Inequality Is Not a Problem
A distribution cannot tell you what is wrong
In 2025, I argued that inequality is not the problem—poverty is. Poverty deprives people of the resources required for effective agency. Inequality merely describes a difference between people. Someone does not become less capable because someone else becomes richer.
That argument remains correct, but it does not go far enough.
Inequality is not only a distraction from poverty. It is an analytically incoherent category of social harm. It combines radically different phenomena under a single statistical description, then encourages governments to treat compression of the distribution as evidence of moral progress.
A society may be unequal because innovators created enormous value, because dictators stole public wealth, because some people saved while others consumed, because monopolists captured the state, or because birth assigned different legal rights. An inequality statistic does not distinguish among these causes. It registers only the resulting distance between holdings.
That distance has no intrinsic moral meaning.
This matters because influential institutions increasingly speak as though inequality were an established scientific problem. A recent Nature editorial was titled “Why there needs to be a global debate on inequality” and declared that researchers have a responsibility to seek consensus on this “crucial issue.” The editorial was responding to a modelling report proposing extensive global redistribution in the name of equality, prosperity, and planetary limits.
The title assumes what would first need to be proved: that inequality itself constitutes a problem requiring collective correction.
It does not.
A distribution is not a moral diagnosis
Imagine two societies with identical wealth distributions.
In the first, the richest citizen acquired his fortune by developing a technology that millions of people voluntarily purchased. The buyers became better off. The inventor became much richer. No one was coerced, deceived, or deprived.
In the second, the richest citizen is a dictator who confiscated productive assets, suppressed competitors, and diverted public revenue into private accounts. The rest of the population is poorer because of his wealth.
The distribution may be numerically identical. The moral character of the two societies is opposite.
Now consider another pair.
In one society, nearly everyone lives at subsistence level and wealth is distributed almost equally. In the other, the poorest citizens enjoy secure housing, abundant food, advanced medicine, and considerable discretionary income, while a small number of people possess immense fortunes.
The second society is more unequal. It is also plainly preferable for nearly everyone who must live in one of them.
These examples expose what inequality statistics necessarily omit. A measure of dispersion contains no information about consent, coercion, production, deprivation, or legitimacy, and cannot tell us how wealth was obtained, whether anyone was harmed, or whether the poorest are flourishing.
Inequality is a measure of difference, not a theory of justice.
The inequality category conceals the actual problems
Arguments about inequality usually move rapidly among several distinct complaints.
Some people lack food, shelter, medical care, or the basic resources required to direct their own lives. That is poverty.
Some fortunes arise from monopoly protection, corruption, regulatory capture, or privileged access to the state. That is rent extraction.
Some people can use wealth to purchase exemptions from law or control political institutions. That is political domination.
Some people are legally confined to lower status by caste, race, sex, or citizenship. That is unequal treatment under law.
Some children are denied education, safety, or access to opportunities through circumstances beyond their control. That is capability deprivation.
Some economies prevent people from improving their condition through licensing barriers, exclusionary zoning, cartelization, or inherited privilege. That is blocked mobility.
Some goods are intrinsically positional. Only a limited number of people can own beachfront property, attend the most prestigious university, or occupy a high-status office. That is positional competition.
These phenomena differ in cause, moral status, and remedy. Poverty may call for assistance. Monopoly calls for competition. Corruption calls for institutional reform and prosecution. Legal hierarchy calls for equal rights. Blocked mobility calls for removal of barriers.
Calling all of them “inequality” discards the distinctions required to address any of them intelligently.
It also introduces a false problem: differences produced by voluntary exchange, innovation, risk-taking, inheritance, and luck are placed in the same category as differences produced by theft and political privilege.
The distribution cannot tell them apart. Public rhetoric generally does not try.
Provenance matters more than position
Suppose two people each possess ten million dollars.
One earned it by building a successful company. One received it through political corruption.
Their positions in the wealth distribution are identical. Their claims to the wealth are not.
Suppose two other people each have very little.
One has suffered an involuntary catastrophe. The other has chosen a low-income life that permits more leisure, artistic work, or personal independence.
Their positions in the income distribution are identical. Their circumstances are not.
Any serious moral evaluation must ask how a condition arose, whether the relevant choices were voluntary, whether anyone’s agency was violated, and whether the resulting claims are legitimate.
Inequality metrics erase this history. A dollar created through productive cooperation and a dollar extracted through coercion enter the same column. A voluntary difference and an imposed hierarchy appear as the same distance.
This is not a minor limitation that can be repaired by adding another variable. It follows from what inequality measures. Dispersion is insensitive to provenance by definition.
The morally relevant questions concern acquisition, consent, deprivation, and power. Relative position can sometimes serve as evidence about those questions, but it cannot answer them.
When inequality becomes evidence
None of this implies that extreme concentration should be ignored.
A sufficiently large fortune can purchase political influence, suppress market entry, shape regulation, or insulate its owners from ordinary legal constraints. Concentrated wealth can therefore become a causal risk factor for domination.
But this does not establish that numerical disparity is itself the harm.
The relevant mechanism is the conversion of economic resources into coercive or quasi-governing power.
That conversion is neither automatic nor unique to wealth. States have been captured by military, bureaucratic, religious, ethnic, and party elites whose power did not originate in private fortunes. Wealth can facilitate capture, but political structure determines how readily influence becomes control.
The structure of the state determines much of what is available to buy. It does not make the state immune to buyers.
Large fortunes can finance campaigns, lobbying organizations, media networks, and durable institutions capable of influencing the rules under which future wealth is acquired. Wealth can purchase existing privileges and help create new ones.
No constitutional order can eliminate this pressure.
A government with vast discretionary powers, opaque institutions, and saleable privileges invites capture. A government constrained by general rules, transparency, decentralization, and independent courts is harder to capture, but never capture-proof.
Concentrated wealth and concentrated state power therefore form a dangerous feedback loop. Each increases the threat posed by the other. Private wealth can bend public authority, while public authority can create and preserve private fortunes.
This makes extreme concentration a legitimate object of scrutiny without making inequality itself the offense.
The relevant question is whether wealth has been converted into unaccountable power: control over law, essential infrastructure, market entry, or resources on which others depend.
Extreme inequality can be evidence that something is wrong. It is never a sufficient diagnosis of what is wrong.
Scale and private power
The distinction between private resources and coercive privilege becomes harder to apply at very large scales, but it does not disappear.
A family bakery and a dominant communications platform are both privately owned assets. They do not exercise the same kind of power.
At sufficient scale, some private assets acquire quasi-governing functions. A dominant payment network, digital platform, transport hub, or utility may control access on which others depend. Its owners may be able to exclude competitors, impose terms without meaningful negotiation, or shape the practical conditions under which others exercise agency.
The concern is not merely that the owners possess a great deal. It is that the asset functions as infrastructure or a bottleneck.
Size alone does not settle the question. A billion-dollar firm in a contestable market may wield less durable power than a smaller regulated utility, local land cartel, or payment processor protected by high switching costs and legal barriers.
The relevant variables are contestability, network effects and switching costs, legal privilege, exclusionary conduct, and how far others depend on the asset. A large fortune may be benign. A smaller fortune attached to a strategic bottleneck may be dangerous.
“Massive” is not a causal mechanism.
Positional goods and relative purchasing power
Relative wealth matters where access is allocated by bidding for fixed or tightly constrained supply.
A richer person can outbid others for beachfront land, housing in a desirable location, elite educational places, or scarce medical capacity. In these cases, one person’s increased purchasing power can reduce another person’s effective access even if the latter’s nominal resources remain unchanged.
That is a real phenomenon, and it still does not convert inequality in general into a social harm.
The questions that matter: Is the good naturally scarce or politically restricted? Can supply expand? Is access allocated through an open market, inherited privilege, or state favoritism? Is the good essential to agency or merely desirable?
Housing scarcity in a city with exclusionary zoning is not adequately diagnosed by observing that rich people can pay more. The decisive mechanism is restricted supply. Manhattan land is naturally finite. Housing in the broader metropolitan region is not fixed to the same degree. A scarce luxury location and basic access to shelter are different problems.
An elite university is positional partly because prestige depends on exclusion. That does not imply that all educational inequality should be eliminated. It suggests that access to basic competence and opportunity should not depend on admission to a tiny set of status institutions.
Medical labour presents a more serious case. In the short run, specialists, operating rooms, and clinical capacity may be highly inelastic. Rich buyers can bid scarce resources away from poorer patients. Where an essential good is allocated primarily by price under fixed supply, relative purchasing power can affect agency directly.
This is a genuine distributive effect. It still requires analysis of the good, the supply constraint, and the allocation rule. Relative purchasing power matters in particular markets, and those markets require particular remedies.
Inheritance and the passage of time
A society cannot be evaluated only at the moment wealth is first created.
Fortunes persist. They are transferred across generations. Descendants may receive resources they did not earn and use them to acquire education, security, and access unavailable to others.
This can reduce mobility, entrench class structure, and magnify the conversion of private wealth into public power.
But “unearned” does not mean “illegitimate.”
A gift does not become wrongful because the recipient did not produce it. People ordinarily have strong claims to decide what happens to property they legitimately own, including the right to transfer it to their children.
The relevant distinction is between inheriting private resources and inheriting coercive privilege.
A child may inherit a house, a business, or a portfolio. No child should inherit legal immunity, monopoly rights, regulatory authority, or control over public institutions.
At large scales, inherited assets may nevertheless create quasi-governing power. A descendant who inherits a dominant platform, infrastructure system, or protected monopoly may inherit control over systems on which others depend. The distinction has not vanished. The inherited asset has crossed into a different category because of its function, not because inheritance itself is illegitimate.
Inherited wealth becomes a structural problem when it secures durable exclusion from essential institutions, suppresses contestability, or purchases political power.
A society committed to mobility should remove artificial barriers, preserve competition, and ensure that poverty does not extinguish basic agency. It need not treat every intergenerational transfer as an injustice merely because it widens the wealth distribution.
Correlation does not establish inequality as the cause
Defenders of inequality reduction often respond that highly unequal societies exhibit worse health, lower trust, more crime, and weaker mobility.
Some of these correlations may be real. They still do not establish that inequality itself causes the associated harms.
A society dominated by corrupt elites may produce both high inequality and low trust. A rigid caste system may produce both high inequality and poor mobility. Weak institutions may produce both concentrated wealth and violent crime. Monopoly, ethnic conflict, and state capture may generate an unequal distribution while independently generating poor outcomes.
In such cases, inequality is a symptom of the causal structure, not the active ingredient.
The distinction is decisive because interventions aimed at the symptom can leave the mechanism intact. A state may tax the nominally rich while preserving the regulations, political connections, and institutional privileges through which wealth is extracted. It may reduce measured income inequality while expanding the discretionary power that enabled capture in the first place.
A correlation between inequality and some harm therefore supports investigation, not compression as a default remedy.
One must still identify the mechanism. Does relative wealth itself reduce health, or does poverty do so? Does dispersion destroy trust, or does a visible system of privilege destroy it? Does inequality cause political capture, or does a state with saleable powers create both capture and concentrated wealth?
Without causal discrimination, “inequality causes bad outcomes” functions less as an explanation than as a permission slip for redistribution.
Redistribution does not become justified because someone benefits
The conceptual confusion becomes most dangerous when inequality is treated as sufficient justification for coercive redistribution.
A policy that takes resources from one person and gives them to another will usually benefit the recipient. That does not establish that the recipient had a valid claim against the person compelled to pay.
Benefit alone cannot justify coercion. If it could, any taking would become permissible whenever the recipient’s expected gain exceeded the owner’s estimated loss. Property, consent, and individual agency would survive only at the discretion of whoever performs the calculation.
Axionic ethics rejects this aggregation. Coercive redistribution harms those subjected to it because it reduces their agency through threats of fines, seizure, or imprisonment. Compassionate intentions do not erase the coercive structure, and benefits to one group do not cancel harms imposed on another.
Restitution is different. If property was acquired through theft, fraud, corruption, or coercive privilege, returning it to those with a valid claim is not justified by a desire for greater equality. It is justified because the current holding is illegitimate.
The distinction is between rectifying a specific wrong and manipulating a distribution.
Poverty creates a strong reason for voluntary aid. It does not by itself prove that a particular stranger owes the person in poverty an enforceable debt. Need and entitlement are not synonyms.
A society concerned with poverty should ask how assistance can be provided while preserving agency and consent. A society concerned with inequality asks how much force is required to produce a preferred statistical pattern.
Those are different projects.
When intervention is justified
The rejection of inequality as a policy target does not imply that all concentrations of wealth must be left untouched.
Intervention may be justified where wealth is sustained by fraud, coercive privilege, monopoly maintained through exclusion, corruption, or actual control over essential bottlenecks that others cannot reasonably avoid.
These are not offenses of possession. They are offenses of domination, exclusion, or illegitimate acquisition.
Antitrust enforcement, forced divestiture, ownership limits, or utility regulation may therefore be defensible in some cases. Their justification must rest on a demonstrated mechanism of harm, not on the size of the fortune alone.
Risk is not guilt.
A person does not become liable to coercion merely because his resources might someday be used badly. Prophylactic intervention requires a much stronger case than discomfort with concentration.
The threshold should be actual or credibly imminent domination: control over market entry, political authority, indispensable infrastructure, or systems on which others depend.
This boundary will always be contested. That does not justify replacing it with a wealth percentile.
The social contract does not settle the question
A common response is that property rights depend on state enforcement, so society may alter those rights through taxation and redistribution.
This argument does not follow.
Courts enforce contracts, bodily autonomy, freedom of association, and property rights. The fact that an institution protects a right does not mean the institution owns the right or may redefine it without limit.
State enforcement may justify charging for the provision of legal order under some theory of political obligation. It does not automatically establish a general authority to transfer resources among citizens whenever officials believe aggregate welfare would improve.
The phrase “social contract” supplies no answer unless the contract’s terms, limits, and basis of consent are specified.
Who agreed to it?
What binds dissenters?
Which powers were delegated?
What prevents the state from continually revising the terms in its own favor?
Why does participation in a legal order create an enforceable duty to finance another person’s consumption rather than merely the institutions that protect equal rights?
These are the disputed questions. Invoking the social contract does not resolve them.
Why institutions prefer inequality
If inequality is such a poor diagnostic concept, its institutional popularity needs explanation.
One reason is measurability. Poverty, coercion, legitimacy, mobility, and political capture are difficult to characterize. They require causal analysis and often resist reduction to a single index. Income shares and Gini coefficients are easy to calculate, compare, and display.
A scalar measure creates the appearance of scientific precision. Countries can be ranked. Targets can be established. Progress can be plotted. A morally and causally heterogeneous society becomes a line on a dashboard.
Another reason is administrative convenience. Governments can change measured post-tax income distributions directly through taxation and transfers. They cannot as easily create high-trust institutions, competent schools, open markets, or governments resistant to capture. A metric that responds immediately to state action is institutionally attractive, even when the action does little to resolve the underlying problem.
Inequality is also politically useful. The term unifies people with very different grievances: the poor, opponents of monopoly, critics of inherited privilege, egalitarians, and those who simply resent the rich. A vague category sustains a broader coalition than a precise diagnosis would.
The language carries a further rhetorical advantage. “Inequality” evokes unequal legal status and discriminatory treatment even when the subject is merely unequal outcomes. Moral condemnation properly directed at caste systems, racial law, and inherited political rank is transferred to ordinary differences in income or wealth.
The resulting politics rarely needs to prove that a particular disparity caused harm. The existence of the disparity is presented as the harm.
Science cannot discover the correct wealth ratio
The Nature editorial calls on researchers to search for consensus on inequality. But researchers cannot empirically discover how unequal a society is morally permitted to be.
Science can investigate factual questions. Economists can estimate how taxes affect investment, migration, and capital formation. Social scientists can study mobility, poverty, monopoly, and political capture. Climate researchers can estimate environmental constraints and externalities.
None of these disciplines can derive an ideal income ratio from data alone.
A claim that no person should possess more than ten, one hundred, or one thousand times the wealth of another is a normative proposition. It depends on assumptions about entitlement, desert, freedom, authority, and the moral significance of relative position.
Researchers may hold views on those subjects. Their scientific credentials do not convert those views into findings.
The demand for consensus obscures the distinction among empirical, moral, and political questions. What causes poverty is an empirical question. Whether one person may be coerced to relieve another person’s poverty is a moral question. Which institution may exercise that coercion is a political question.
A model can estimate consequences under specified assumptions. It cannot establish the legitimacy of those assumptions by producing numerically precise outputs.
When a desired distribution is built into the objective function, the model does not discover that distribution. It operationalizes it.
Compression is not progress
Once inequality becomes a policy target, reduced dispersion is treated as evidence of success.
But a society can become more equal through mass impoverishment. It can become more unequal through broad prosperity combined with exceptional gains at the top. It can reduce income inequality by destroying productive industries, driving away skilled workers, or preventing successful firms from scaling. It can increase wealth inequality because a new technology creates vast consumer value and rewards its creators.
The direction of the inequality statistic does not tell us whether the change was good.
Nor does it tell us who gained agency. A policy may compress incomes while making nearly everyone poorer. Another may increase dispersion while raising the living standards of the poorest. A third may leave the distribution unchanged while replacing corrupt fortunes with legitimate ones.
If the policy objective is human flourishing, agency, or justice, inequality is at best incidental.
Reduction in inequality is not evidence of success. Increased inequality is not evidence of failure.
Postscript
A society should not ask whether it is too unequal.
It should ask whether people possess the resources required for meaningful agency. It should ask whether wealth was acquired through production or extraction, whether markets are open to competition, whether political power can be purchased, and whether legal rights apply equally.
It should ask whether concentrated wealth controls essential bottlenecks, whether private assets have acquired quasi-governing functions, and whether inherited fortunes are preserving privilege rather than merely transferring property.
It should ask whether the existing distribution is feeding back into the conditions that produced it. Can concentrated wealth purchase coercive power? Can private resources secure public privilege? Can economic dominance become legal dominance?
No state can be made immune to capture. No distribution can guarantee freedom. The objective is continuous contestability: dispersed authority, open entry, independent courts, and direct action against coercive control where it appears.
These questions do not collapse into a single metric. That is a virtue. Moral and political reality is not obligated to become one-dimensional for administrative convenience.
Inequality can indicate that something has gone wrong. Smoke can indicate fire. But smoke may come from a fireplace, a factory, or a fog machine. A measurement that cannot distinguish creation from extraction, abundance from deprivation, or freedom from domination cannot serve as a social objective.
Relieve poverty. Remove coercive privilege. Punish corruption. Preserve equality before the law. Prevent private wealth from becoming unaccountable public power. Defend voluntary cooperation.
Judge the resulting distribution by the conditions that produced and sustain it, not by its shape.