Goodhart's Law

Goodhart's Law explains why setting targets often backfires. When you turn a measure into a goal, people focus on hitting the number rather than achieving what you actually care about. It's like a call center tracking call length—agents rush people off the phone to keep their average time low, but customer satisfaction actually drops because problems don't get solved. Notice this: whenever a metric becomes a target, check what's being sacrificed to hit it.

Goodhart's Law, named after economist Charles Goodhart who observed this pattern in 1975 monetary policy, captures a universal problem: measurement changes behavior. When a metric becomes a target, people optimize for the metric instead of the underlying value it's meant to track.

In education, this shows up when teachers are evaluated on standardized test scores. They may narrow their teaching to test content, skip deeper conceptual learning, or even encourage low-performing students to stay home on test days. The scores look good on paper, but actual learning suffers.

Business sees similar effects. Sales teams rewarded solely on revenue quotas may push products customers don't need, offer unsustainable discounts, or close sales that quickly cancel. Healthcare systems judged on patient wait times may prioritize easy quick cases over complex urgent ones, or discharge patients prematurely leading to readmissions.

The law doesn't mean metrics are useless—it means single measures are never perfect. Smart organizations use multiple indicators, rotate metrics to prevent gaming, and combine quantitative measures with qualitative judgment. The key is recognizing that any measurement system can and will be manipulated once people understand it.

Goodhart's Law originated in monetary policy but applies universally to any system where metrics drive decisions. The core insight: once people know what's being measured and that it affects their outcomes, they'll optimize for the measure—often at the expense of the actual goal.

• Goodhart (1975): Originally observed that statistical relationships used for monetary policy collapse when used as targets, as market participants adapt to the policy regime [1].
• Strathern (1997): Popularized the formulation "When a measure becomes a target, it ceases to be a good measure" in the context of British university accountability systems [2].
• Manheim & Garrabrant (2018): Systematically categorize four distinct mechanisms of Goodhart effects—regressional, causal, and adversarial subtypes—showing the problem has multiple failure modes, not one [3].
• Muller (2018): Documents how metric fixation in education, medicine, policing, and business displaces professional judgment and creates perverse incentives despite good intentions [4].

Goodhart's Law isn't inevitable—it depends on how metrics are designed and used. Some limitations and exceptions: When metrics closely match true goals and aren't easily gamed (like direct outcome measures), they may remain effective. Professional cultures with strong ethical norms can resist pressure to optimize metrics at the expense of real quality. Short-term target-setting for specific initiatives can work when combined with broader qualitative assessment. The law also doesn't specify how much degradation occurs—some metrics remain partially useful even when targeted. Related concepts like Campbell's Law focus specifically on social indicators and corruption pressures, while the Lucas critique in economics addresses how policy changes invalidate historical econometric relationships.