There is No Meritocracy Without Lottocracy

Why random selection is necessary to create stable meritocratic institutions

Campbell's Law (a variant of Goodhart's Law) states that the more a metric is used for social decision-making, the more it will be subject to corruption which distorts and corrupts not only the metric itself, but the very social processes it was meant to measure. Selection criteria for a position of authority are one example of such a metric. When selection criteria are opaque, it is difficult for them to become a target, preserving their utility as measures. For governance positions however, it's often impractical or impossible for selection criteria to be too opaque. Therefore, they end up being gamed, ceasing to be effective measures and undermining the legitimacy of both the position and the process. Positions of authority end up in the hands of those who are best at gaming the system rather than those who are most qualified.

Representative republics are a perfect example. The most effective administrators and policymakers tend to be intellectually curious, well-versed in law and economics, good at taking feedback, at least somewhat good at thinking creatively, and morally upstanding. In principle, good looks, oratory eloquence, a charming personality, well-connectedness, and personal wealth are not particularly useful to creating and executing government policy. But the primary selection criterion for public office is the ability to win elections. Thus, that is the target, not being a good policymaker. And being charming is far more useful than being studious when it comes to attaining public authority.

Hereditary monarchy is of course quite prone to this problem as well. You cannot make the selection criteria opaque, because that would cause constant succession crises. So, it has to be something transparent, such as a hierarchical list of heirs which determines who has legitimacy as a new ruler and who doesn't. Even if the rules of succession are immutable, they can still be gamed, by assassination for example. Therefore, factors such as getting away with poisoning your elder siblings or sending your nephews to battle on suicide missions become crucial in determining who gets authority.

Max Planck claimed that science advances one funeral at a time. Such sentiments are not uncommon across professions and disciplines. Relationships often outweigh merit when it comes to determinations of who gets authority. This is why KPIs (key performance indicators) and other data-driven approaches have become popular at many big firms. They know that absent such metrics, relationship-building ability ("politicking") would inevitably become a major, if not primary means of attaining advancement within the corporate hierarchy, and that this ability is often not very correlated with job performance. KPIs may be impersonal, lacking in depth, and still subject to being gamed, but at least they are more reliable indicators of performance than political ability.

Metrics such as KPIs may be adequate in narrow circumstances such as approximately evaluating an employee's ability to do a specific type of job, but they have a ceiling of complexity they are able to measure. Highly complex tasks with many dimensions of competency (often including an unknown number of unknown dimensions) cannot be adequately legibilized by quantitative metrics. There is also the matter of who gets to decide what performance indicators are used, how they are measured, and how they are used to inform decisions which are ultimately made by humans. You could design a purely mechanistic system which attempts to cut out human decision-making, but such systems lack flexibility and nuance and are typically abandoned quickly for good reason. In the end, simply having good quantitative metrics isn't enough on its own. It is the ultimate decision-making process itself that must be improved.

Randomness helps address these problems because it eliminates most marginal advantage from subversive scheming. With random selection, no action or investment can meaningfully improve one's chances, rendering efforts to manipulate the system worthless. This nullifies political capital and ensures that authority is not seized by those adept merely at influencing outcomes through charm, money, or connections. Instead, it creates a system where competency and merit have a genuine chance to rise naturally, unhindered by strategic manipulation. Moreover, randomness systematically dismantles entrenched crony networks by constantly disrupting established relationships.

How might meritocratic institutions actually harness randomness in practice though? The possibilities are virtually limitless, but here are some ideas:

1. Place critical appointment/hiring processes into the hands of randomly selected oversight boards. These boards manage appointments, evaluations, and dismissals, mitigating biases and discouraging the formation of insular power groups.

2. Directly select candidates at random for positions from an eligibility pool. Set and maintain the eligibility standard (such as an exam) by randomly selected oversight board to keep it updated and prevent the standard from being manipulated or gamed.

3. Firms could randomly select employees or shareholders to serve on their boards. These members can significantly dilute insider collusion and introduce perspectives often overlooked by traditionally selected executives.

4. Use stratified sampling to select committees, ensuring diverse representation of viewpoints, backgrounds, and expertise, contributing to balanced decision-making.

5. Create randomly composed auditing and oversight committees, deterring corrupt practices through constant unpredictability in oversight.

Common objections to randomness-based systems, such as fears of incompetence, accountability gaps, and scaling issues, are not that difficult to address. Research consistently demonstrates that groups, under the right conditions, routinely outperform individual experts. Training, sufficient group sizes, and coordination technology can further mitigate the risk of incompetence. Accountability can be ensured through transparent decision-making processes, post-hoc reviews, and recall mechanisms. Concerns about scale can be resolved through nested selection—multi-layered lotteries that maintain manageable group sizes even within large, complex systems.

Numerous historical and contemporary examples demonstrate the successful application of randomness to governance. Juries, widely trusted to impartially deliver justice, are the most familiar instance. The Republic of Venice maintained effective governance for centuries using sophisticated lottery mechanisms to select its councils and doge. Classical Athens relied extensively on randomly assigned administrative and judicial roles, curtailing factionalism and corruption. Modern citizens' assemblies similarly prove that random selection can generate thoughtful, informed policy decisions. Some counties in the state of Georgia use grand juries to appoint and oversee local officials whose duties are too niche and arcane to be held accountable by voters.

Randomness-based governance mechanisms provide a number of system-level benefits. They prevent intellectual stagnation by continuously introducing fresh mental models into positions of authority. This dynamic churn naturally fosters diversity without the need for cumbersome diversity bureaucracies. Even more significantly, randomness introduces corruption entropy—rapidly decreasing the lifespan of capture schemes, as each new lottery resets established relationships and loyalties. This continuous renewal grants resilience to institutions, much like genetic recombination in biological systems provides resistance against environmental shocks.

Lottery-based mechanisms do not inhibit excellence but rather serve as its safeguard. They prevent meritocracies from devolving into court societies dominated by schemers and sycophants. Integrating chance into selection processes creates a firewall, allowing genuine talent and ability to flourish, uncorrupted by the perverse incentives of traditional selection processes.