Investigated
What is Goodhart's Law?
Evidence strength: The core definition is well established: applying pressure to control a measure can make the underlying statistical relationship deteriorate. Debate concerns how broadly the principle applies and whether it is a useful warning rather than an absolute law.
Short answer
Image: pmc.ncbi.nlm.nih.gov
Goodhart’s Law says that when a measure becomes a target, people and institutions adapt to it, so it may stop reflecting the underlying goal. Charles Goodhart originally described statistical regularities collapsing under pressure for control; the shorter wording was later popularized by Marilyn Strathern. It is a warning to treat metrics as imperfect proxies, especially when incentives encourage optimization or gaming. The effect has also been studied experimentally in reinforcement learning, where optimizing an imperfect reward can reduce performance on the true objective.
The views on this question
Two lenses on the same evidence. Source weight and the primary source ratio show what each rests on.
Alternative
Source weight 50%Lens adapted to this topic: Limits, qualifications, and practical use
A qualified reading argues that the slogan is too broad if treated as an automatic law. Targets can sometimes remain useful, and the practical challenge is to design measures, incentives, and review processes that reduce gaming and detect when a metric no longer tracks the goal. The law is therefore a cautionary heuristic, not a ban on targets.
- The slogan “when a measure becomes a target” can overstate the case if interpreted as claiming that targets inevitably become useless.
- Goodhart’s original formulation was narrower than the popular version and arose from monetary-policy indicators whose relationships changed after they became targets.
- Statistical regularities may be genuine yet still depend heavily on the institutional conditions under which the data were produced.
- Practical responses include using balanced sets of metrics and reassessing whether indicators still represent the intended objective.
Follow the white rabbit
Deeper threads worth pulling on next.
- How is Goodhart's Law applied in AI?Different angle, same storyExplore how reward hacking and proxy optimisation cause AI to exploit flawed metrics, with RL examples.
- What are examples in education?A different rabbit holeLook at how high-stakes testing changed teacher and student behavior, causing teaching-to-the-test and score inflation.
- How does healthcare face Goodhart problems?