Goodhart's Law

Origin and meaning

British economist Charles Goodhart (1975) observed that once the Bank of England began targeting the money supply, the previously stable relationship between money and prices and incomes broke down. He wrote that "any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes," because financial institutions and markets changed their behavior and products in response to the new controls (see monetary aggregates). Anthropologist Marilyn Strathern (1997), analyzing British university audits, generalized it as "when a measure becomes a target, it ceases to be a good measure," the phrasing widely known today. Around the same time, the Lucas critique (1976) gave a macroeconomic argument that policy changes alter expectations and behavior, so past relationships do not hold.

Campbell's law and incentive theory

Social psychologist Donald Campbell (1979) proposed that "the more any quantitative social indicator is used for social decision-making, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor." In management, Steven Kerr (1975) described the same problem as "the folly of rewarding A, while hoping for B." In economics, Holmström and Milgrom's (1991) multitask principal–agent model explains it: when work has several dimensions but only some are measured and rewarded, people shift effort toward measured tasks and neglect the rest, so strong pay-for-performance is not always desirable. Manheim and Garrabrant (2018) classified Goodhart effects into cases where a chance correlation breaks at extremes, where the causal structure is misunderstood, and where agents deliberately game the metric.

Cases and current issues

Bevan and Hood (2006) reported that targets in the English health system, such as four-hour emergency waits, improved the numbers but produced "gaming," such as holding patients in ambulances. Jacob and Levitt (2003) statistically detected teachers manipulating answer sheets after high-stakes testing was introduced in Chicago public schools. In academia, incentives rewarding publication counts and significant results have been argued to spread poor research practices (Smaldino and McElreath, 2016), and Jerry Muller (2018) criticized the wider pattern as "the tyranny of metrics." More recently, "reward hacking" in artificial intelligence, where overoptimizing a reward model lowers true quality, has been studied as a case of Goodhart's law (Gao et al., 2023). The lesson is not to abandon measurement but to use multiple indicators, check validity and combine KPIs with qualitative judgment.

FAQ

Are Goodhart's law and Campbell's law the same?

They are very similar. Goodhart started from statistical relationships in monetary policy that collapse once targeted; Campbell started from social indicators that face corruption and distortion pressures once used for decisions.

Should performance metrics be abandoned, then?

No. Metrics remain useful, but it helps to avoid attaching large rewards to a single indicator, to combine several indicators with qualitative assessment, and to keep checking whether they are being gamed.

Sources reviewed: 2026-10-26

References

1. Goodhart, C. A. E. (1975). Problems of monetary management: The U.K. experience. Papers in Monetary Economics, 1. Sydney: Reserve Bank of Australia.
2. Campbell, D. T. (1979). Assessing the impact of planned social change. Evaluation and Program Planning, 2(1), 67–90.
3. Lucas, R. E. (1976). Econometric policy evaluation: A critique. Carnegie-Rochester Conference Series on Public Policy, 1, 19–46.
4. Strathern, M. (1997). 'Improving ratings': Audit in the British university system. European Review, 5(3), 305–321.
5. Kerr, S. (1975). On the folly of rewarding A, while hoping for B. Academy of Management Journal, 18(4), 769–783.
6. Holmström, B., & Milgrom, P. (1991). Multitask principal–agent analyses: Incentive contracts, asset ownership, and job design. Journal of Law, Economics, and Organization, 7, 24–52.
7. Bevan, G., & Hood, C. (2006). What's measured is what matters: Targets and gaming in the English public health care system. Public Administration, 84(3), 517–538.
8. Jacob, B. A., & Levitt, S. D. (2003). Rotten apples: An investigation of the prevalence and predictors of teacher cheating. Quarterly Journal of Economics, 118(3), 843–877.
9. Manheim, D., & Garrabrant, S. (2018). Categorizing variants of Goodhart's law. arXiv:1803.04585.
10. Muller, J. Z. (2018). The Tyranny of Metrics. Princeton, NJ: Princeton University Press.
11. Smaldino, P. E., & McElreath, R. (2016). The natural selection of bad science. Royal Society Open Science, 3(9), 160384.
12. Gao, L., Schulman, J., & Hilton, J. (2023). Scaling laws for reward model overoptimization. Proceedings of ICML 2023, PMLR 202, 10835–10866.