The Unaccountability Machine

Davies's subject is decisions that no one makes. Large organizations hand choices to rule books, procedures and software, creating what he calls accountability sinks: places where the feedback link between the people who bear a decision's consequences and the people who make it is cut. One example he uses is the 1999 episode in which KLM staff at Schiphol shredded 440 squirrels shipped without proper paperwork, after which the airline said the staff had acted correctly under the rules.

Davies worked as a regulatory economist at the Bank of England and as an analyst at several investment banks, and he worked on cases including the LIBOR and FX scandals and the collapse of Anglo Irish Bank. His earlier book, Lying for Money, was a field guide to financial fraud. He writes for the Financial Times and The New Yorker.

The book is part biography of Stafford Beer, the British management cyberneticist whose most ambitious project, Cybersyn in Salvador Allende's Chile, ended with the 1973 coup. It is also part critique of economics, which in Davies's account squeezes messy information into models that must produce an optimal answer. It was longlisted for the FT Business Book of the Year, and The Economist counted it among 2024's most useful books. Ed Smith in the New Statesman thought Davies blamed economics too heavily and doubted that Beer deserves a hero's role, and Mandy Brown found the proposed fix of curbing private equity too easy.

Risk managers, operations leaders, regulators and private equity investors will get the most out of it.

1. Look for the accountability sink. When a bad outcome cannot be traced to any person, the decision has been handed to a rule book or system. Finding where the feedback is lost is the first step to fixing it.
2. Accountability follows the power to change. Davies's core principle is that you can only be accountable for a decision to the extent you can change it. Assigning blame without that power is theater.
3. Treat complex organizations as black boxes. Following Beer, it is better to watch what a large system actually does than to assume its org chart or model explains it.
4. Optimizing for one metric blinds an organization. Davies cites the Dominion v. Fox News case as a chain of choices shaped by an implicit priority on audience metrics. Nobody in particular had to decide anything.
5. Cost cuts can remove a firm's ears. Stripping slack, such as thinning a hotel's staff until guests have no one to complain to, saves money but also cuts the channels that tell management something is wrong. Davies ties this kind of lean, metric-driven management to the shareholder-first era.

The book has four parts, each followed by a short intermission.

1. The Nature of the Crisis. Opening chapters, including one on Stafford Beer and one called Aliens Among Us, set out the idea that organizations behave like non-human intelligences that make decisions none of their members intend. An intermission covers computing ponds and rabbit holes.
2. Pathologies of the System. Davies explains how to psychoanalyse a non-human intelligence and presents cybernetics without diagrams, introducing Beer's distinction between systems you can fully understand and black boxes you can only observe. The intermission uses Beer's image of a decerebrate cat, which can still move but cannot act with purpose.
3. The Blind Spots. Chapters on how economics got the way it is, and on why rich institutions are not necessarily smart ones, argue that the discipline built a large blind spot about information and organization. The intermission returns to Chile.
4. What Happened Next? Enter Friedman traces the shareholder-value turn, The Morbid Symptoms catalogs today's dysfunction, and What Is to Be Done? proposes remedies, including reining in private equity.