The modern prosecutor is formally treated as the constitutional custodian of fair disclosure, but functionally operates inside an information environment that is fragmented across agencies, databases, personnel systems, investigative teams, laboratories, supervisory chains, and informal channels of institutional memory. That structural contradiction lies at the center of the Brady economy. The law speaks as though the prosecution can know what the government knows. The bureaucracy is organized so that no single actor naturally does. The result is not merely administrative inconvenience. It is a recurring constitutional defect in which favorable evidence, impeachment material, credibility concerns, and context essential to adversarial fairness are dispersed across a system that rewards compartmentalization more than integration. Brady Doctrine assumes a prosecutorial obligation broad enough to protect due process, while ordinary justice-bureaucracy design distributes information in ways that make that obligation difficult, selective, and often strategically incomplete. Brady established the disclosure duty for material favorable evidence, and Kyles later made explicit that the individual prosecutor has a duty to learn of favorable evidence known to others acting on the government’s behalf in the case. The Department of Justice’s own policy reflects that same premise by requiring federal prosecutors to seek exculpatory and impeachment information from all members of the prosecution team, including participating federal, state, and local officers.
This chapter matters because the problem of prosecutorial knowledge is not, in the first instance, a problem of intellect or diligence. It is a problem of institutional architecture. The prosecutor may be intelligent, conscientious, and formally aware of Brady while still standing at the center of a system designed to withhold integrated knowledge from any one participant. Police agencies retain internal affairs records. Laboratories hold validation concerns, bench notes, and quality issues. Jails and custodial units may record witness benefits, contacts, or coercive conditions. Supervisors know about patterns of officer unreliability that never migrate into trial files. Human-resources or disciplinary systems classify conduct as personnel information rather than impeachment information. Informal knowledge about a witness’s bias, inducements, or prior falsehoods may circulate conversationally without ever entering a database that a line prosecutor would search. When those informational fragments remain institutionally separated, the prosecutor’s formal duty expands while the prosecutor’s practical visibility contracts. The legal system then treats nondisclosure as a failure of prosecutorial obligation even though the underlying cause is often a bureaucratic structure that makes full knowledge improbable by design.
Kyles is decisive here because it rejects the idea that the State may benefit from its own internal divisions. The Court made clear that the prosecutor’s duty reaches favorable evidence known to others acting on the government’s behalf in the case, including the police. That holding is often read as an expansion of duty, but its deeper significance is structural. It recognizes that the State cannot constitutionally compartmentalize information and then invoke that compartmentalization as a defense to nondisclosure. The point is not simply that prosecutors should try harder. The point is that the government must be treated as a constitutional whole when it prosecutes, because fragmented state knowledge otherwise becomes an engine of concealment. The Department of Justice mirrors this logic in policy by defining the prosecution team broadly and requiring prosecutors to seek exculpatory and impeachment information from those participating in the investigation and prosecution.
Yet the law’s recognition of collective governmental knowledge does not eliminate the administrative reality that information is stored, classified, and transmitted through institutions that are not built primarily for disclosure. Rule 16 of the Federal Rules of Criminal Procedure provides a discovery framework, but it is not coextensive with Brady, and Brady itself was never transformed into a comprehensive system of open-file constitutional administration. The Constitution requires certain disclosure; bureaucracies continue to manage records according to operational, personnel, security, labor, and litigation logics that only partially overlap with that constitutional demand. The result is a constant slippage between legal obligation and record architecture. A prosecutor can comply with routine file review while still missing information never placed in the litigation file. An office can adopt discovery policy while still relying on participating agencies whose internal systems were designed for discipline, employment, or supervision rather than for constitutional disclosure. That is why information fragmentation is not cured merely by reciting Brady or by training prosecutors to ask for favorable evidence in the abstract. The problem is embedded in the way justice institutions create and store knowledge.
The fragmentation problem is especially acute with impeachment material. Giglio and Bagley make clear that impeachment information falls within the disclosure obligation, but impeachment information is often the category most vulnerable to bureaucratic dispersal. Promises to witnesses may be held in prosecutor communications, case notes, or law-enforcement briefing files. Prior misconduct by officers may be held in disciplinary systems insulated by confidentiality rules or labor practices. Informant benefits may exist in payment records, custodial records, debrief memoranda, or separate investigative files. Bias, coercion, and inconsistent accounts may be reflected in rough notes or interagency communications never integrated into the formal prosecution file. Because impeachment information often concerns credibility rather than the core narrative of offense conduct, institutions are particularly tempted to treat it as peripheral, disputed, or administratively separate. But constitutionally it is not peripheral. The credibility of government witnesses is frequently central to guilt, punishment, plea leverage, suppression issues, and the defense’s ability to investigate. The more the system relies on witness credibility, the more dangerous it becomes to store credibility information outside the channels through which prosecutors actually learn and disclose.
This leads to one of the defining contradictions of the Brady economy: the prosecutor is made responsible for information that the institution does not reliably make visible. That contradiction generates two different forms of failure. The first is genuine informational failure. The prosecutor does not know because the structure did not reveal what the prosecutor was duty-bound to know. The second is managed ignorance. The office knows enough about the fragmentation problem to understand that certain kinds of information are likely to exist elsewhere, yet it preserves procedures narrow enough to avoid actual institutional knowledge becoming too definite, too searchable, or too disclosable. Managed ignorance is more than negligence. It is a bureaucratic equilibrium in which the system avoids building the very structures that would convert scattered warnings into undeniable disclosure obligations. The institution benefits because ambiguity preserves operational flexibility. So long as information remains distributed and unintegrated, officials can describe nondisclosure as oversight, incompleteness, or uncertainty rather than as conscious suppression.
That is why prosecutorial knowledge must be analyzed as an institutional phenomenon rather than a purely personal one. The classic image of the prosecutor as a single actor reviewing a file and making disclosure decisions obscures the actual conditions of modern prosecution. Prosecutors operate through task forces, multi-agency investigations, laboratory submissions, digital evidence systems, parallel administrative channels, and intergovernmental cooperation. The state’s knowledge is therefore networked. But its disclosure duties are still often operationalized through office-centered review practices that assume the relevant information has already flowed inward. When that assumption fails, the prosecutor becomes the nominal bearer of a duty without possessing the infrastructural means to discharge it fully. The constitutional order thus depends on something the bureaucracy does not naturally generate: integrated state knowledge at the point of disclosure.
The justice system often responds to this problem by emphasizing the phrase “prosecution team.” That phrase is useful, but it can also conceal more than it solves. To describe police, prosecutors, and participating officials as one team for Brady purposes is doctrinally necessary; otherwise the government could endlessly evade responsibility through internal division. But the label itself does not create real informational unity. Agencies may have incompatible systems, different retention practices, conflicting incentives, and divergent understandings of what counts as disclosure material. Some participants regard impeachment information as confidential employment data. Others treat it as tactical material. Others do not reduce important credibility concerns to writing at all. A formal team can therefore remain an informational archipelago. The constitutional fiction of unity is indispensable to prevent evasion, yet the administrative reality often remains fractured enough to generate recurring nondisclosure.
The problem becomes even more severe in plea-driven systems. United States v. Ruiz held that the Constitution does not require pre-plea disclosure of impeachment information in the same way it does at trial. Whatever the reasoning of that decision on its own terms, its structural effect is significant. In a system dominated by pleas, delayed visibility of impeachment material lowers the practical cost of fragmented knowledge. If the institution can obtain disposition before credibility information is surfaced, the incentive to build aggressive, integrated pre-plea disclosure systems weakens. A fragmented bureaucracy already struggles to move impeachment information to the prosecutor; a plea-centered case flow reduces the immediate institutional penalty for that failure. The result is a system in which informational disorder and expedited resolution can reinforce each other. A defendant may resolve a case without ever receiving the full benefit of material that the government, in a more integrated structure, should have located and understood.
The ethical and administrative posture of federal policy underscores that this problem is widely recognized. The Justice Manual states that the prosecution’s duty to disclose is governed not only by Brady and Giglio but also by internal Department policy, and it directs prosecutors to seek information from the prosecution team rather than passively await production. That is an important admission. It acknowledges that prosecutors cannot simply review their immediate files and declare compliance. They must affirmatively search upstream and outward. But the necessity of that policy also reveals the depth of the structural problem. When a system must expressly command prosecutors to go looking across institutional boundaries, it is admitting that ordinary workflow will not reliably deliver constitutional knowledge on its own.
Information fragmentation also distorts responsibility. When favorable evidence surfaces late, each actor can describe the problem as residing elsewhere. Police may say they were never asked in precise enough terms. Prosecutors may say the agency never transmitted the material. Supervisors may say the information was not classified as Brady or Giglio material. Records custodians may say they handle employment or administrative files, not criminal disclosure. Laboratories may say the prosecutor did not request the relevant underlying data. The institution thereby converts a single constitutional failure into a chain of partial disclaimers. No one actor appears to have suppressed the truth in a singular, dramatic way, yet the defendant was still denied the benefit of information the State, taken as a whole, possessed or could readily have assembled. Fragmentation thus functions not only as an evidentiary problem but as an accountability solvent. It dissolves concentrated blame.
This solvent has major consequences for later litigation. A plaintiff challenging suppression or disclosure failure must often prove not just that favorable information existed, but that it was attributable to the prosecution team, materially significant, and suppressed in a way that affected the fairness of the proceeding. When records are fragmented, proving that chain becomes more difficult. The institution’s own disorder becomes a litigation advantage. Missing integration makes causal narratives harder to reconstruct. Ambiguous custodianship permits arguments about what the prosecutor did or did not have. Classification disputes enable claims that certain files were outside the disclosure universe. Even where Kyles rejects narrow compartmentalization in principle, practical fragmentation still raises the cost of establishing exactly how institutional knowledge should have been assembled. The result is a self-protective system: fragmented information creates nondisclosure, and the same fragmentation later makes the nondisclosure harder to litigate.
The deeper issue, then, is not simply whether prosecutors know enough. It is how institutions decide what counts as knowable. A bureaucracy can choose to build integrated credibility systems, searchable disclosure repositories, mandatory interagency reporting rules, auditable transmission logs, and supervisory review structures that force information to converge. Or it can preserve patchwork systems in which disclosure depends on memory, informal relationships, discretionary reporting, and incomplete file assembly. The first model treats prosecutorial knowledge as a design problem. The second treats it as a personal burden laid atop a fragmented organization. The Brady economy has too often preferred the second model because it is cheaper in the short term and more protective of institutional flexibility. It avoids the full administrative cost of constitutional integration while preserving the ability to speak the language of compliance.
That short-term economy is deceptive. The costs reappear later as reversals, post-conviction litigation, civil-rights exposure, disciplinary controversy, public distrust, and the broader corrosion of adjudicative legitimacy. Connick v. Thompson illustrates the doctrinal difficulty of converting disclosure failure into municipal liability through failure-to-train theory, but the decision also reveals the structural stakes. When courts make municipal liability difficult to establish absent a pattern or other strong proof, institutions may underinvest in the infrastructure needed to transform fragmented state knowledge into reliable prosecutorial knowledge. The danger is not merely doctrinal under-compensation. It is administrative under-construction. If the legal system does not strongly price fragmented disclosure architecture, governments may continue to operate without building the systems their constitutional duties presuppose.
The solution must therefore be structural. Prosecutorial knowledge cannot depend on heroic individual diligence inside institutions organized for partial visibility. The constitutional duty announced in Brady and sharpened in Kyles requires administrative consequences. Participating agencies must be obligated to transmit favorable and impeachment information through durable, searchable systems rather than informal communication. Credibility-related records must be classified with disclosure consequences in mind, not merely employment or disciplinary categories. Supervisory review must focus on whether information actually moved, not merely whether policies exist. Audit trails must be strong enough to reveal where fragmentation persists. Most importantly, the institution must abandon the comfort of managed ignorance. A prosecutor cannot constitutionally be left as the symbolic owner of knowledge the system refuses to integrate.
Within this volume, the significance of prosecutorial knowledge and information fragmentation is unmistakable. The Brady economy is sustained not only by concealment in the dramatic sense, but by the quieter and more durable practice of distributed non-knowing. The government’s disclosure obligation is broad because the danger of fragmented state knowledge is broad. When institutions preserve informational separation while retaining the benefits of collective prosecution, they transform administrative architecture into constitutional risk. The prosecutor then becomes the visible face of a much deeper problem: a justice bureaucracy that wants the advantages of aggregated governmental power without the burdens of aggregated governmental knowledge. That imbalance is not a minor operational defect. It is one of the core mechanisms by which constitutional fairness is weakened, responsibility is diffused, and the Brady economy perpetuates itself.