clearly established federal law. District Judge Franklin D. Burgess issued a brief order approving the Report and entered judgment for respondents. Petitioners appealed.
II.
Petitioners argue that the Washington Supreme Court’s decision was “contrary to” clearly established federal law and, thus, under the Antiterrorism and Effective Death Penalty Act of 1996 (“AEDPA”), their applications for writs of habeas corpus must be granted. We agree that the state court’s decision was contrary to clearly established federal law.
AEDPA amended 28 U.S.C. § 2254(d) to provide that no habeas relief may be granted to a person in custody pursuant to a judgment of a state court unless the underlying state adjudication: (1) resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established federal law, as determined by the Supreme Court of the United States; or (2) resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence presented in the state court proceeding. 28 U.S.C. § 2254(d). Neither party challenges the state court’s determination of the facts. Our inquiry, thus, focuses first on whether there was clearly established federal law governing the issue presented here: whether the allowance of lesser early-release credits to defendants detained pre-trial in county jails because of financial inability to post bail than to more fortunate defendants, whose financial resources permitted them to wait to begin serving their time until after commitment, post-sentencing, to a state correctional facility, violates equal protection. If there was clearly established federal law that applies to this question, then we must next decide whether the Washington Supreme Court’s decision was contrary to, or constituted an unreasonable application of, it.
III.
In Bearden, 461 U.S. at 671-72, 103 S.Ct. 2064, the Supreme Court held that incarceration should not be imposed on an indigent individual solely because of that person’s inability to pay. We conclude that Bearden, decided in 1983, and grounded in thirty years of precedents, was clearly established law governing the issue presented here at the time of the Washington Supreme Court decision denying petitioners’ state habeas petitions.9
The petitioner in Bearden pleaded guilty to burglary and theft by receiving stolen property. The trial court sentenced him to three years’ probation for the burglary charge, a concurrent one-year term of probation on the theft charge, and a $500 fine and $250 in' restitution. Petitioner was to pay $100 on the day of sentencing, $100 on the following day, and the $550 balance within four months. Petitioner paid the first $200, but then lost his job, was unable to find other work and, therefore, notified the probation office that he would be late with his payment of the balance. After an evidentiary hearing, the trial court revoked probation for failure to pay the balance of the fine and restitution, entered a conviction, and sentenced petitioner to serve the remaining portion of the probationary period in prison. The state courts rejected petitioner’s claim that imprisoning him for inability to pay the fine violated [1139]*1139the Equal Protection Clause of the Fourteenth Amendment.
In analyzing the case, the Supreme Court stressed the balance between the goal of punishment and the impermissibility of imprisoning a defendant solely because of a lack of financial resources. Id. at 661, 103 S.Ct. 2064. Bearden delineated the factors to be balanced in examining an equal protection claim involving a lack of financial resources and the working of the criminal justice system: it requires a careful inquiry into (1) the nature of the individual interest affected; (2) the extent to which it is affected; (3) the rationality of the connection between legislative means and purpose; and, (4) the existence of alternative means for effectuating the purpose. Id. at 666, 103 S.Ct. 2064. Thus, Bearden both articulated guiding principles for analyzing equal protection claims brought by indigent defendants and applied them to the specific issue presented, which, like the issue presented here, concerned increased incarceration because of indigency. Bearden stresses the necessity of weighing alternatives to increased incarceration as a means of punishing the indigent. Id. at 671-72, 103 S.Ct. 2064. In no uncertain terms, the Court held that to deprive a defendant of his freedom simply because, through no fault of his own, he cannot pay would be contrary to the Fourteenth Amendment’s requirement of fundamental fairness. Id. at 672, 103 S.Ct. 2064.
Ninth Circuit case law also confirms that Bearden is clearly established law on the issue of incarceration based on indigency. For example, in United States v. Parks, 89 F.3d 570, 572-73 (9th Cir.1996), we applied the principles of Bearden to find a constitutional violation when a defendant was assessed two extra criminal history points because, at the time he committed his offense, he was under a sentence of “legal financial obligation” to pay state law fines, and the increase in his criminal history due to this outstanding financial obligation increased his Guidelines sentencing range for the new offense from 37-46 months to 46-57 months. We found that it was unconstitutional to add eight months of incarceration to a defendant’s sentence solely because of indigence, stating that “[t]he application of Washington state’s [legal financial obligation] as a criminal justice sentence ... creates two classes of defendants for federal sentencing purposes: those who could afford to pay their state law fines immediately, and those who required a period of time to do so.” 89 F.3d at 572 & n. 5.
Bearden controls the analysis of whether and under what circumstances an individual can be subjected to greater incarceration solely because of indigency. Respondents, however, argue that no specific Supreme Court case compelled the Washington Supreme Court to apply Bear-den to MacFarlane’s and Fogle’s petitions: “Petitioners fail to identify any United States Supreme Court case which applies the Bearden test to a disparate award of good conduct time between pre-trial detainees in county jails and convicted felons in a state prison system.” Bearden, as respondents note, does not address the identical factual situation presented here. If that were the standard, however, hardly any law would be “clearly established,” for only cases presenting precisely the same facts could govern later cases. The Supreme Court has not defined “clearly established” in the AEDPA context, nor has this court directly opined on what constitutes “clearly established law.” We have, however, in several cases found “clearly established law”, and, in doing so, have not required that the precedent arise from the identical factual circumstances. For example, in Jeffries v. Wood, 114 F.3d 1484, 1500-01 (9th Cir.1997) (en banc), we held that Chapman v. California, 368 U.S. 18 (1967), was clearly established law regarding the burden of proof on prejudice in habeas cases. In Davis v. Kramer, 167 F.3d 494 (9th Cir.1999), we held that Anders v. California, 386 U.S. 738, 87 S.Ct. 1396, 18 L.Ed.2d 493 (1967), was clearly established law regarding the constitutionally mandated procedure that appellate counsel must follow if he determines that an appeal would be frivolous. See also [1140]*1140Moore v. Calderon, 108 F.3d 261, 265 (9th Cir.1997) (Faretta v. California, 422 U.S. 806, 95 S.Ct. 2525, 45 L.Ed.2d 562 (1975), “clearly established” a rule of law regarding the timeliness of a request for self-representation). These cases comport with the definition of “clearly established” in the Teague context, relied on by other circuits to determine whether there is clearly established law in the AEDPA context.10 We agree in this respect with the First Circuit, which, drawing on the Teag-ue line of cases, stated that a petitioner need not point a habeas court to a factually identical precedent, but rather that “clearly established” Supreme Court law often erects a framework intended for application to variant factual situations. See O’Brien v. Dubois, 145 F.3d 16, 24 (1st Cir.1998). See also Neelley v. Nagle, 138 F.3d 917, 922-23 (11th Cir.1998) (“clearly established” terminology echoes the concerns underlying Teague and its progeny).
Respondents fail to show that there is any meaningful distinction between the case of a poor person unable to pay a fíne, who, as a result, suffers longer incarceration than a similarly situated affluent person (Bearden) and the case of a poor person unable to post bail pre-trial who, as a result, suffers longer incarceration than a similarly situated affluent person (MacFarlane and Fogle). Bearden erected the framework intended for application where a defendant suffers increased incarceration because of indigency and, thus, is clearly established law governing the issue presented here.11
[1141]*1141IV.
We turn now to the question whether the Washington Supreme Court’s denial of MacFarlane’s and Fogle’s habeas petitions was contrary to, or an unreasonable application of, Bearden. The Washington Supreme Court conducted a full review of MacFarlane’s and Fogle’s claims, but did not mention Bearden, nor any other federal case in its opinion. The state court applied an intermediate level of scrutiny; it was on this basis that the district court concluded that the state court’s decision was consistent with Bearden. But heightened scrutiny is not all that Bearden stands for: the state court failed to identify or apply the factors that must be weighed under Bearden in determining whether equal protection has been violated by a wealth-based classification, namely, “the nature of the individual interest affected, the extent to which it is affected, the rationality of the connection between legislative means and purpose, [and] the existence of alternative means for effectuating the purpose.” 461 U.S. at 666-67, 103 S.Ct. 2064. Indeed, the state court failed to consider petitioners’ interest at all, or to examine the connection between the legislative means and objectives. In addition, the state court did not consider any alternative means for accomplishing the penological purposes that it identified. When the state court fails to apply a controlling Supreme Court authority, we must examine its decision in light of that authority. Here, we must apply the Bearden factors to determine whether petitioners’ right to equal protection was violated by the counties’ early-release policies, which limited pre-sentence detainees to a maximum good-conduct credit of 15% of the sentence imposed and prohibited them from participating in good-performance programs. We will consider the good-conduct credit first, because good performance is less directly involved here and involves significantly different issues for those in charge "of the jails.
First, the nature of the individual interest affected here—physical liberty— weighs heavily in the constitutional calculus; the state’s interests must be highly significant if they are to outweigh this fundamental interest. The Supreme Court has repeatedly found wealth discrimination unconstitutional when additional incarceration is caused solely by disparities in wealth. See e.g., Williams v. Illinois, 399 U.S. 235, 90 S.Ct. 2018, 26 L.Ed.2d 586 (1970) (unconstitutional for state to add time to maximum prison term solely because the defendant involuntarily failed to pay fine or costs); Tate v. Short, 401 U.S. 395, 398, 91 S.Ct. 668, 28 L.Ed.2d 130 (1971) (jailing the defendant for involuntary failure to pay fines on non-jail offenses unconstitutionally imprisons “solely because of ... -indigency”). The crucial factor here is that the additional incarceration is caused solely by the petitioners’ indigency: had they posted bail and served their entire sentences in a state facility, they would have been required to serve less time.
Second, as to the extent to which petitioners’ interest (their physical liberty) was affected, although from our perspective 30, 40, or 50 days more or less on a 5-year or 6-year term of imprisonment may seem negligible, even a single extra day of incarceration is of substantial significance for constitutional purposes. Had Fogle posted bail, served no time in the Pierce County Jail, and been committed directly to the state facility after sentencing, his incarceration would be shorter by 16 days. Had MacFarlane posted bail, served no time in the Clark County Jail, and been committed directly to the state facility after sentencing, his incarceration would be reduced by approximately 23 days.
Third, the state court’s failure to examine the rationality of the connection between the claimed.legislative purposes and the good-conduct policy contributed significantly to its error. According to the state, two primary purposes motivate the sentencing statute in general, and the county jails’ good-conduct (and good-performance) policies in particular: discipline [1142]*1142and local control. The first purpose, maintaining prisoner discipline, has no rational connection to affording county pretrial detainees less good-conduct credit; in fact, one might reasonably assume the contrary, that increasing the amount of good-conduct credit available would increase inmates’ incentive to behave well and, thus, reduce disciplinary problems. The second purpose, local control, is only weakly connected to the counties’ calculation of good time credit for detainees who serve most of their sentences in state facilities. The county jails have little rational interest in how much credit for good conduct is ultimately to be received by their former inmates who have moved to the state system to serve the principal portions of their sentences. The county jails can, and do, exert individual control over what constitutes good conduct, and whether a particular inmate has met that standard, but it is illogical to contend that the ultimate calculation of the maximum amount of good conduct time to which a prisoner who has earned good conduct credit and who is no longer in county jail is entitled could endanger' — or, alternatively, further — local control. Moreover, to the minor degree, if any, an interest in local control is implicated, that interest is substantially outweighed by the prisoner’s interests in liberty and freedom from punishment on account of indigency.
Finally, there are alternative means to accomplish the legislative purposes that would not involve additional incarceration time for those who are unable to make bail and whose conduct is satisfactory during their stays in county jail. As we have noted, to assuage any concerns regarding discipline, the counties can (and do) deny good-conduct time entirely to detainees who present disciplinary problems, and counties can revoke or restrict the privileges of such individuals. As to local control, a minimal interest, at most, underlies the good-conduct policy. In addition, as the dissenting justices of the Washington Supreme Court pointed out, when a prisoner is transferred post-sentencing from a county jail to a state facility, the county could forward to the Department of Corrections a statement containing the total number of days the prisoner spent in the jail and the number of days on which he did not earn good-conduct credit. The Department of Corrections could then itself calculate the total amount of credit for good conduct that the prisoner should receive under the applicable state policies and credit that amount of time against the time to be served. In re Fogle, 128 Wash.2d at 73, 904 P.2d 722.
V.
Our application of the legal rules prescribed by Bearden clearly requires the conclusion that the good conduct policies challenged by MacFarlane and Fogle violate equal protection. The Washington Supreme Court’s failure to apply Bearden, and its resultant decision that the county jails’ good-conduct policies did not violate equal protection, was “contrary to ... clearly established ... Federal law” within the meaning of 28 U.S.C. § 2254(d). Such a failure to identify or apply the governing Supreme Court standard unquestionably falls within the “contrary to” provision of AEDPA. See Jeffries, 114 F.3d at 1500-01 (where the Washington Supreme Court had failed even to cite Chapman or to identify the correct harmless error standard set forth in that case, we held that the “contrary to” law provision applied). This court has not, however, attempted to define the outer boundaries of the “contrary to” provision, nor need we do so here. We also have not sought to draw a rigid distinction between the “contrary to” and “unreasonable application of’ provisions, both of which reflect the same general requirement that federal courts not disturb state court determinations unless the state court has failed to follow the law as explicated by the Supreme Court. Indeed, we have stated that there is an overlap in the scope of the two provisions. See Davis, 167 F.3d 494. Again, however, because the “contrary to” provision is clearly met, we need not consider whether such an overlap exists here.
[1143]*1143VI.
MacFarlane and Fogle also challenge the counties’ good-performance policies, which raise the same constitutional question as the good-conduct policies, but require a different answer.
Under the policy in force in both Pierce County and Clark County, pre-sen-tence detainees are ineligible for participation in work or other programs through which they could earn good-performance credit. This ineligibility prevents those detainees from earning the additional early-release credit that can be earned by persons who serve their entire sentences in a state facility. The nature of the individual interest affected by the good-performance policy is, of course, the same as that affected by the good-conduct policy: physical liberty and freedom from incarceration. The crucial difference, for purposes of our Bearden analysis, lies in another factor: the relationship between the policy’s objectives and the means that it employs to achieve them. The counties assert that concerns over community safety and the risk of flight primarily underlie their refusal to permit pre-sentence detainees to participate in good-performance programs, which frequently entail allowing inmates to work outside the jails. They point out significant differences in the composition of the pre-sentence detainee population, which may include murderers, rapists, and career offenders, and the larger population of convicted inmates who have been sentenced to serve a comparatively small amount of time (one year or less) in the county jail. The counties have established a strong rational connection between the legislative means and purpose. Similarly, an individual county’s ability to decide on eligibility to participate in a good-performance program, and in what circumstances, during periods of detention in that county’s jail, is closely tied to local control, given the myriad variables involved in allowing a prisoner to work or attend an academic program and the commitment of local resources that may be required to administer the programs. Furthermore, unlike in the case of the good conduct policies, no alternative means are readily identifiable for accomplishing the counties’ objectives—ensuring community safety and preventing flight— while permitting inmates to participate in work or other “performance” programs. We conclude, therefore, that as to the good-performance policies the counties’ interests in community safety and prevention of flight outweigh petitioners’ interest in physical liberty. Thus, the good-performance policies, unlike the good-conduct credit policies, do not violate the Equal Protection Clause.
VII.
The district court’s decision is reversed as to the counties’ good-conduct policies and affirmed as to their good-performance policies. The Department of Corrections must calculate the amount of credit for good conduct that the petitioners would have received had they been subject to the state good-conduct policy for the period of time they spent in the county jails, and must credit that amount of time against the time served.
REVERSED IN PART, AFFIRMED IN PART.