Professional services — law, medicine, accounting, consulting, financial advice, engineering — share a defining feature: clients purchase judgment they cannot themselves evaluate. This asymmetry is the reason the professions exist, and also the reason they are dangerous. A client who could verify the quality of legal reasoning would not need the lawyer; a patient who could assess the diagnosis would not need the physician. Trust fills the gap — but trust unsupported by verification decays into either naïveté or cynicism. Auditability is what allows trust to be earned rather than merely demanded. It is the property of a professional process that lets a third party reconstruct, after the fact, what was done, why, on what evidence, and by whom. Its importance is not bureaucratic. It is foundational to how professional services can be honest, improvable, and legitimate at all.
The verification problem in credence goods
Economists classify professional services as credence goods: their quality cannot be fully assessed even after consumption. If the audit was clean, the surgery successful, the merger advice sound — how would you know? The outcome is confounded by luck, and the counterfactual (what a better professional would have done) is invisible. Darby and Karni's classic analysis showed that credence markets are structurally prone to overtreatment, undertreatment, and overcharging, because the seller both diagnoses the need and supplies the remedy while the buyer can verify neither.
In such markets, quality cannot be disciplined by consumer choice the way restaurants are disciplined by diners. Something else must do the disciplinary work. Historically that "something" has been a bundle: licensure, professional ethics, reputation, and liability. But every element of that bundle depends, at bottom, on auditability. A malpractice claim cannot be adjudicated without records showing what the physician knew and when. A reputation cannot be accurate if failures are undetectable. An ethics code is aspirational poetry unless conduct can be reconstructed and compared against it. Auditability is the load-bearing infrastructure beneath every other trust mechanism.
What auditability actually consists of
Auditability is often confused with surveillance or paperwork. It is better understood as four linked properties:
Documentation of reasoning, not just conclusions. An auditable opinion records not merely "we recommend X" but the alternatives considered, the evidence weighed, the assumptions made, and the standards applied. The working papers behind an audit opinion, the differential diagnosis behind a treatment, the memo behind a legal position — these transform a bare assertion into an inspectable argument.
Attribution. Auditable work identifies who did what and who approved it. Signature creates accountability; anonymity dissolves it. This is why engineers stamp drawings and audit partners sign opinions personally — and why diffuse, unattributed collective work is where professional failures hide.
Traceability of evidence. Claims connect to sources: the lab result, the precedent, the confirmation from the bank, the data underlying the model. A conclusion whose evidentiary chain cannot be followed cannot be distinguished from a fabrication.
Independence of review. Records matter only if someone structurally positioned to find fault can examine them — a peer reviewer, an opposing counsel, a regulator, an external auditor. Auditability without independent audit is a diary, not a discipline.
Why it matters: four functions
First, deterrence and honesty. The expectation of ex-post reconstruction changes ex-ante behavior. Professionals who know their reasoning will be examined by a skeptical peer reason more carefully — the well-documented "audience effect." Conversely, the professional certain that no one will ever check operates under criminogenic conditions; incentive distortions (the subject of a vast literature on conflicts of interest) flourish precisely where verification is impossible. Auditability does not make people virtuous; it makes the dishonest path more costly and the honest path demonstrable.
Second, error detection and learning. Even perfectly honest professionals err, and complex judgment errs systematically. Without records, errors are unrecoverable: the misdiagnosis is buried with the chart that was never written. Aviation's transformation into the safest complex industry in history rested on radical auditability — flight data recorders, mandatory incident reporting, and blame-aware but learning-oriented investigation. Medicine's patient-safety movement, launched by the Institute of Medicine's finding that tens of thousands die annually from preventable error, explicitly imported this logic: you cannot fix what you cannot reconstruct. Auditability converts individual mistakes into institutional knowledge.
Third, fair adjudication. When things go wrong — and in probabilistic domains, they will even when everyone performed well — records are what allow the crucial distinction between bad outcomes and bad decisions. This protects clients from negligent professionals, but equally protects competent professionals from hindsight bias. The surgeon whose documented reasoning was sound at the time has a defense; the one with an empty chart does not, however good their actual judgment was. Auditability is thus not only a sword for clients but a shield for practitioners — a point professionals who resent documentation chronically underappreciate.
Fourth, legitimacy of the professional bargain. Professions hold state-granted monopolies (only licensed lawyers may practice law) in exchange for self-regulation in the public interest. That bargain is credible only if self-regulation is inspectable. When auditability fails at scale, the bargain collapses and regulation arrives from outside — the pattern is almost mechanical. Enron's implosion, enabled by an audit firm that shredded its own working papers, produced Sarbanes-Oxley and an end to pure self-regulation of American auditing (the PCAOB now audits the auditors). The 2008 ratings failures produced Dodd-Frank scrutiny of agencies. Each scandal follows the same script: unauditable judgment, catastrophic failure, forfeited autonomy.
The objections, and their limits
Auditability has real costs, and pretending otherwise discredits the case for it. Documentation consumes time that could serve clients; defensive record-keeping can degenerate into liability theater ("charting for the lawyers") that obscures rather than reveals; metrics invite gaming (Goodhart's law applies — auditable proxies can displace the unauditable substance they proxy for); and pervasive review can erode the professional autonomy and trust that make the work attractive to talented people. Onora O'Neill's influential critique of "audit culture" warns that indiscriminate accountability regimes can breed box-ticking while actual trustworthiness withers.
These objections are best read not as arguments against auditability but as specifications for good auditability: proportionate to stakes (the routine tax return needs less than the megamerger opinion), focused on reasoning rather than ritual, designed for learning rather than pure blame (the aviation model), and paired with genuine independence rather than performed compliance. The failure mode of too much bad auditing does not vindicate the alternative of unexaminable judgment; Enron had plenty of paperwork and no auditability of the kind that matters.
The stakes are rising
Two developments make auditability more urgent, not less. First, professional judgment is increasingly mediated by models and algorithms — diagnostic AI, legal research tools, robo-advisors, automated underwriting. These systems can be either the most auditable artifacts ever created (every input, weight, and output loggable) or the least (proprietary black boxes laundering unexaminable judgment through a vendor). Which of these futures arrives is a design and regulatory choice being made now, and the professions that embraced auditability for humans have every reason to demand it of machines performing the same functions. Second, services are globalizing and disaggregating — work passes through outsourced chains where no single professional sees the whole. In such chains, auditability is the only thread connecting the client to the actual work.
Conclusion
Professional services rest on an unavoidable asymmetry: the client cannot verify what they are buying. Auditability is civilization's answer to that asymmetry — not because inspection happens often, but because the possibility of inspection disciplines conduct, enables learning, permits just adjudication, and legitimates professional privilege. A profession that resists auditability is asking for trust while dismantling the only machinery by which trust can be checked against reality. The demand "trust me" and the demand "verify me" are not opposites; in expert domains, the second is the only durable foundation for the first. The mark of a mature profession is not that it is never wrong, but that when it is wrong, one can find out — and the profession itself wants to.
References
- Darby, M. R., & Karni, E. (1973). "Free Competition and the Optimal Amount of Fraud." Journal of Law and Economics, 16(1), 67–88.
- Kohn, L. T., Corrigan, J. M., & Donaldson, M. S. (Eds.) (2000). To Err Is Human: Building a Safer Health System. Institute of Medicine, National Academies Press.
- O'Neill, O. (2002). A Question of Trust: The BBC Reith Lectures 2002. Cambridge University Press.
- Power, M. (1997). The Audit Society: Rituals of Verification. Oxford University Press.