Goodhart's Law

Goodhart's law is the observation that "when a measure becomes a target, it ceases to be a good measure." The compact phrasing is due to anthropologist Marilyn Strathern; the underlying insight comes from British economist Charles Goodhart, who in 1975 noted that any statistical regularity used as a policy target will collapse once policy starts pressing on it. The mechanism is simple: actors optimize for what is measured rather than what the measure was meant to proxy, so the correlation between measure and underlying goal breaks down. Classic illustrations include schools optimizing for standardized test scores rather than learning, hospitals optimizing for surgery-survival statistics by refusing risky cases, and machine-learning systems exploiting loopholes in reward functions. In academia, h-index, impact factor, and raw citation counts are now widely cited examples: once these became hiring and promotion criteria, gaming followed — coercive citation, citation cartels, salami-slicing papers to inflate counts, and inflated co-authorship. Goodhart's law is not a claim that metrics are useless; it is a warning that any metric durable enough to be used for incentives will eventually be optimized against, and that ecosystems relying on metrics for coordination need fresh, harder-to-game signals over time.