The gap between potential and actual Pareto improvement is where most real politics happens

Many proposed reforms are Kaldor-Hicks efficient (winners could compensate losers) but not actually Pareto-improving (losers aren’t in fact compensated). The distinction matters enormously, and clever mechanism design often can’t close the gap because those with power to implement the mechanism are those who benefit from compensation not being paid.

Explanandum

Why do policies that are demonstrably positive-sum in aggregate consistently produce uncompensated losers? Is this a design failure or a structural feature?

Substance

Kaldor-Hicks efficiency (1939) was developed precisely to justify policies where some people lose — the argument being that if aggregate gains exceed aggregate losses, the policy is efficient regardless of whether compensation actually occurs. This became the intellectual foundation for cost-benefit analysis in British public policy.

All three Works in Progress articles propose what look like positive-sum interventions. Land readjustment claims everyone gets a more valuable plot. Desalination claims to free downstream nations without confronting upstream powers. Better competition metrics claim to help regulators make better decisions. But in each case, there are potential losers: the landowner whose non-financial attachment to their plot isn’t captured by value metrics; the inland farmer who can’t access coastal desalination; the workers in zombie firms that “should” have died.

English enclosure is the historical archetype: aggregate agricultural output increased dramatically, but compensation for displaced commoners was minimal. Whether this is “necessary modernisation” or “class robbery with parliamentary sanction” depends on your politics, but the Pareto improvement claim was false for a substantial population.

As Amartya Sen argues, a criterion that declares a policy efficient regardless of whether losers are compensated is really just a way of saying efficiency and justice are separate questions — then answering only the efficiency question while implying you’ve answered both.

Supports

• Enclosure increased aggregate output but displaced commoners — the canonical case of Kaldor-Hicks without actual compensation
• UK rail privatisation created “efficiency” gains for some while producing coordination failures and rent extraction for others
• Land readjustment’s value-based compensation doesn’t capture subjective/non-financial welfare losses

Challenges

• Some Pareto improvements genuinely are achievable — the question is how often, not whether ever
• The demand for actual Pareto improvement in all cases is itself a form of conservatism that blocks beneficial change
• Robust compensation mechanisms (retraining programmes, transition funds) can close the gap in practice, even if imperfectly

Open Questions

• Is the gap between potential and actual Pareto improvement a contingent policy failure (we could design better compensation) or a structural feature of political economy (those who control the process benefit from non-compensation)?
• Does the Ise principle help — if everyone expects periodic restructuring, is the perceived loss from any single round smaller?
• Can AI-assisted optimisation narrow the gap by better identifying and targeting compensation?

Source Context

Emerged as a meta-theme across all three articles — each proposes a positive-sum mechanism that, on closer inspection, involves potential losers whose compensation is uncertain.