Responsibility Laundering

Consider an arrangement with the following features. Policy is set centrally. Information is held centrally. Benefit accrues centrally. Execution, along with its risk, is pushed outward to agents who see only their own part. When harm results, no individual at the centre made the decision that caused it, and each executor can point truthfully to the narrowness of what was within his control.

The arrangement is not a failure of design. It is a design. Chapter Seven, section XIV, gives the manoeuvre a name: responsibility laundering, which occurs when power designs fragmentation and later invokes that fragmentation as evidence that no powerful agent is answerable.

The principle it violates

The distribution principle holds that the degree and form of answerability should track the agent’s situated share of causal contribution, epistemic access, practical control, role authority, benefit, exposure to cost, and capacity to alter the conditions of action.

It is not an algorithm, and the book is explicit that it is not. The factors can conflict, and their weight depends on the type of demand at issue. Its function is negative: to prevent responsibility from following a single variable such as intentional action, formal office, causal proximity, or metaphysical freedom.

Two of the seven factors do most of the work against laundering.

Practical control concerns the capacity to initiate, stop, redirect, or repair conduct, including control over procedures, budgets, personnel, records, and channels of escalation. Formal titles are evidence of control but not conclusive evidence. A nominal decision maker can be constrained by withheld information; an adviser without final authority can exercise decisive influence; a governing body can design delegation so that no individual experiences the outcome as chosen. The inquiry follows actual power rather than organizational labels.

Capacity to alter the conditions of action is the most important prospective factor. Agents with authority to change information flows, incentives, sanctions, or avenues of appeal bear responsibilities not exhausted by their personal conduct within the existing arrangement. They can be answerable for the architecture itself. Power creates responsibility not merely because it increases causal efficacy, but because it enables the agent to reorganize the practical fields of others.

The presumption that follows

The principle yields a presumption against downward displacement. An institution should not centralize benefit and policy control while dispersing risky execution to less powerful agents, and then treat the executors as the sole bearers of responsibility when harm occurs. Nor should it require those harmed to prove an architecture whose records and categories the institution itself controls.

That second clause matters more than it may appear. Where an institution holds the documents, defines the categories, and decides what is recorded, an evidentiary standard that appears neutral operates as a shield. The distribution of epistemic access is not a background condition of the inquiry. It is part of what the inquiry is about.

The opposite error

The book is careful to name the failure on the other side, and the care is worth reproducing, because an account that only ever finds responsibility upward is as crude as one that only finds it downward.

Indiscriminate structural exculpation is the opposite error. To say that a practice is systemic does not mean that no participant can understand or resist it. Structure can increase the responsibility of actors who maintain, conceal, or profit from it while retaining the capacity to act otherwise. Situated judgment therefore rejects both the fantasy of the sovereign individual and the fiction of a structure that acts without persons.

Institutions as architectures

Chapter Eight completes the thought by reading the practico-inert as an architecture of answerability. Material arrangements preserve prior praxis and confront later agents as necessity. They also preserve allocations of knowledge, authority, access, and risk. A form, a database, a schedule, a reporting line, a credential, or a property rule can determine who sees a problem, who may name it, and who can act on it.

These objects are not responsible agents. They organize the conditions under which agents become responsible. Five institutional functions are especially important: the epistemic function distributes information and interpretive categories; the authoritative function distributes powers of decision, refusal, and review; the cost function distributes the burdens of compliance and resistance; the justificatory function determines who must give reasons to whom and which reasons count; the remedial function distributes the capacity to correct, compensate, and redesign.

An institution’s responsibility profile can be analysed through those five. Laundering, described in that vocabulary, is the deliberate arrangement of the epistemic and justificatory functions so that the remedial function has nowhere to attach.

Developed in Chapter Seven, sections XIII and XIV, and Chapter Eight, sections XIV and XV.