concurring:
The court affirms the district court’s award of more than $1.4 million in attorneys’ fees. This is in addition to an earlier *1093award of $1.8 million in this class action, which has now lasted fourteen years. These serial fee disputes are the latest chapter in a saga the court describes as “stretching back at least forty years.” Maj. Op. 1089. Although the legal result is defensible, the practical effect is reprehensible. The only segment of the population well-served by this aspect of the IDEA program is the lawyers who litigate these cases.
The District argues that the $4,000 fee cap provides a ceiling even for a class action; the plaintiffs insist the fee cap does not apply to class actions at all. I agree with the plaintiffs. But this court apparently concludes that while the cap does not limit fees in class actions to $4,000, it establishes an upper bound. See Maj. Op. 1090, 1092; Williams, J., concurring, at 1099-1100. Thus, counsel in a class action would be entitled to reasonable attorneys’ fees so long as the fees do not exceed “$4,000 per student in the class.” Williams, J., concurring, at 1099. Yet the fee cap legislation — reenacted several times — says nothing about class actions. In the absence of any express prohibition, a court could decide the fee cap is irrelevant to the question of class action fees and resort to a reasonable lodestar calculation. See, e.g., Clackamas Gastroenterology Assocs., P.C v. Wells, 538 U.S. 440, 447, 123 S.Ct. 1673, 155 L.Ed.2d 615 (2003) (“[Cjongressional silence often reflects an expectation that courts will look to the common law to fill gaps in statutory text....”). Because under either approach the bottom line would be indistinguishable, I concur in the result. But I write separately because I wonder if class actions — which fall outside the fee cap paradigm — should be certified for IDEA claims and because the focus on fee caps seems both futile and pernicious.
I
■ I begin with the issue of statutory interpretation. We could — and probably should — derive a different principle from congressional silence on the subject of class actions. Since Congress clearly did not contemplate such representative actions and made no provision for compensating broad systemic attacks, the statute’s silence challenges the propriety of class actions under the auspices of IDEA. Indeed, the very structure of the statutory scheme undermines the validity of class actions in this context.
When IDEA was enacted in 1975, its goals were lofty but attainable; IDEA aimed to provide a “free appropriate public education” for children with special educational needs. Education of All Handicapped Children Act, Pub.L. No. 94-142, § 612, 89 Stat. 773 (1975) (codified as amended as the Individuals with Disabilities Education Act, 20 U.S.C. § 1400 et seq.). (1982). To achieve this goal, IDEA mandated that participating states identify these children and then work cooperatively with parents, teachers, and administrators to develop individualized education plans suitable to each child’s needs and abilities. Pub.L. No. 94-142, § 613-14.
IDEA’S remedial system was not intended to be adversarial; although Congress anticipated that disagreements would arise that required legal intervention, Congress’s aim was to resolve disputes amicably — litigation seemingly was intended as a last resort. See Pub.L. No. 94-142, § 615 (listing a due process hearing last on the list of procedural safeguards); see also 20 U.S.C. § 1415(e)-(f) (1997). And Congress certainly did not anticipate that attorneys’ fees would become the focus of any litigation that ensued. See H.R.Rep. No. 105-670, at 21 (Committee on Appropriations) (deciding in 1999 to cap attorneys’ fees so that “the District’s compli*1094anee with [IDEA] will focus more clearly on teaching and learning rather than on litigation and expensive legal fees”). In fact, Congress affirmed this understanding of IDEA’S original purpose as recently as 2005, when it amended the statute to allow a state or local educational agency to recoup attorneys’ fees from parents who file frivolous or unfounded complaints. 20 U.S.C. § 1415(h)(3)(B)(i)(II)-(III). By amending the statute, Congress implicitly recognized that IDEA was being distorted; it was no longer simply a means to improve primary education, but was becoming a breeding ground for vexatious and costly litigation. See id. (permitting state and local agencies to recoup fees when the parent’s cause of action was intended “to harass, to cause unnecessary delay, or to needlessly increase the cost of litigation”).
IDEA’S singular focus on an individualized, cooperative educational approach providing customized remedies makes Congress’s neglect of broad-based actions understandable and renders IDEA actions ill-suited to class-wide relief. The Federal Rules of Civil Procedure demonstrate perfectly why this is so. Federal Rule of Civil Procedure 23(a) sets forth four prerequisites for the certification of a class. If those prerequisites are met, a class will only be certified if the plaintiffs can establish their case fits within one of the three categories of subsection (b).
Paragraph (b)(2) of Rule 23, the basis of class certification in this case, see Oct. 22, 1997 Mem. & Order Certifying the Class at 2, permits a class action if “the party opposing the class has acted or refused to act on grounds that apply generally to the class, so that final injunctive relief or corresponding declaratory relief is appropriate respecting the class as a whole.” To satisfy this test, class claims must be “sufficiently cohesive to warrant adjudication by representation.” Amchem Products, Inc. v. Windsor, 521 U.S. 591, 623, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997). This requirement stems from the understanding that even unnamed class members are bound by a Rule 23(b)(2) class action without the opportunity to opt out. See In re Veneman, 309 F.3d 789, 792 (D.C.Cir.2002). Accordingly, cohesiveness is a significant touchstone of a(b)(2) class. See Barnes v. Am. Tobacco Co., 161 F.3d 127, 142-43 & n. 18 (3d Cir.1998).
As is probably true for most IDEA-spawned class actions, this class action raises a significant question about whether the class claims are sufficiently cohesive to merit certification. Certainly, at some level of abstraction, a degree of cohesion will exist in almost any putative class. But this does not mean it is prudent to generalize to such a degree, especially when IDEA operates from the premise that each child will have unique disabilities and presumes that each program will be personalized. See Pub.L. No. 94-142, § 613-14; cf. J.B. ex rel. Hart v. Valdez, 186 F.3d 1280, 1288-90 (10th Cir.1999) (affirming denial of class certification when alleged “systemic failures” too broadly defined the claims at issue, amounting to a lack of commonality under 23(a)(2)).
Moreover, it is patently evident the District is frequently unable to comply with IDEA. See, e.g., Letter from Patricia Guard, Acting Director of Special Education Programs to Superintendent Gist, at 1 (June 15, 2007), http://www2.ed.gov/fund/ data/report/idea/partbspap/2007/de-aprltr2007b.doc (last visited Jan. 12, 2011) (“[T]he District of Columbia needs intervention in meeting the requirements of Part B of the IDEA”). The question then is not one of fault but one of remedy — and any remedy will necessarily be unique to a particular child. This means the District’s systemic failures have little impact on the ultimate question courts now face; even assuming the District violated its IDEA *1095obligations as to each plaintiff, any relief will be “dependent upon that particular studentfs] needs, capabilities, and the IEP in place for that child.” Blunt v. Lower Merion Sch. Dist., 262 F.R.D. 481, 489-90 (E.D.Pa.2009). In fact, even the class certification and ultimate liability determination in this case did little to achieve the personalized remedies plaintiffs sought; many class plaintiffs proceeded to file individual motions for preliminary injunctions “based on them unique circumstances.” Appellants’ Br. at 10. That this class action failed to provide the specialized relief plaintiffs desired only further supports the notion that the class action is an inappropriate vehicle for these claims.
Indeed, we have recognized the individual nature of IDEA’S remedies, and consequently have declined to fashion “cookie-cutter” compensatory relief. Reid ex rel. Reid v. District of Columbia, 401 F.3d 516, 524 (D.C.Cir.2005) (“[J]ust as IEPs focus on disabled students’ individual needs, so must awards compensating past violations rely on individualized assessments”). I see no reason why these considerations should not be decisive when it comes to class certification, and decisions in several other jurisdictions support this view. See Blunt, 262 F.R.D. at 489-90 (refusing to certify a class in part because “individual determinations, which must be made to determine whether a particular student falls within the class definition and whether such student has a cause of action, weigh against certifying this class”); M.A. ex rel. E.S. v. Newark Public Schs., No. 01-3389, 2009 WL 4799291, at *15 (D.N.J. Dec. 7, 2009) (unpublished) (declining to certify a(b)(2) class for lack of cohesion, which resulted from the necessity of individualized circumstances and remedies); cf. McClendon v. Sch. Dist. of Phila., No. 04-1250, 2005 WL 549532, at *4-5 (E.D.Pa. Mar. 7, 2005) (unpublished) (denying class certification because named class members failed to satisfy Rule 23(a)(4)’s typicality requirement in that the remedy sought was individualized in nature). Because the ultimate relief in an IDEA action is unique to each plaintiff, any putative IDEA class quite arguably suffers from a lack of cohesion. Although it is too late for this court to consider the issue, hereafter courts should give significant consideration to the question of whether a class action is the appropriate vehicle for an IDEA case — at least for one seeking individual remedies.
II
The problem of IDEA litigation is not the only problem that needs a remedy. IDEA’S focus from the outset was the child — recognizing that individual children required individualized educational strategies to ensure their learning success, IDEA outlined a cooperative statutory scheme. See Pub.L. No. 94-142, § 614 (“A local educational agency ... shall ... establish a goal of providing full educational opportunities to all handicapped children, including ... the participation and consultation of the parents or guardian of such children.... ”). But the individualized, cooperative educational system IDEA envisioned was quickly overwhelmed in the District of Columbia.
Here the IDEA system is anything but cooperative. Because the District is perpetually unable to comply with IDEA’S mandates, parents have sought to enforce their rights in court — a situation IDEA contemplates, but one surely not intended to be the norm. See 20 U.S.C. § 1415. It has become so routine to “lawyer-up” and go to court, IDEA has morphed from a system intended to benefit children to a system that provides full employment to a specialized cadre of lawyers. And this case is a perfect (yet terrible) illustration of the transformation IDEA has undergone within the sixty-eight square miles of the nation’s capital. Rather than fulfilling *1096IDEA’S promise to educate the individual child, this lawsuit involves more than six-thousand plaintiffs. It has lasted more than a dozen years. Plaintiffs who were in grade school when this ease began long ago graduated (or dropped out). Some may have children of their own who are being victimized by the District’s dysfunction. If we have learned anything from decades of IDEA litigation, it is this: when the system is broken, litigation will not fix it.
Congress itself has recognized the problems plaguing IDEA litigation in the District, and has attempted to re-order the program’s incentives over the last twelve years by enacting statutory caps on the fees attorneys may be awarded for their work representing IDEA plaintiffs. But the fee cap has been limited; each renewal faced strong opposition. Some years the cap was tabled entirely, and as it currently stands, the fee cap only applies retroactively to cases initiated prior to 2009. Omnibus Appropriations Act, 2009, Pub.L. No. 111-8, 123 Stat. 524, 697-98 (2009). As this case and others like it demonstrate, Congress’s efforts seem to have been misguided. Congress has focused on attorneys’ fees, but fees are only a visible symptom of a more fundamental failure. The District is frequently unable to provide an adequate education for special needs children because the District struggles to provide an adequate education for any children.
Here, the per-pupil costs of education are astronomical, yet the District suffers from an inversely proportionate success rate. For the 2006-2007 school year, the Institute of Education Sciences (“IES”) calculated educational spending in the District at $20,596 per pupil, National Center for Education Statistics, http://nces.ed.gov/ programs/digest/d09/tables/dt09_186.asp (last visited Jan. 12, 2011) (hereinafter “Table 186”); some estimates are even higher, Andrew Coulson, Do You Still Think DC Spends only $15,000/Pupil?, (Feb. 19, 2010) http://www.cato-at-liberty. org/do-you-still-think-dc-spends-only-15000pupil/ (last visited Jan. 12, 2011) (follow “Excel spreadsheet file” link to 2008-2009 data sheet) (calculating the annual cost per pupil to be $28,169.91 in 2009). Even using the lower figure for consistency across comparisons, this means the District outspends some of the most heavily populated states, including California ($9,364), Florida ($9,391), Pennsylvania ($12,440), New York ($17,818), and Texas ($8,798). Table 186. In fact, according to the IES compilation, the District outpaces every single state in per-pupil spending. Id. But the District’s graduation rate is well-below the national average, which was 74.7% in 2004-2005 (the last year for which data was available). National Center for Education Statistics, http://nces.ed.gov/ pubs2009/dropout07/tables/table_13.asp (last visited Jan. 12, 2011). For that same period, the District’s graduation rate was 68.8% — the twelfth lowest rate in the country. Id. And IDEA litigation costs the District in excess of $10 million each year in attorneys’ fees. D.C. Office of the Att’y Gen. http://newsroom.dc.gov/show.aspx/ agency/oec/section/2/release/15655/year/ 2008 (last visited Jan. 12, 2011).
All that money has failed to stem the tide of IDEA litigation. This tiny district with fewer than 50,000 pupils has more IDEA suits than the largest school systems in the nation. See Deborah K. Nichols, Auditor, Office of the District of Columbia Auditor, Flawed Processes and Ineffective Systems of Accountability Pertaining to DCPS’ Special Education Program Have Resulted in Costly Legal Fees and Exorbitant Charges for Related Services and Nonpublic Tuition, at 24 (May 9, 2003) http://dcauditor.org/dca/reports/ dca0303.pdf (last visited Jan. 12, 2011) (explaining that in 2002, the District of *1097Columbia Public Schools received more due process hearing requests (3,044) than the entire state of California (2,670), even though the District’s entire student population numbers less than 50,000, while in California the special education population alone numbers 670,000 students).
The solution to the District’s failure as an educational provider is not mysterious. It does not depend on how well or how poorly lawyers get paid for representing dissatisfied parents. Even in the District, there are bright spots. For example, the Seed School, which is a charter school serving the District of Columbia, has experienced a 96% rate of college acceptance and a 95% rate of college enrollment among its students. The Seed School of Washington, http://www.seedschooldc.org/ (last visited Jan. 12, 2011). And the Seed School achieves these successes educating the very same children the District is unable to assist; in fact, the Seed School accepts applications on a lottery basis, ensuring that selection is not based on demonstrated academic ability. The Seed School of Washington, Admissions, http:// www.seedschooldc.org/page.php7pid=63 (last visited Jan. 12, 2011). Similarly, the Friendship Public Charter School, which is publicly funded in its entirety, has produced significant results; students have exceeded state academic requirements for six years running. Friendship Public Charter School, About: Welcome from the Chairman, http://www.friendshipsehools. org/RelId/606513/ISvars/default/Weleome_ fromMhe_Chairman.htm (last visited Jan. 12, 2011). Programs like those offered by the Seed School and the Friendship Public Charter School demonstrate that these children can achieve academically, if given the proper guidance. But such guidance need not necessarily take the form of paid legal representation. The sixty-eight square miles of the federal city probably has more JDs and PhDs per capita than any other spot in America, yet it lags behind states like Pennsylvania and Ohio where parents are often guided through the IDEA process by a corps of parent volunteers. See, e.g., Dr. Edward Feinberg, The Role of Attorneys in Special Education Mediation, http://www.mediate. com/articles/cadre4.cfm (last visited Jan. 12, 2011) (discussing Pennsylvania’s system of non-attorney “advocates” and Ohio’s successful parent mentor program).
The District of Columbia Public Schools’ biggest problem should be having more pro bono services offered than it can use. Instead, in the bizarro world inhabited by the District, the more its student population shrinks, the bigger its legal bills grow. If there is an answer to this problem, it will likely come from concerned parents, committed teachers, and conscientious volunteers. One thing is clear. It will not come from adjusting the spigot directing the flow of public funds to lawyers.