A question as old as time

“Quis custodiet ipsos custodes?” - the famous Latin phrase traditionally attributed to the Roman satirist Juvenal, roughly translates as ‘Who will guard the Guards?”

Interestingly, the original phrase was used in the context of marital fidelity, of the impossibility of enforcing moral behaviour on women when the enforcers themselves were so susceptible to corruption (although the contextual history is disputed, and some scholars consider the relevant lines to be a later interpolation). Over the centuries, the phrase escaped its domestic origins and came to express a much broader problem: how do we hold to account those to whom we entrust the power to hold others accountable?

The question now finds application wherever power is entrusted with the expectation that it will watch over and prevent its subjects from corruption and overreach. These include governments and government bodies, judiciaries, regulators and corporations, auditors and boards, police forces and public institutions. Ultimately, it is a question of who exercises power, and who oversees those empowered to oversee others.

There is no dearth of examples where those in charge of oversight failed to constraint those they were supposed to oversee, uphold the rules of justice and fairness and protect the vulnerable. In this essay series Who will guard the Guards? I will explore fifty cases of institutional failures from history in which those charged with oversight – the guardians– failed to prevent or contain a material adverse outcome, and attempt to answer why the failures occurred, with each essay focusing on one ‘failure’.

Scope:

For the purposes of this project, a ‘guardian’ is a person, body or institution entrusted with a formal responsibility to supervise, regulate, assure, inspect, challenge, constrain or otherwise hold another actor or institution accountable.

For analytical purposes, I will consider oversight ‘failure’ along two dimensions:

  1. Guardian failure:  the guardian possessed reasonable capacity, authority, information and resources but failed to exercise its oversight function effectively.
  2. System failure: structural limitations in mandate, authority, independence, information, resources or institutional design materially impaired the guardian’s ability to exercise effective oversight.

These two categories are not mutually exclusive; a case may exhibit elements of both.

To conduct a comparative analysis of institutional oversight failures, I will select my sample of fifty cases across the following five dimensions of variation:

1. Institutional domain: Large and medium-sized private and publicly listed corporations, financial institutions, government and state-owned entities, NGOs and charities, media, arts and cultural institutions, and sports bodies.

2. Guardian type: Boards, audit committees and internal auditors, external auditors, regulators, rating agencies, inspectors, trustees, parliamentary and public oversight bodies, independent commissions.

3. Era: 1970-89, 1990-2009, 2010 – present.

4. Geography: North America, Western Europe, Eastern Europe & Russia, Middle East, South Asia, East Asia, Africa, Latin America and Oceania.

5. Nature/consequence of failure: Financial misconduct, corruption and abuse of power, safety incidents, ethical misconduct, operational catastrophes, concealment/misinformation, injury to humans, animals or the environment.

There is no dearth of examples where those in charge of oversight failed to constraint those they were supposed to oversee, uphold the rules of justice and fairness and protect the vulnerable

Selection criteria and inclusion protocol:

The decision to select fifty cases was made to provide sufficient breadth across the dimensions of variation while retaining the depth necessary to reconstruct and analyze each oversight failure individually. The sample is purposive rather than representative; it is designed to identify and compare mechanisms of oversight failure, not to estimate the frequency with which oversight institutions or bodies fail generally. Because the study examines failures rather than a matched sample of successful oversight interventions, I will treat recurring characteristics as mechanisms associated with failure rather than as evidence of causation or prevalence. The cases I select, however, will need to fulfill all of the following criteria before they enter the case pool:

  1. There was an identifiable guardian: Some persons or institutions had an explicit mandate to supervise, regulate, assure, inspect, challenge or hold another actor accountable.
  2. There was a material failure: Mere criticism or imperfect decision-making by a guardian will not qualify. There must be a material adverse outcome that the guardian could reasonably have been expected, within its mandate, to prevent, detect, escalate or contain.
  3. The oversight failure is documented independently: The failure and the relevant oversight circumstances are documented through an official inquiry, court findings, parliamentary investigation, commission reports or equivalent credible records (either through public or private bodies). Contemporaneous reporting by reputable news organizations may be used as supplementary evidence, particularly where primary records are unavailable, but material conclusions about the guardian’s conduct will, where practicable, be corroborated through independent sources.
  4. There is enough evidence to reconstruct the oversight process. Information is available about what the guardian knew, what authority it possessed, what warnings existed, what actions it took, and what happened after the actions were taken.

For each of the fifty cases, I will try to reconstruct what happened, and more importantly, why the institutions entrusted with oversight failed to stop it.

I will look at:

a. What the guardian was supposed to do: What was it protecting, what powers did it have and what was expected of it?

b. What it knew: What warning signs existed, when did they appear, and did the right people see them?

c. What it did: Did it challenge, investigate or intervene? If not, why not?

d. What got in the way: Were there conflicts of interest, lack of independence, powerful personalities, inadequate expertise or resources, fear or intimidation, complacency, or simply the assumption that someone else would act?

e. Who was watching the guardian: Did the secondary oversight fail as well, and why?

f. What happened afterwards: What changed, and did it work?

An ancient question

Centuries ago, Plato was already wrestling with a version of the same problem in The Republic: how could those entrusted with guarding the ideal city themselves be kept from corruption? His answer lay largely in the character, education, selection and way of life of the Guardians themselves. They were to be carefully selected, repeatedly tested and denied the private wealth that might turn Guardians into rulers pursuing their own interests.  At one point, discussing the discipline expected of the Guardians, Socrates remarks that a Guardian should never become so intoxicated that he no longer knows where he is. Glaucon, Plato’s brother and one of Socrates’ principal interlocutors in The Republic, finds the idea absurd: “Truly it would be ridiculous for a guardian to require a guard”, Glaucon retorts.

More than two thousand years later, we seem less certain. We have surrounded our guardians with other guardians: boards with regulators, auditors with oversight bodies, public institutions with inspectors, governments with courts and committees. Yet when oversight fails, our instinct is often to add another layer.

Over the course of these fifty cases, I want to understand what went wrong. Did the guardians themselves need stronger oversight? Were they too close to those they were supposed to watch? Did they lack independence, information or courage to act? Or did we create guardians and then deny them the authority and resources they needed to guard anything at all?

Or, most troublingly, is institutional decay inevitable? Given enough time, will every system of oversight tend towards complacency, ineffectiveness or irrelevance?

I don’t know yet what the fifty cases will reveal. That is rather the point.