Kaldor-Hicks Efficiency

The compensation principle

Pareto's criterion is unforgiving: a policy passes only if literally no one is worse off. By that standard, almost no real-world policy passes — every tax, every regulation, every infrastructure project hurts someone. Welfare economics needed a less stringent test.

Kaldor and Hicks supplied one in 1939. Their idea: ignore whether anyone is actually compensated. Ask only whether the policy generates enough surplus that the winners could in principle compensate the losers and still come out ahead. If yes, the policy is "Kaldor-Hicks efficient" regardless of whether transfers happen.

Operationally, this means: monetize every gain and every loss, sum them, and approve the policy if the net is positive. This is cost-benefit analysis. The Kaldor-Hicks criterion lurks behind every Environmental Protection Agency rule-making document, every OMB Circular A-4 review, every EU regulatory impact assessment.

Two distinct tests fall under the umbrella:

• Kaldor's test: The change from A to B is efficient if winners (those better off in B) could compensate losers (those worse off in B) and still be better off than in A.
• Hicks's test: The change is efficient if losers in B could not bribe winners enough to make them prefer staying in A.

Usually the two tests agree. When they disagree — when Kaldor approves the move from A to B and Hicks approves the reverse from B to A — you have the Scitovsky paradox.

Worked example: a road through a town

A new highway will save 1,000 commuters $200/year each in time savings. It will displace 20 households who valued their homes at $40,000 above the market price (sentimental attachment, network ties).

Group	Per-person impact	Group size	Total welfare effect
Commuters (winners)	+$200/year × 30 years = +$6,000	1,000	+$6,000,000
Displaced households (losers)	−$40,000	20	−$800,000
Net			+$5,200,000

Kaldor-Hicks: the project is efficient. Winners gain $6 million; losers lose $800k. If we taxed each commuter $1,000 over 30 years and paid each displaced household $50,000 — every party would be better off than the no-project baseline. That hypothetical compensation makes the project pass.

The catch: in real life, no compensation is paid. The town builds the road, the commuters keep their savings, the 20 households move. The "efficiency" is realized, but 20 families bear the entire cost.

This is exactly where Kaldor-Hicks earns its critics. The criterion declares the project efficient based on aggregate willingness-to-pay even though the distribution is patently unjust without transfers. Modern cost-benefit practice often supplements raw Kaldor-Hicks with distributional weights or environmental justice screens to address this.

Welfare criteria, side by side

	Pareto	Kaldor-Hicks	Scitovsky	Cost-benefit (weighted)	Utilitarian	Rawlsian
Requires actual compensation	—	No	No	Sometimes	No	No
Allows losers	No	Yes	Yes	Yes	Yes	Only if min improves
Avoids Scitovsky reversal	Yes	No	Yes	Yes	Yes	Yes
Distribution-sensitive	Implicit	No	No	Yes (weights)	No	Yes
Used by OMB / EPA	No	Yes	Rarely	Yes	Rarely	No
Information needs	Ordinal	Cardinal $	Cardinal $	Cardinal $ + weights	Cardinal utility	Min utility

Kaldor-Hicks's appeal is its tractability: every gain and loss reduces to a dollar value, the math is unambiguous, and the answer is decisive. Its weakness is the same: by reducing everything to dollars, it submerges distributional concerns.

The Scitovsky paradox

Tibor Scitovsky (1941) noticed a logical glitch. Suppose state A has a different distribution than state B. The Kaldor test asks whether, in B, winners could compensate losers using B's prices. But moving from A to B changes prices, which changes what compensation looks like.

This means: a move from A to B can pass Kaldor (winners in B could compensate losers in B at B's prices) and the reverse move from B to A can also pass Kaldor (winners in A — formerly losers in B — could compensate losers in A — formerly winners in B — at A's prices). Both moves "improve" efficiency. By transitivity, the criterion is incoherent.

Scitovsky proposed a double-test fix: a change is Scitovsky-efficient if Kaldor approves the forward move and Kaldor does not approve the reverse. This eliminates the cycle but is rarely applied in practice — analysts almost always use raw Kaldor-Hicks.

Empirically, Scitovsky reversals are most common when the policy substantially changes relative prices (e.g., trade policy that reshapes the cost of imports), and rarer for small projects. This is some comfort, but no theoretical solution.

Critiques

Distributional blindness. Counting a $1 gain to a billionaire as worth the same as a $1 loss to someone in poverty contradicts most ethical frameworks. Defenders argue that explicit distributional weighting can be added on top; critics argue that doing so abandons the criterion's claimed objectivity.

The wealth effect. Willingness to pay (used to monetize benefits) and willingness to accept (used to monetize losses) depend on existing wealth. A policy that helps the rich and hurts the poor passes Kaldor-Hicks more easily than one with the reverse pattern, because rich winners "value" the gain more in dollar terms.

Hypothetical consent. Critics like Ronald Dworkin argue that imposing costs on people who could have been compensated but weren't isn't real consent. Real Pareto improvement requires real compensation; Kaldor-Hicks substitutes a thought experiment for moral standing.

Aggregation problems. Even within Kaldor-Hicks's frame, summing dollar gains across people requires assumptions about constant marginal utility of income. Empirical evidence suggests this fails — a dollar means more to a poor person than a rich one.

Variants

• Scitovsky double-test: Kaldor-positive forward and Kaldor-negative reverse. Closes the cycling loophole.
• Little criterion (1950): Adds a distributional check — the new income distribution must not be worse than the old.
• Distributional weighting: Assigns higher weight to gains/losses for poorer groups. Ramsey/Atkinson/Stern weighting schemes formalize this.
• Required compensation (actual Pareto improvement): Insist that winners actually pay losers. The result is Pareto-improving but politically and administratively expensive.
• Hicks-Allais discount: Adjusts for the fact that current losers and future winners may not be the same generation.

Common pitfalls

• Forgetting compensation is hypothetical. "Kaldor-Hicks efficient" doesn't mean "fair" — it means "could be made fair if anyone bothered to transfer."
• Treating WTP as ground truth. Willingness to pay reflects budget constraints, not need. The same lung-cancer death is "worth" more if it happens to a rich person on the WTP measure.
• Ignoring Scitovsky reversal. If your policy reshapes prices significantly, check whether the reverse passes too.
• Conflating Kaldor-Hicks with utilitarianism. Utilitarianism sums utility; Kaldor-Hicks sums dollar willingness-to-pay. The two coincide only when marginal utility of income is constant — a strong assumption.
• Skipping distributional analysis. Modern regulatory practice supplements Kaldor-Hicks with explicit distributional and equity assessments. Reporting a single net-present-value number is no longer best practice in serious policy work.