The Conditional Frontier.

---

There is a photograph, taken from orbit, of the Korean peninsula at night. The southern half is a sheet of light. The northern half is black, save for a single pale smear over Pyongyang. The two halves share a language, a thousand years of common history, a climate, a mountain range, and — until 1945 — a single economy. They were separated by a line drawn through an institutional space, and three generations later one of them is among the richest societies on earth and the other cannot keep its lights on. If any single image looks like it settles the argument between capitalism and socialism, it is this one.

This paper argues that the image does not settle the argument, and that understanding why it does not is the beginning of an honest comparison rather than the end of one. The literal question — which system is better — is unanswerable as posed, for two reasons that are usually run together and must be pulled apart. Part of "better" is empirical, a matter of what institutions actually deliver; and part is normative, a matter of which deliveries we have most reason to want. And each term names not a system but a family of systems: the distance between Singapore and Sweden, or between Yugoslav self-management and the Soviet plan, is larger than the distance between several capitalist–socialist pairs. A comparison that does not first decompose the labels will spend its life relitigating which cases were real capitalism and which were real socialism.

The method here is to fix a transparent, multi-criterion evaluation before looking at outcomes, apply it with deliberately symmetric standards to both families, and report not a winner but a conditional performance frontier: which institutional configurations dominate, on which criteria, under which enabling conditions. The contribution is integrative rather than paradigmatic. The empirical findings are honestly asymmetric where the evidence is asymmetric — market coordination decisively out-delivers comprehensive central planning on prosperity and resilience; liberal democracy out-delivers every authoritarian configuration on liberty; and, least congenial to capitalism's confident defenders, coordinated and social-democratic capitalism out-delivers the liberal-market variety on distribution, human development, and subjective wellbeing — while a normative sensitivity analysis shows the global verdict moving with one's values, but only along the liberal-market–to–social-democratic axis. No plausible weighting rehabilitates the command economy. The paper closes by naming the load-bearing assumption that even the winning configuration leaves unexamined — that an economy's reward architecture can be inherited rather than designed — and hands that question to its companion in the Motivation thread, Beyond the Binary, which takes it up.

Keywords: comparative political economy, varieties of capitalism, socialist calculation debate, welfare regimes, institutional configurations, multi-criteria evaluation, natural experiments

---

§ IThe Scoreboard That Cannot Settle

Return to the photograph. What makes the Korean night so rhetorically powerful is exactly what makes it methodologically valuable: it is, as close as history offers, a controlled experiment. Take one people, hold constant their culture and their geography and their inheritance, vary the institutions, and wait fifty years. The South took private property, market prices, and — eventually — liberal democracy; the North took state ownership, central planning, and a hereditary dictatorship. The South's per-capita income is now roughly thirty times the North's (Bank of Korea, 2023). The light is the dependent variable.

And yet the photograph proves less than it appears to, because the North is not "socialism" in the way a fair comparison needs the term to behave. It is one configuration — the most authoritarian, most autarkic, most thoroughly planned configuration the twentieth century produced — and to let it stand for every system to the left of the market is to commit, in advance, the very error that makes this debate interminable. The defender of socialism will say, correctly, that the Pyongyang regime is a dynastic tyranny that no serious socialist would endorse. The defender of capitalism will say, equally correctly, that the failures of Pinochet's Chile or the Gilded Age are not indictments of markets as such. Each side keeps its exemplars and disowns its embarrassments, and the argument runs forever because the units of comparison are slogans, not structures.

It would be comfortable to treat all this as a settled quarrel, a relic of a century that ended in 1991. It is not settled, and the evidence that it is not is generational. A growing share of the young in the liberal-market democracies tell pollsters they regard socialism favourably and capitalism with suspicion; the financial crisis of 2008, a decade of flat real wages, the visible dysfunction of the housing and healthcare markets, and the pandemic's demonstration that the state can in fact mobilise on a vast scale when it chooses have between them reopened a question their parents thought closed. At the same time, the authoritarian-capitalist model presents itself, with some success, as a working rival to liberal democracy as such. The debate is live; the stakes are the institutions of the next half-century; and the quality of public argument about it is low, conducted for the most part in exactly the slogan-against-slogan register this paper is trying to leave behind. That is the case for doing the comparison properly — which is to say slowly, with the criteria fixed in advance and the evidence made to clear the same bar on both sides.

This is the first thing a serious evaluation has to fix. We are not going to compare capitalism with socialism. We are going to compare institutional configurations — specific, namable arrangements of who owns the means of production, how production is coordinated, how the proceeds are distributed, and how political power is held — and we are going to let each configuration's results count against it whether or not its partisans wish to claim it. A case earns its place in the comparison by its structure, not by its reputation. This single move dissolves the no true Scotsman manoeuvre on both sides before it can begin, and it is the reason the Korean photograph, for all its force, is the start of the analysis rather than its conclusion.

The second thing to fix is the smuggling of values. "Which is better" is not one question but a family of questions indexed to what one is measuring. A libertarian who weights economic liberty above all will not reach the same verdict as a Rawlsian who weights the position of the worst-off, and neither is committing an error; they are applying different and legitimate value-weightings to the same facts. The dishonest move — and it is committed across the spectrum — is to hide one weighting inside a confident-sounding empirical claim, so that a contestable judgment about what matters masquerades as a finding about what works. The remedy is not to pretend to neutrality. It is to make the weighting explicit, run the comparison under several of them, and report how the verdict moves. A conclusion that survives only one weighting is a preference; a conclusion that survives all of them is a finding.

What follows from these two fixes is a particular shape of answer. If the unit of analysis is the configuration and the verdict is indexed to values, then the output of an honest comparison cannot be a scoreboard with a single number at the bottom. It must be a map: a statement of which configurations dominate on which criteria, under which conditions, together with an account of how the overall ranking shifts as the value-weighting shifts. We call this map the conditional performance frontier, and constructing it — criterion by criterion, configuration by configuration, weighting by weighting — is the work of this paper.

Three commitments govern that work, and it is worth stating them at the outset because each is a guard against a characteristic failure. First, equal standards are not equal verdicts. Both families face the same evidentiary bar, the same steelman, the same critique; but where the evidence is genuinely asymmetric, the conclusion is permitted to be asymmetric. The manufactured even-handedness that concludes "both have their strengths and weaknesses" is not balance — it is a refusal to report a result, and we treat it as a failure mode rather than a virtue. Second, the unit is the configuration, not the word. We will define a small typology and hold to it, so that "capitalism" and "socialism" never appear in the analysis as undefined primitives. Third, values are shown, not assumed. The verdict is reported as a function of the weighting vector, and the reader who disagrees with our weightings can read their own verdict off the same frontier.

A note on what this paper is and is not. It is the empirical, evaluative member of a pair. Its companion in this archive's Motivation thread, Beyond the Binary (Cahill, 2026a), argues that capitalism and socialism share a deeper error than either's failures — that both misread human motivation, and that the institutions of the next economy will have to be designed around motivation rather than against it. That is a claim about what to build. The present paper is upstream of it: a claim about what the evidence on the existing configurations actually shows, and therefore about which direction the building should start from. The two papers can be read in either order, but they were written to be read together, and §9 hands off explicitly from the one to the other. The discipline throughout is the discipline the archive imposes on itself: name the load-bearing assumption the dominant framing will not, and then say something specific enough to be wrong about.

§ IIConfigurations, Not Camps

The words "capitalism" and "socialism" are best treated not as the names of two systems but as labels for two regions of a larger space — a space whose dimensions we can specify, and within which real economies can be located by their structure rather than by their flag. Four dimensions do most of the work.

The first is ownership of the means of production: private, state, cooperative, or some mixture. The second is the coordination mechanism — how the millions of daily decisions about what to make and for whom get reconciled. The candidates are the market price system, the central plan, and negotiated coordination among organised actors (employer associations, unions, sectoral bodies). The third is distribution and decommodification: the degree to which a person's livelihood tracks their market position, as against being secured by a social wage or an egalitarian rule. The term of art is Esping-Andersen's (1990): decommodification, "the degree to which individuals can maintain a livelihood without reliance on the market." The fourth is political form: liberal-democratic, authoritarian, or participatory.

The dimensions are partly independent, and their independence is precisely what the labels obscure. Markets can be yoked to private ownership (the United States), to extensive state ownership and direction (China after 1978; Naughton, 2018), or — at least on paper — to social ownership, which is the entire content of the market-socialist tradition from Lange (1936) to Roemer (1994). Decommodification can run high under unambiguously capitalist ownership (the Nordic social democracies) and low under it (the Anglo-American economies). Hold the dimensions apart and the space comes into focus; collapse them into a binary and the most interesting cases vanish into the seam.

Of the four dimensions, decommodification is the one whose independence is least intuitive and most consequential, because it is the axis on which the Nordic economies are perpetually misfiled. A country can leave the ownership of capital almost entirely private, leave allocation almost entirely to the market, and still decommodify a vast share of life — health, education, childcare, old age, the income floor beneath everyone — by funding it collectively and providing it as a right rather than as a purchase. That is the Nordic move, and it is why those countries appear, to a casual eye, to straddle the line between the systems: they are capitalist on ownership and on coordination, and social-democratic on distribution and decommodification. The binary has no cell for this combination, which is exactly why the binary keeps mishandling the most instructive case on the board — filing Sweden as a kind of socialism by readers on the right and as a kind of socialism by readers on the left, when it is neither. The typology has a cell for it. As it turns out, a great deal of this paper's central finding lives in that cell. Placing real systems into this space yields five reference configurations, and these five recur through the rest of the paper:

Liberal-market capitalism (LMC) — private ownership, market coordination, market-determined distribution, liberal democracy. The United States and the United Kingdom are the exemplars; in the varieties-of-capitalism scheme they are the liberal market economies, coordinating chiefly through competition and arm's-length contract (Hall & Soskice, 2001).

Coordinated and social-democratic capitalism (CSC) — predominantly private ownership, but with negotiated coordination and high decommodification through a large welfare state. The Nordic countries and Germany are the exemplars. This configuration pairs Hall and Soskice's coordinated market economy with Esping-Andersen's social-democratic welfare regime, and — as will become important — it is a market economy, not a socialist one. The Nordic states are among the most economically free on earth; what they socialise is a large share of consumption and risk, not the ownership of capital.

State (political) capitalism (SC) — mixed ownership with heavy state direction of credit and strategic sectors, market coordination across much of the economy, under authoritarian political control. China since 1978 and Singapore are the exemplars. Milanovic (2019) calls this political capitalism; Nölke and colleagues (2019), studying its emerging-economy variant, call it state-permeated capitalism. The crucial point for our purposes is that it is a species of capitalism, not of socialism: its dynamism arrived with the market, not with the plan.

Command socialism (CS) — comprehensive state ownership, central planning, egalitarian-indexed distribution, authoritarian political form. The Soviet Union, the German Democratic Republic, North Korea, Maoist China, and Cuba are the exemplars. Kornai's (1992) anatomy of "the classical socialist system" remains the canonical description of how its parts fit together — and, as §5 argues, of why they fit together as they do.

Market socialism and worker self-management (MS) — social or cooperative ownership combined with market coordination. Yugoslav self-management, the Mondragón cooperatives at the scale of a single firm, and the Lange–Lerner model as a theoretical limit are the reference points. This is the configuration the evidence can say least about, because — unlike the other four — it has barely been run at national scale, and §7 marks its cells in the frontier as under-identified rather than dominated for exactly that reason.

One feature of the space deserves emphasis before we leave it: real economies move through it, and the same country can occupy different regions in different decades, which is a further reason the camp-labels mislead. China travelled from command socialism to state capitalism without a change of ruling party. Britain shifted from its coordinated postwar settlement toward the liberal-market pole across the Thatcher decade; Sweden made the opposite journey, toward social democracy, across its mid-century one, and then trimmed back from it. The "third way" debates of the 1990s were arguments about which coordinates to occupy, conducted largely by people who agreed the market would do the allocating and differed over how much to decommodify. To fix a country to a single eternal label is to mistake a trajectory for a point. The configurations are positions in a space that economies traverse, and what follows is a comparison of positions — scored wherever and whenever a real economy actually sat in one long enough to read its results.

Two payoffs follow immediately from coding by structure rather than by camp. The first is that the persistent empirical finding of the comparative-capitalisms literature — that national institutional configurations remain stubbornly diverse and do not converge on a single best model even under decades of globalising pressure (Wood & Allen, 2020) — tells us in advance that a single global ranking is the wrong shape of answer. The differences among capitalisms are not noise around a mean; they are the signal. The second payoff is that the typology disarms both of the symmetrical evasions that keep this debate alive. To the claim that "that wasn't real socialism" and to the claim that "that was merely crony capitalism," the same reply now applies: a case is admitted by where it sits on the four dimensions, and once admitted, its results are allowed to count. The Soviet Union is command socialism whether or not today's socialists wish to own it; the United States is liberal-market capitalism whether or not today's libertarians wish to disown its inequalities. Structure first, reputation never.

§ IIIThe Same Standards

If the configurations are the rows of our map, the criteria are its columns, and the integrity of the whole exercise depends on fixing those columns before we look at the cells. This is the pre-registration discipline — standard in clinical research, increasingly standard in experimental social science, and almost unheard-of in comparative political economy, where the temptation to choose the criteria that flatter one's prior conclusion is overwhelming precisely because it is so easy to do unnoticed. We commit, in advance, to seven criteria.

Material prosperity and dynamism (C1): output per head, long-run growth, productivity, and the rate of innovation — with chronic shortage and the soft-budget pathology (Kornai, 1986) as the negative pole. Distribution (C2): inequality, measured by the Gini coefficient and by the income shares of the top tenth and bottom half, together with poverty. Human development and capabilities (C3): life expectancy, infant mortality, education, and the Human Development Index, read through the lens of Sen's (1985, 1999) capability approach, on which development is the expansion of what people are actually able to be and do. Subjective wellbeing (C4): the population's own evaluation of its life, via the Cantril ladder. Freedom (C5), kept deliberately as two axes that do not move together: political and civil liberty (C5a) and economic freedom (C5b). Resilience and stability (C6): the volatility of output and the depth of, and recovery from, crises. Ecological sustainability (C7): emissions intensity and resource throughput measured against planetary boundaries.

Seven is a chosen number, and the choice is a compromise we defend rather than hide. Fewer criteria would look selective — an evaluation on three axes can be gerrymandered to almost any conclusion. More would dilute, burying the load-bearing comparisons under a litter of weakly-measured ones. Seven is enough to be comprehensive and few enough to be legible, and the particular seven are chosen to span the value-systems that actually contend in this debate, so that no major tradition can complain its central concern was left off the page.

It is worth dwelling on why fixing the criteria in advance matters so much, because the discipline is not a ritual imported for its own sake. The characteristic way these comparisons are rigged is not by lying about the numbers; it is by choosing, after the fact and usually without quite noticing, the criteria on which one's preferred system happens to win, and quietly declining to tabulate the ones on which it loses. The libertarian reaches for growth and economic freedom; the socialist reaches for inequality and health; each assembles an honest-looking dossier out of real statistics, and the two dossiers never meet, because they were built to different specifications chosen to flatter different priors. Pre-registration forecloses this. By writing down the seven criteria, the metrics, and the case-inclusion rules before the outcome data are examined — and, in this paper's preparation, by gathering the framework and the performance figures in two separate passes — one removes the degree of freedom through which the result is usually fixed. The reader does not have to take on trust that we resisted the temptation to cherry-pick the criteria; the structure was built to make that particular move visible if it occurred.

Two honesties keep this claim from overreaching, because pre-registration buys less than its enthusiasts sometimes sell. The first is that fixing the criteria in advance does not fix their operationalisation, and the operationalisation is where most of the residual discretion lives — which index, which year, which exemplar. We chose disposable-income rather than market-income Gini; we read political liberty off V-Dem and Freedom House while reporting the disagreement between the two economic-freedom indices as a finding; and the Cuban infant-mortality figure, as §6.3 will show, swings sharply with the year one selects. None of that was pre-committed at the level of the metric, and a determined skeptic can argue the cherry-picking simply moved one storey down. We have tried to disarm the charge case by case — reporting ranges, flagging the contested numbers, preferring a multi-year pattern to a convenient single year — but the guarantee is weaker than "unavailable by construction," and we would rather say so than imply a rigour the procedure does not deliver. The second honesty is deeper, and §10 returns to it: the seven criteria themselves encode a broadly liberal value-frame — they ask after prosperity, distribution, capability, freedom, and sustainability, the goods a liberal already counts — so "the same standards" means the same liberal-ish standards applied evenly to both families, not a view from nowhere. A Marxian who wanted exploitation and alienation on the scorecard, or a communitarian who wanted solidarity, would say it was tilted before the first number was entered. That objection cannot be fully answered, only conceded: the evenhandedness on offer here is evenhandedness within a value-frame the paper did not itself derive.

That last point is the crux, because the criteria are not value-neutral and we do not pretend they are. A libertarian cares most about C5; a Rawlsian about C2 and C3; an ecological economist about C7. To privilege any one weighting would be to do the smuggling we have promised not to do. So instead of one weighting we use six, and we report the verdict under each. The efficiency/growth vector weights C1 heavily and is anchored in the classical-liberal tradition (Friedman, 1962). The egalitarian vector weights C2 and C3, after Rawls (1971) and Cohen (2009). The capabilities vector weights C3 and C4, after Sen (1999) and Nussbaum (2011). The sustainability vector weights C7 and discounts raw growth. The liberty vector weights both axes of C5, after Nozick (1974). And an equal-weight baseline averages all seven — offered for reference, and explicitly not privileged as the neutral truth, because there is no neutral truth here, only an honest accounting of how the answer depends on what one is asking. The machinery of varying weights and observing how a multi-criteria ranking responds is standard decision analysis (Greco, Ehrgott, & Figueira, 2016); what is unusual is applying it to this question with the weightings shown rather than concealed.

The scoring is ordinal, not cardinal, and this too is a discipline rather than a limitation. On each criterion a configuration is ranked dominant, intermediate, or dominated, with the supporting figure and its source attached, rather than forced onto a single numerical index whose weights would then do the arguing in disguise. A composite score that announces "capitalism: 7.4, socialism: 5.1" has not measured anything; it has hidden a hundred contestable judgments inside two decimal points. The map we are drawing has no such number at the bottom of it, and its refusal to produce one is the point.

§ IVReading the Natural Experiments

The central problem of inference in this field is confounding. Rich countries differ from poor ones in geography, in resource endowment, in demographic structure, in external environment, and in the quality of the institutions they inherited (North, 1990; Acemoglu, Johnson, & Robinson, 2001; Acemoglu & Robinson, 2012) — and any of these can wear the costume of a "system effect." A naïve comparison that found capitalist countries richer than socialist ones would have established almost nothing, because the capitalist countries were, on average, richer to begin with and better-placed in a dozen ways that have nothing to do with their ownership structure. The entire methodological burden of an honest comparison is to strip the confounders away and report only what survives.

An example shows how easily the naïve comparison goes wrong, and in both directions. Suppose one observed that oil-rich Norway is both heavily decommodified and very rich, and concluded that the welfare state causes the wealth. The inference would be unsafe, because Norway's wealth has an obvious alternative source under the North Sea, and the resource endowment is quietly taking credit the welfare state is being given. Now suppose one observed that the command economies were concentrated among societies that were poor and agrarian at the moment they adopted the plan, and concluded — as several of capitalism's defenders implicitly do — that the plan made them poor. That inference is unsafe too, because the causation may run the other way: it was the poverty that made these societies receptive to a revolutionary promise in the first place. Every comparison in this paper is exposed to errors of this shape, and the defence is always the same — find a design in which the confounder is held fixed, and trust the comparison only to the degree that it is. It is why the divided nations do so much of the work here, and why a cross-country scatter-plot does almost none.

The strongest tool for this is the divided-nation natural experiment, and it is why this paper began with a photograph. When a single nation — one culture, one language, one history, one resource base — is partitioned by system, the confounders are held very nearly fixed, and the divergence that follows can be attributed to the institutions with a confidence no cross-country regression can match. East and West Germany, North and South Korea, and mainland China against Taiwan and Hong Kong are the three great cases, and we lean on them heavily, while remaining candid that there are only three and that one of them — China against Taiwan — is complicated by scale and by the external support each side received. A second tool is the transition panel: the post-1989 collapse of the Eastern Bloc created a natural before-and-after within dozens of countries, and the variation in how the transition was sequenced (shock therapy against gradualism) is itself informative. A third is the matched cross-country comparison, holding development level and region roughly constant — the Nordic economies against the Anglo-American ones; Cuba against its Latin-American peers. And a fourth, easily overlooked, is the voluntary micro-scale case: the worker cooperative, the kibbutz, the Mondragón federation. These isolate the ownership mechanism in its purest form — collective ownership without state coercion and without the suppression of prices — and so they test the part of the socialist proposition that the command economies confounded beyond recovery.

A word on the discipline behind the numbers, because it bears on how much weight the reader should place on them. The case universe in this paper was fixed by structural rules — the four admission criteria above — before any outcome data were gathered, and the data were collected in two streams deliberately separated in time: the framework and theory first, the performance figures only afterward. The point of the separation is to prevent the case list from drifting, even unconsciously, toward the cases that would flatter a conclusion — the precise rigging each camp has always accused the other of. And because parts of this work were assembled with the help of a large language model, which will fabricate a plausible citation or invent a confident statistic without the slightest signal that it has done so, every figure and source in what follows was logged against a resolvable identifier and then independently re-verified against the primary source; the handful of numbers that did not survive that audit were corrected or struck. The reader is entitled to assume that every statistic below traces to a real source that says what we claim it says, because that assumption was made to hold by construction rather than by trust.

This is not a fashionable caution to nod at and move past; it changes what the reader is being asked to believe. In a domain this ideologically charged, the usual reason to distrust a comparison is the suspicion that its author went looking for the numbers that suited them — and the usual defence, trust me, I was even-handed, is worth precisely nothing, because everyone offers it and the partisans offer it most loudly. The two-stream rule and the source ledger replace that worthless assurance with a procedure: the criteria and the case list were fixed before the outcomes were seen; the outcomes were logged against resolvable sources; and the load-bearing figures were re-checked against the originals by a separate pass whose only assignment was to find errors — which it did, and which were corrected before this draft. None of that machinery makes the paper's judgments correct. It makes the paper's facts auditable, which is the half a reader cannot check for themselves, and therefore the half that has to be built to be trusted rather than asserted to be trustworthy.

§ VThe Calculation Debate, Treated as Live

No comparison of these two families can decline the one theoretical question on which the whole quarrel was once thought to turn, and which a surprising amount of contemporary writing waves at rather than engages: can a socialist economy calculate? The question is not rhetorical, and it is not settled, and pretending otherwise in either direction is a tell that the writer has a side rather than an argument.

The case was opened by Mises (1920/1935), who argued that rational economic calculation is strictly impossible under full socialism. His reasoning was not about incentives or about bureaucratic sloth; it was about information. Production requires choosing among indefinitely many technically feasible ways of combining capital goods, and choosing rationally requires knowing their relative scarcities — which is to say, their prices. But prices for capital goods emerge only from the exchange of privately owned capital goods. Abolish private ownership of the means of production and you abolish the market for them, and with it the prices, and with them the only available measure of whether one production plan is more wasteful than another. The planner is left, in Mises's image, "groping in the dark." Hayek (1945) deepened the argument into its mature form. The knowledge a planner would need is not merely large but dispersed — it exists only as the particular, fleeting, often inarticulate awareness that millions of individuals have of their own local circumstances, the knowledge "of the particular circumstances of time and place," and it cannot be gathered into any central organ because much of it is never even rendered into words. The price system, on Hayek's account, is not a device for allocating known resources; it is a communication mechanism that lets a society act on knowledge no one of its members possesses in full.

The socialist reply, and it is a serious one, came from Lange (1936, 1937) and Lerner (1944). Grant Mises and Hayek their premise that prices carry indispensable information; it does not follow, said Lange, that the prices must come from private ownership. A central planning board could simulate the market: set provisional prices, observe the resulting surpluses and shortages, adjust the prices up where goods are scarce and down where they glut, and iterate toward the same equilibrium a market would reach — managers being instructed, by Lerner's rule, to produce where price equals marginal cost. The market, on this view, is one algorithm for solving the allocation problem; the planning board running a tâtonnement is another, and there is no reason in principle the second cannot reach the same answer while social rather than private ownership reaps the returns. Roemer (1994) later gave the tradition its most sophisticated modern form, a coupon-based market socialism in which capital markets are retained but the right to the profit stream is distributed equally rather than concentrated.

So the theory reaches a genuine impasse, and one must turn to the evidence — which is where Kornai (1980, 1986, 1992) becomes indispensable, because he studied the actually-existing socialist economies not as a partisan but as an anatomist. What he found was not that the planners were stupid or lazy but that the system generated chronic shortage as a structural property, independent of the competence of any individual within it. The mechanism he named is the soft budget constraint: because a state enterprise that runs into trouble expects to be rescued rather than allowed to fail, the discipline that a hard budget imposes on a market firm is absent, and the enterprise rationally hoards inputs, pads its requests, and ignores cost. Shortage is not a failure of the plan; it is the equilibrium the plan's incentive structure produces. This is the empirical residue of the calculation debate, and it weighs heavily against the Lange–Lerner reply — not because the tâtonnement is incoherent in theory but because the institutions that would have to run it could not, in practice, generate the hard constraints and the truthful signals the simulation requires.

There is, moreover, a political corollary to the informational argument, and the realised record bears it out with uncomfortable consistency. If rational planning requires assembling dispersed knowledge into a single centre, then comprehensive planning requires concentrating not merely information but decision — the authority to direct resources — in that same centre, and a concentration of economic power on that scale has a way of fusing with the concentration of political power. This was the burden of Hayek's broader argument, and whatever one makes of its more sweeping formulations, the pattern is hard to wave away: every economy that abolished the market and the price system in favour of comprehensive planning also, in the event, abolished political liberty. The correlation is not a proof of necessity — the data in §6.5 cannot rule out a planned economy that stayed free — but it is striking that the experiment of comprehensive planning under democracy, like the experiment of market socialism at scale, is one that history never actually ran. The configurations that suppressed prices suppressed votes; whether the two suppressions are causally linked or merely kept company, an honest comparison has to record that they always travelled together.

Intellectual honesty requires three concessions that keep the question live rather than closing it with a verdict the evidence does not fully support. The first is that the case is sometimes overstated: the claim that the Austrians simply won, that the matter is closed, is itself a position (Lavoie, 1985) rather than a neutral summary, and it is contested by serious people. The second is that a computational revival has reopened the feasibility question on new ground. Cockshott and Cottrell (1993), and more recently Dapprich and Cockshott (2023), argue that modern computing and machine-learning forecasting change the arithmetic — that the in-natura calculation Mises declared impossible is now tractable, and that the dispersed-knowledge problem can be attacked with feedback control rather than with prices. The third concession runs the other way: the most careful recent answer to the revival (Lambert & Fegley, 2023) contends that the new computing power misses Mises's actual point, which was never that the calculation was arithmetically too hard but that, without genuine markets in capital goods, there are no true prices to compute with — no amount of processing power generates the data that private exchange alone produces. The one historically serious attempt to build a real-time cybernetic planning system, Allende's Project Cybersyn, is instructive precisely because its reach so far exceeded its grasp (Medina, 2011); it is a monument to the ambition and to the gap between the ambition and the wiring.

What would it take to settle the question the twentieth century left open? Not, on the evidence, more computing power alone — the most careful critics are right that processing speed was never the binding constraint (Lambert & Fegley, 2023). It would take an economy that retained genuine markets in consumer goods and in labour, and therefore genuine prices, while socialising the ownership of capital and the profit stream — Roemer's coupon economy, or something close to it — run at national scale, under a democracy, and for long enough to see whether the investment function can be disciplined without a private capital market to price it. That experiment has not been run. The command economies suppressed the prices and so cannot speak to it; the cooperatives ran inside capitalist economies and so tested ownership but not the macro-architecture; Yugoslavia came nearest and was confounded first by its own illiquidity and then by war. The market-socialist proposition is, in the strict sense, not refuted but unfalsified-because-untried — and a field that prides itself on evidence ought to find that gap interesting rather than convenient.

We therefore record the calculation debate not as a punchline but as a live question with a heavily-weighted body of evidence on one side. The realised command economies failed at dynamism and at the absorption of shocks in exactly the pattern the Austrian argument predicts, and that pattern, documented in §6, is the strongest single thing the twentieth century has to say on the matter. But the market-socialist counterfactual — markets retained, ownership socialised — has not in fact been run at national scale, and the computational revival has not been tested at the technological frontier, and an honest map marks those as open territory rather than conquered ground.

§ VIWhat the Evidence Says, Criterion by Criterion

We now apply the seven criteria, one at a time, with the same standard turned on each configuration. The reader will notice that the verdicts do not all run the same way — that capitalism's defenders and socialism's are each vindicated on some criteria and rebuked on others — and that this is not a contrived balance but the actual shape of the evidence. Where a result is asymmetric we report it as asymmetric; where it is contested we report the contest.

§6.1 Material prosperity and dynamism (C1)

The evidence here is as close to decisive as this field offers, and it runs one way. Begin with the divided nations, because they control the confounders the cross-country data cannot. At reunification, East German output had not merely lagged the West's; it had collapsed, falling by more than forty per cent between 1989 and 1991 — a deeper contraction than the United States suffered in the entire Great Depression — and even after a decade of vast transfers, East German labour productivity converged only to about fifty-five per cent of the Western level before stalling there (Sinn, 2000). Korea is starker still. The Bank of Korea's estimates put North Korean income per head at roughly one-thirtieth of the South's (Bank of Korea, 2023), and the divergence is not only material: at the depth of the 1990s famine, North Korean male life expectancy fell by about seven years in a single year, and the most credible demographic work puts famine-related excess deaths between six hundred thousand and a million (Choi et al., 2023; Goodkind & West, 2001). Two halves of one nation, one of them ablaze and one of them dark.

The Chinese world ran its own divided-nation experiment, less photogenic than Korea's but pointing the same way. Taiwan and Hong Kong — the Chinese societies that took markets and private property — reached rich-world living standards across precisely the decades in which Maoist central planning held the mainland near subsistence, and the mainland's own ascent began only when, after 1978, it abandoned the plan for the model its market-Chinese neighbours had already proven. The transition record then makes the point a third time, from the inside. China's pivot from the plan toward market coordination after 1978 coincided with average growth above nine per cent a year and a fall of close to eight hundred million in the number of its people below the international extreme-poverty line — the largest and fastest escape from poverty in recorded history (World Bank, 2022, 2024). And the decisive detail, for the calculation debate, is the timing: the gains followed the introduction of markets and prices, not any improvement in planning. China's record is a vindication of market coordination operating within an authoritarian state-capitalist shell (Naughton, 2018; Milanovic, 2019), not of the plan it replaced.

This is the empirical residue of §5. The pattern — command economies able to mobilise resources for a few prioritised objectives but unable to sustain economy-wide dynamism — is exactly what Kornai's (1986) account of shortage and the soft budget constraint predicts, and it is too large and too consistent across independent cases to be explained away by geography or starting conditions. On C1, the market-coordinated configurations — liberal-market, coordinated, and state-capitalist alike — dominate command socialism, and the live question is no longer whether markets out-produce comprehensive planning but how much of the intra-capitalist gap is real. On that narrower question the United States leads the Nordics on headline output per head — about $85,810 against Sweden's $71,031 in 2023, at purchasing-power parity (World Bank) — though the gap narrows once one adjusts for hours worked and for distribution, and though the comparison should distinguish frontier innovation from catch-up. The liberal-market economies, the United States above all, sit at the technological frontier — the research universities, the venture finance, the firms that produced the digital and biotechnological revolutions — while the coordinated economies have more often been fast and efficient adopters and incremental improvers. That is a genuine liberal-market advantage, and headline income only partly captures it.

It is worth being precise about what that advantage does and does not establish, because it is the liberal-market case's strongest card and is routinely overplayed. It establishes that the institutions of the liberal-market economy — deep and risk-tolerant capital markets, a culture of firm entry and exit, research universities coupled to private finance — are unusually good at producing the radical, uninsurable innovations that push the technological frontier outward. It does not establish that those same institutions are good at distributing the gains from that innovation, which the next criterion shows they are not; nor that the frontier-pushing must be bundled with the liberal-market configuration's thin safety net rather than the coordinated configuration's thick one. The interesting and under-asked question is whether a society can keep the innovation engine while decommodifying the downside — whether the dynamism is a property of the whole liberal-market package, or only of some separable part of it that a coordinated economy could in principle bolt on. The synthesis of §9 bets, tentatively and with its reasons, on the latter.

§6.2 Distribution (C2)

Here the asymmetry reverses, and it is the part of the record that capitalism's confident defenders most often decline to concede. Within the capitalist family, the coordinated and social-democratic configurations deliver markedly more equal distributions than the liberal-market one — at the same level of development, and with no loss of liberal-democratic standing. The United States' disposable-income Gini stood at 0.39 in 2023, against 0.26 to 0.29 across the Nordic economies — Norway 0.26, Finland 0.27, Denmark 0.28, Sweden 0.29 — a gap of eleven to thirteen Gini points (OECD, 2025). The pre-tax distribution diverges further: the top tenth takes about forty-seven per cent of US national income, with the bottom half left on thirteen, against roughly twenty-nine to thirty per cent for the top tenth in Norway and Sweden (World Inequality Lab, 2025). The long-run drift has a candidate engine in the tendency for the return on capital to outpace growth — Piketty's (2014) r > g — though we lean on it only lightly, since whether r > g is a genuine mechanism or merely a compact description of one is contested in the subsequent literature, and the inequality figures stand on their own without it. The cross-sectional pattern is firmer than any mechanism: the societies with the widest income gaps tend also to be the ones in which a child's eventual position is most tightly predicted by the parents' — the correlation sometimes called the Great Gatsby curve — and while the direction of causation there is debated, the liberal-market configuration's higher inequality at least travels with lower measured mobility, which sits awkwardly against the "land of opportunity" self-image.

Command socialism did compress measured money incomes — but the comparison is contaminated past the point of usefulness by suppressed prices, by chronic shortage, and by the substitution of political privilege for monetary inequality, so that the nomenklatura's access to goods unavailable at any price to ordinary citizens never appears in a Gini coefficient. On C2, coordinated and social-democratic capitalism dominates the liberal-market variety; command socialism's nominal equality does not survive correction for shortage and non-monetary privilege; and the egalitarian achievement that does survive scrutiny is the social-democratic welfare state operating inside a market economy, not the abolition of the market. That last clause matters for everything downstream, and we return to it.

A caution on identification belongs here, because it governs the next three criteria as much as this one. The verdict that coordinated capitalism beats the liberal-market variety on distribution rests not on a divided-nation experiment but on matched cross-country comparison — Nordic against Anglo-American — which is the softest tool in §4's kit: a handful of rich democracies that differ in scale, in population homogeneity, in social trust, and (for Norway) in oil, as well as in the size of their welfare states. We therefore hold the distributional verdict, and the developmental and wellbeing verdicts that follow it, with less confidence than the prosperity verdict of §6.1, whose divided-nation design is far cleaner. What raises our confidence above the level a single soft comparison would warrant is corroboration from better-identified evidence pointing the same way: the within-country shifts as Britain moved toward the liberal-market pole and Sweden toward the social-democratic one, and the 2008 and 2020 episodes of §6.6, where the coordinated architecture's effect on employment is visible inside a single shock rather than across dissimilar countries. The matched comparison is suggestive; it is the triangulation that makes it credible.

§6.3 Human development and capabilities (C3)

Read through Sen's (1999) capability lens, output is instrumental and the real question is what a society's people are able to be and do. On the Human Development Index the social democracies again lead: Norway at 0.970, Denmark at 0.962, Sweden and Germany at 0.959, ahead of the United Kingdom at 0.946 and the United States at 0.938 — seventeenth in the world — on 2023 data (UNDP, 2025). The analytically richest case, though, is Cuba, because it is the cleanest test of whether socialist public provision can buy capabilities at low income. It can, in part: Cuba reaches an HDI of 0.762 — "high human development" — on income per head of around $8,415, roughly a tenth of the US level, and its life expectancy of about 77.6 years essentially matches the United States' 77.4 on the same series (UNDP, 2025; World Bank). That is a real and under-acknowledged achievement of prioritised spending on health and education.

But it must be reported as a bundle and not as a cherry, and the infant-mortality figure in particular invites exactly the cherry-picking this paper warns against. Cuba's official rate has often run at or below the United States' — about 4.0 per thousand in 2024 against the US's 5.6 — but it is volatile and contested: it spiked above the US level during the 2022–23 economic crisis, reaching 6.6 in 2023, so a single year can be chosen to flatter either side, and a peer-reviewed analysis argues the official Cuban figure is deflated outright, by reclassifying early neonatal deaths as late fetal deaths — which, corrected, would raise the rate from a reported 5.8 to somewhere between 7.5 and 11.2 per thousand, up to nearly double (Berdine, Geloso, & Powell, 2018). The honest reading takes neither the convenient year nor the official statistic at face value: Cuba's infant-survival record is genuinely good for its income and genuinely uncertain in its books. Its measured income, meanwhile, is a fraction of its regional peers'; and — as §6.5 documents — its capability gains are purchased alongside the suppression of the very freedoms that the capability approach counts not as luxuries but as constituents of development. A life expectancy of seventy-eight years inside a closed political order is a genuine good achieved at a genuine cost, and Sen's framework, properly applied, registers both. Two confounders cut in opposite directions and are worth naming. The decades-long US embargo depressed Cuban incomes in ways not attributable to the socialist model, and a fair accounting sets that aside; but at the same time, the concentration of a poor economy's resources onto two highly visible achievements — a doctor on every block, near-universal literacy — is exactly the kind of prioritised provision a command system can deliver, precisely because it is not attempting the thousand other things a market quietly does. The Cuban lesson, read carefully, is narrow and real: a determined state can buy specific capabilities at low income, and it pays for them in liberty, in dynamism, and in the uncounted goods that are never made. On C3, coordinated and social-democratic capitalism leads; the Cuban case shows targeted public provision can lift health and education at low income, but it does not generalise into a command-economy advantage once the full capability set — agency and liberty included — is scored.

§6.4 Subjective wellbeing (C4)

The evidence on how people evaluate their own lives tracks the distributional and capability findings closely. In the World Happiness Report 2025, the Nordic social democracies hold the top places — Finland first at 7.736 on the Cantril ladder, for the eighth consecutive year, with Denmark and Iceland next — while the United States has fallen to twenty-fourth, its lowest position on record (Helliwell et al., 2025). High average income, it turns out, is compatible with middling reported wellbeing; the coordinated configuration's combination of security, low inequality, and high social trust appears to do better on a margin that aggregate output does not capture. On C4, coordinated and social-democratic capitalism dominates the liberal-market variety; the command and state-capitalist cases are measured more sparsely and less comparably — the survey does cover China, where the ladder score is middling rather than absent, but the socialist cases are thin and the data politically mediated — so they are scored intermediate-to-dominated with lower confidence, an honest gap in the evidence rather than a finding.

§6.5 Freedom, on two axes (C5)

Keeping the two axes of freedom apart is essential, because they genuinely do not move together, and most of the confusion in popular argument about this comes from collapsing them. On political and civil liberty (C5a), the liberal democracies dominate every authoritarian configuration without a single exception. Freedom House's 2025 scores place Finland at 100, Sweden and Norway at 99, and the United States at 81, against 9 apiece for China and Cuba; V-Dem's Liberal Democracy Index tells the same story in finer grain, with 2025 values around 0.88 for Denmark, 0.85 for Sweden and Norway, and 0.04 and 0.06 for China and Cuba at the floor — while the United States, far above the autocracies but well below the Nordics, fell sharply in the most recent coding, from 0.75 in 2024 to 0.57 in 2025 (Freedom House, 2025; V-Dem, 2026). This is the decisive count against command socialism and authoritarian state capitalism alike: whatever else they achieved, they were built on the suppression of the freedoms that the capability tradition treats as ends in themselves, and no quantity of health clinics buys that back.

On economic freedom (C5b) the ranking shifts, but it stays within the capitalist family. The two leading indices disagree in a way that is itself instructive. The Heritage Index for 2025 places Singapore first at 84.1, the Nordics in its "mostly free" band (Denmark 79.1, Norway 78.3, Sweden 77.9, Finland 77.0), and the United States down at 70.2 — twenty-sixth — with Cuba at 25.4 and North Korea at 3.0 at the bottom (Heritage Foundation, 2025). But Fraser's measure, on 2023 data, ranks the United States fifth at 8.10 while high-tax Sweden and Norway fall to thirty-fifth and forty-eighth on its size-of-government component — the reverse of Heritage's ordering of the same countries (Fraser Institute, 2025). The disagreement is not noise; it is a disclosure of what each index means by "economic freedom," one penalising a large public sector and the other crediting a country that maintains open trade and sound money despite one. We report both rather than choosing, because the divergence is part of the finding. On C5a, the liberal-democratic configurations dominate absolutely; on C5b, all the capitalist configurations dominate the socialist ones, while the two market families trade places depending on whether a large welfare state is scored as a subtraction from freedom or as something orthogonal to it.

§6.6 Resilience and stability (C6)

The post-communist transitions are the sharpest stress test on record, and they distinguish sequencing as much as system. Output fell everywhere when the plan was dismantled, but the depth and duration varied enormously. Russia's recorded output fell on the order of forty to forty-five per cent across 1989–98, and Ukraine's by roughly half or more, while Poland — after a sharp initial contraction in 1990, variously estimated between seven and twelve per cent depending on the series — returned to growth within two years and stood about twenty per cent above its 1989 level by 1999 (IMF, 2014; Fischer, Sahay, & Végh, 1992). The human cost of the worst-managed transitions was severe: Stuckler, King, and McKee (2009) find rapid mass privatisation associated with a 12.8 per cent rise in adult male mortality, an effect muted where membership in social organisations was high. We flag that as an association whose causal reading was challenged in the subsequent correspondence, and we note what it actually indicts: a particular transition policy — shock privatisation without institutional cushioning — rather than market coordination as such.

The same lesson recurs in the rich democracies' two most recent great shocks. In the financial crisis of 2008, and again in the pandemic of 2020, the configurations that carried strong automatic stabilisers and short-time-work institutions — Germany's Kurzarbeit is the cleanest example — held employment far more steadily through the trough than the liberal-market economies, where the adjustment ran through layoffs and the recovery through a slow reabsorption of the unemployed. The coordinated economies did not escape the shocks; they distributed them differently, socialising through the welfare state the risk that the liberal-market configuration leaves sitting with the individual worker. Whether that distribution is a virtue depends on one's weighting, but that the employment relationship is more resilient through a downturn under the coordinated architecture is not in serious dispute — and it is a property of the welfare-state layer, not of the market mechanism the two families hold in common. State capitalism's own resilience is harder to read, and we score it with low confidence: China kept growing through the 2008 crisis on the back of an enormous state-directed stimulus, which is a kind of resilience but a fiscally and financially expensive one, whose bill — local-government debt, overbuilt capacity — came due later. The intermediate mark in its cell carries that caveat, and the evidence behind it is thinner than for the democratic cases. On C6, the evidence rewards the configuration that pairs market coordination with strong automatic stabilisers — coordinated capitalism — over both unbuffered liberalisation and the plan; resilience turns out to be less a capitalism-versus-socialism axis than a coordinated-versus-unbuffered one.

§6.7 Ecological sustainability (C7)

Sustainability is the criterion on which no configuration can claim a win, and saying so plainly is part of holding the standard. Among rich economies the coordinated configurations have the lower emissions intensity — territorial carbon dioxide per head in 2023 ran about 3.43 tonnes in Sweden and 4.56 in Denmark, against 13.83 in the United States and 9.24 in China (Crippa et al., 2024) — but even the best performers sit at or above the world average of 4.86 tonnes, which is itself far above any sustainable level, and two caveats sharpen the picture. Consumption-based accounting, which charges each economy for the emissions embodied in what it imports, erodes part of the Nordic lead, because trade-exposed, service-weighted rich economies offshore their industrial emissions. And on an ecological-economics reading, sustainability is not one tradeable criterion among seven but a binding constraint: if the planetary boundary is hard, a configuration that out-performs on the other six while breaching this one has not won but borrowed against a future that will foreclose. Command socialism's record here was worse, not better — the Soviet bloc was notoriously resource- and pollution-intensive, consistent with the same soft-budget incentive to hoard inputs that produced its shortages (Kornai, 1986). On C7, coordinated capitalism leads on emissions intensity, but the honest finding is a frontier gap rather than a victory: no realised configuration is yet biophysically viable at high-income consumption, and this is a failure the two traditions share.

§6.8 The cooperative at scale, and the case the command economies buried

The seven criteria, applied across the four well-documented configurations, leave one configuration thinly evidenced, and it happens to be the one the socialist tradition cares about most. Command socialism abolished the market along with private ownership, and so it confounded two distinct propositions — that capital should be socially owned, and that allocation should be planned rather than priced — beyond any hope of telling them apart. The market-socialist tradition from Lange to Roemer always insisted that these come apart: that one could socialise the ownership of capital while retaining the market for its products. The command economies never ran that experiment. To see the ownership mechanism on its own — collective ownership without the suppression of prices and without the coercion of the one-party state — one has to look smaller, at the voluntary cooperative, and at the single serious attempt to scale worker self-management to a national economy.

Start with the attempt that scaled. Yugoslavia, alone among the socialist states, built its economy on worker self-management rather than central command: enterprises were, in principle, run by their workers, who elected councils and claimed the residual. It is the closest thing history offers to market socialism at national scale, and its record is genuinely mixed rather than simply damning — growth was respectable for two decades — but it displayed, in practice, the pathologies the theory of the labour-managed firm predicts in principle. A firm owned by its current workers has a reason to under-invest in anything whose payoff outlives their tenure, because they cannot sell or bequeath their stake; it has a reason to restrict new hiring, because each new member dilutes the existing members' residual; and absent a capital market that prices the firm's assets, investment decisions lose the discipline that the calculation debate identified as indispensable. Yugoslavia exhibited all three — chronic under-investment, unemployment alongside labour hoarding, and an inflationary softness that grew worse over time — and while its eventual collapse owed more to ethnic and political fracture than to economics, its economic record is the strongest single piece of evidence we have on the national-scale MS configuration, and it is a cautionary one.

Now the voluntary cases, which isolate ownership most cleanly because no one is compelled into them. The Mondragón federation in the Basque Country is the existence proof that worker ownership scales and endures: some eighty-one cooperatives, around 70,500 workers, and €11 billion in sales in 2023 (Mondragón, 2023), built over seven decades and surviving the bankruptcy of its own flagship without abandoning the model. The kibbutz movement is the existence proof that it can also drift: founded on near-total collective ownership, many kibbutzim have over time privatised wages and differentiated pay, a real-world instance of the "degeneration" that the cooperative literature has long worried over, in which successful cooperatives gradually take on conventional features. And on the narrow question of whether worker ownership costs productivity, the evidence is genuinely split and must be reported split: Doucouliagos's (1995) meta-analysis of forty-three studies finds participation and ownership positively associated with productivity, the effect strongest in worker cooperatives, while Craig and Pencavel's (1995) close study of the Pacific Northwest plywood industry finds no significant productivity difference between cooperatives and comparable conventional firms. The defensible reading — and it is the same reading §9 will carry into the synthesis — is that broadened ownership is not a productivity penalty and is plausibly a distributional gain, which is reason to widen the room for it but not warrant to compel it. On the MS configuration, then, the honest verdict is the one the frontier records as a row of asterisks: cooperative ownership is viable, durable, and not a drag on output at the scale of the firm; at the scale of the nation it has been tried essentially once, under conditions that confounded it; and the proposition the socialist tradition most wants tested — markets plus social ownership, at scale, under democracy — remains, remarkably, almost untested. This is not a vindication and it is not a refutation. It is an open question that the command economies' failure has too often been allowed to foreclose, and intellectual honesty requires holding it open.

§6.9 The hard case, examined: state capitalism and the China question

One configuration threatens to swallow this paper, and a comparison that did not meet it head-on would not deserve to be believed. China is the hardest case in comparative political economy: an authoritarian state, with mixed ownership and party-directed credit, that has produced the largest sustained growth and the greatest absolute poverty reduction in human history. If liberty and prosperity were as tightly linked as the liberal story suggests, China should not exist. It does, and the temptation is to read it as proof that the authoritarian road to development works — that the freedoms the previous sections treated as ends can be deferred, indefinitely, in the name of growth. We think that reading is wrong, and saying why is the last piece of the empirical case.

The first thing to see is that China's growth is evidence for the market thesis, not against it. The authoritarian plan, in its pure Maoist form, produced the largest famine in recorded history and three decades of stagnation; what changed after 1978 was not the regime's authoritarianism, which persisted, but its economics, which marketised. The growth arrived with the price system, the household responsibility system, the special economic zones, and the partial opening to private enterprise — which is to say it arrived with capitalism's coordination mechanism, grafted onto a one-party state (Naughton, 2018; Milanovic, 2019). China is not a refutation of the claim that markets out-produce plans; it is one of its largest confirmations. What China does put in question is narrower and sharper: whether the political form — authoritarian rather than liberal-democratic — is compatible with sustained prosperity at the frontier, once the catch-up phase of copying and importing the world's existing technology is exhausted.

On that narrower question the evidence is genuinely not yet in, and honesty requires saying so rather than predicting a fall the data have not delivered. But the theoretical reasons for doubt are the same ones this paper has already developed. State-directed credit is the soft budget constraint wearing a newer suit: when the allocation of capital answers to political rather than to price signals, the calculation problem reasserts itself, not as famine but as malinvestment — the empty cities, the overbuilt sector, the local-government debt that a hard budget would have disciplined. Frontier innovation, as distinct from catch-up, has historically wanted the open contestation that authoritarian systems structurally suppress. And the political form carries its own, non-instrumental verdict: on the liberty and capability vectors, China's authoritarianism is not a neutral means to a prosperous end but a standing deprivation, scored as such, regardless of the growth rate. The frontier records all of this without distortion — China sits at intermediate on prosperity, with the frontier question flagged as open; dominated on political liberty; intermediate on economic freedom — and the sensitivity analysis shows exactly where it falls in each value-frame. Treated as a focal case, China can be made to swallow any comparison; treated as a robustness case, it confirms the frontier rather than overturning it — a capitalism that grew by importing the market, under an authoritarianism whose costs the liberty and capability weightings register in full. That is why this paper holds it at robustness distance: not to dodge the hard case, but because letting the single hardest case set the terms is how comparisons of this kind are usually lost.

§ VIIThe Conditional Frontier

Assemble the seven verdicts and what emerges is not a winner but a map. The table below records, for each reference configuration, whether it is dominant (▲), intermediate (●), or dominated (▽) on each criterion, with the enabling condition that makes the result hold. The cells are ordinal summaries of §6; each rests on a sourced figure.

Criterion	LMC (US/UK)	CSC (Nordic/Ger.)	SC (China/Sing.)	CS (USSR/DPRK/Cuba)	MS (Yugo./Mondragón)
C1 Prosperity & dynamism	▲	▲	● fast catch-up; frontier untested	▽	● micro-scale only
C2 Distribution	▽	▲	▽	●* nominal; shortage-contaminated	▲ within firm
C3 Human development	●	▲	●	●* Cuba health/education	●
C4 Subjective wellbeing	●	▲	● sparse data	● sparse data	● sparse data
C5a Political/civil liberty	▲	▲	▽	▽	● varies with political form
C5b Economic freedom	▲	●	●	▽	●
C6 Resilience & stability	● unbuffered shocks	▲ strong stabilisers	●	▽ transition collapse	▽ self-management instability
C7 Sustainability	▽	● best of a failing set	▽	▽	— insufficient data

A word on the coding, since the cells are the spine of the argument. A configuration is marked dominant on a criterion when it leads the field on that criterion's primary metric by a margin that survives the confounding controls of §4 and is not reversed by a data caveat; dominated when it is clearly last on the same terms; and intermediate otherwise, including wherever wide error bands or sparse data forbid a confident ranking. Two cautions are built into the table and must not be read past. The asterisked cells carry the explicit caveats of §6 — the shortage-contaminated nominal equality of command socialism on C2, and the liberty-bundled, partly contested Cuban result on C3 — and are not clean parity. And the market-socialism column is under-identified, not dominated: because markets-plus-social-ownership has never been run at national scale, its non-dominant cells record the absence of evidence rather than evidence of failure, and its micro-scale marks rest on cooperatives operating inside wider market economies. That column is the paper's largest open question, not one of its settled results.

Read the table by its rows and a shape appears that no single number could carry. The prosperity row is a wall of market coordination: every configuration that uses the price system leads or holds, and the one that abolished it trails alone. The liberty row sorts cleanly by political form rather than by ownership — the two democratic configurations at the top, the authoritarian and one-party ones below, the ownership structure almost irrelevant to where a configuration lands. The distribution and wellbeing rows invert the prosperity story within the capitalist family, the coordinated pole pulling clear of the liberal-market one. And the sustainability row is a line of failures relieved only by the faint "best of a failing set" that is the most any configuration earns. The verdicts do not line up behind one banner; they line up behind mechanisms — the price system, the ballot, the welfare state, the unpriced atmosphere — each of which rewards or punishes whichever configuration happens to embody it, regardless of the camp that claims it.

Three features of the frontier deserve emphasis. The first is that no configuration dominates the field: even the strongest performer, coordinated and social-democratic capitalism, leads outright on only five of the eight criterion-rows and shares the sustainability failure with everyone else. The second is that the dominant axis of variation runs through the capitalist family, not between capitalism and command socialism. The interesting trade-offs — equality and wellbeing against headline output and economic freedom — are intra-capitalist, played out between the liberal-market and the coordinated poles. The third is that command socialism is dominated precisely where the comparison is cleanest — on prosperity, on liberty, on resilience, the criteria the divided-nation experiments measure best — and reaches parity only where measurement is hardest or where a single case carries the weight. Strong where the design is weak, weak where the design is strong: the pattern is itself a finding.

There is a temptation, looking at a frontier rather than a verdict, to hear relativism — to conclude that since every configuration wins somewhere, the choice among them is finally a matter of taste. That is not what the frontier says, and the difference is worth stating precisely. A map on which one configuration won every criterion would license a single ranking; a map on which the wins fell at random would license the shrug. This map is neither. The wins are structured: they cluster, they follow the logic of the mechanisms that produce them, and they are bounded — whole regions of the configuration space, the authoritarian and the comprehensively planned, lose almost everywhere, while the live contest narrows to one well-defined axis within the capitalist family. A bounded, structured disagreement is not the same animal as "everyone has a point." It is a finding about where the genuine disagreement lives, and locating the disagreement precisely is more useful than a manufactured verdict and more honest than a shrug.

§ VIIIThe Verdict Is a Function of Values

The frontier says where each configuration wins. It cannot, by itself, say which configuration is best, because that word imports a weighting, and we promised to show the weighting rather than smuggle it. So we run the comparison under each of the six vectors of §3 and report how the answer moves — but doing so honestly means executing a step the frontier's ordinal marks have so far let us skip, and executing it in the open.

To ask how the verdict moves is to ask how the configurations rank once the frontier's ordinal marks are turned into numbers and combined under each value-frame — and that turning is itself a modeling choice, of exactly the kind this paper has warned against. So we make it in the open. The appendix records the procedure: score each dominant mark as two, each intermediate as one, each dominated as zero; set market socialism aside, since its cells are absences of evidence rather than scores; weight the seven criteria according to each value-frame; read off the ranking; then check it against alternative scorings. The exercise is a stylised illustration and not a measurement — its worth is in what it forces into the open, not in any number it returns.

What it forces into the open is sharper than a purely verbal reading would have been, and in one respect less comfortable for the liberal-market case than §6 implied. The whole result turns on a single decision: whether a value-frame down-weights the criteria it cares less about, or zeroes them. On the inclusive reading — where a libertarian still counts equality and health and sustainability, merely weighting growth and economic freedom more heavily — the coordinated configuration wins under every one of the six vectors, efficiency and liberty included, because its breadth across the criteria swamps the liberal-market economy's narrow leads on raw growth and economic freedom. The liberal-market configuration wins no value-frame at all. At the other end, the exclusive reading — where the libertarian attends to growth and economic freedom and assigns everything else a weight of zero, and the egalitarian likewise looks only at distribution — returns the familiar story: efficiency and liberty fall to the liberal-market pole, the rest to the coordinated one. But these two readings are points on a continuum, not a true dichotomy — the weight a frame places on its priorities is a dial, not a switch — and the appendix locates the crossover between them exactly. What matters here is the shape of the dependence: the result is fixed not by the evidence but by how heavily one frame is allowed to drown out the others, and the "verdict moves with values" thesis is true precisely, and only, to that extent.

Two findings then survive every version of the exercise, and they are what the sensitivity analysis is for. The first is close to invariant and decisive: command socialism finishes last, or tied for last, under every vector and every scoring scheme — down-weighted or zeroed, narrow gaps or wide. The one exception is instructive rather than exculpatory: on the exclusive egalitarian frame, counting only distribution and human development and crediting command socialism's nominal equality at face value, it edges the liberal-market economy — but that nominal equality is exactly the figure §6.2 showed does not survive correction for shortage and privilege, and scoring it honestly returns command socialism to last even there. The second finding is the genuine open contest, now located with precision: the live disagreement is not between capitalism and the plan but between the two poles of capitalism — and the appendix locates it on a dial rather than a switch. The liberal-market economy reclaims efficiency and liberty not only at the zeroing extreme but anywhere it weights its two priority criteria more than roughly five or six times the rest; below that threshold the coordinated configuration's breadth carries even those two frames, and it carries the other four at every weight. So the honest headline is the reverse of a both-sides truce: the coordinated, social-democratic configuration is the robust winner across most of the value-space; the liberal-market configuration prevails only for a growth-or-liberty-first valuer willing to weight those goods several times over above equality, health, and the rest — heavy, but a position a serious economist can hold; and command socialism prevails nowhere.

One caution must be stated as plainly as the finding, because the finding is so easily misread. The artillery that makes the command-socialism verdict so decisive — the divided-nation experiments, the shortage economics, the transition record — bears on command socialism, the comprehensively planned configuration. It does not bear on the configuration the live contemporary debate actually defends. When a young person in a liberal-market democracy says they prefer socialism, they rarely mean Gosplan; they mean, vaguely or precisely, some version of markets-plus-social-ownership, or simply a thicker social-democratic welfare state — and the first of those, as §6.8 conceded, is almost untested, while the second is the very configuration this analysis crowns. The asymmetry, decisive as it is, is therefore not the rebuttal to contemporary socialism it can be dressed up as. It refutes the model almost no one in the room is defending, and it endorses much of what the room is actually asking for. The honest way to carry this result is not as a victory over the left but as a redirection of it: the evidence is hostile to the plan and friendly to the decommodifying market, which is nearer most contemporary socialists' real instincts than the slogans on any side concede.

One clarification guards the result against misreading from both sides at once. The configuration that wins across most of the value-space is a market economy with large-scale redistribution and decommodification, not an economy of social ownership: the Nordic states are among the freer economies on earth on both major indices, and what they socialise is a large share of consumption and risk, not the means of production. The victory recorded here is a victory for redistribution within capitalism — which is why it neither vindicates socialism in the ownership sense nor licenses the conclusion that markets are dispensable. And it is bounded in a further way the next section takes up: the synthesis it points to is the winner of the inclusive weighting, the one a centrist would recognise, and the libertarian who zeroes equality and the degrowther who treats sustainability as an absolute veto would both, with reason, refuse it.

Step back from the particular verdicts, and the method itself makes a claim worth stating plainly. On a question that is part empirical and part normative — which is to say, on most of the questions that actually matter — the honest output of social science is not a verdict but a function: a mapping from what one values to what the evidence then recommends. To collapse that function into a single answer is to do one of two dishonest things — to hide a value-judgment inside a finding, or to pretend the facts are more indeterminate than they are. The conditional frontier refuses both. It reports the facts as determinate where they are determinate — comprehensive planning loses, authoritarianism loses, the welfare state delivers — and the values as open where they are genuinely open, in the pricing of liberty against equality and of growth against sustainability. The reader is entitled to their own weighting and may read their own verdict off the same map. What they are not entitled to, once the frontier is on the table, is the older luxury of mistaking their preference for a discovery.

§ IXBest of Breed, and Where It Breaks

A frontier and a sensitivity analysis between them point toward a synthesis, and we state it deliberately as a modest, evidence-summary recommendation rather than a grand design, because the modesty is part of the honesty. The configuration that performs best across the widest span of value-weightings is recognisable, and it is broadly the coordinated and social-democratic one, hardened at the two margins where even it underperforms. But we should name the condition on that crown, since §8 made it unavoidable: the synthesis is the winner of the inclusive weighting — the reading on which every value still counts for something — and a centrist would accept it precisely because they decline both to price equality at zero and to treat the ecological boundary as an absolute veto. The libertarian who does the first and the degrowther who does the second would each, with reason, refuse the synthesis below. It is not the values-free output of the evidence; it is the evidence read through a moderate, plural weighting that the paper holds but does not pretend to have derived from nature.

Concretely, that configuration combines market coordination and the price system as the default allocation mechanism, because the evidence on prosperity and resilience and the logic of the calculation debate converge on its necessity; predominantly private but genuinely mixed ownership, leaving room for public provision in natural monopolies and for broadened worker ownership at the level of the firm; strong decommodification — universal health, education, and income security — because that is the demonstrable source of the coordinated configuration's advantage on distribution, development, and wellbeing (Esping-Andersen, 1990); a liberal-democratic political form, non-negotiable given the evidence that every authoritarian configuration suppressed the freedoms development is supposed to expand (Sen, 1999); and ecological guardrails — carbon pricing, throughput limits — because sustainability is failed by all comers and will not be solved by the choice of configuration alone. Two mechanisms from the socialist tradition survive the comparison and earn a place, but only in their market-compatible forms: decommodification, socialist in ancestry but realised most successfully inside capitalist economies; and broadened ownership — worker cooperatives and employee ownership, the institutional core of the "real utopias" programme (Wright, 2010) — whose evidence is genuinely mixed and must be reported so. Mondragón demonstrates that worker ownership scales and endures (some 81 cooperatives, 70,500 workers, and €11 billion in sales in 2023), and Doucouliagos's (1995) meta-analysis finds participation positively associated with productivity; but Craig and Pencavel (1995) find no productivity difference between plywood cooperatives and conventional firms. The defensible reading is that broadened ownership is not a productivity penalty and may be a distributional gain — reason enough to widen the room for it, but not evidence for compelling it.

It is worth stating the symmetry plainly, because a synthesis that named its debts to only one side would be a partisan document wearing a centrist hat. From the capitalist tradition the synthesis keeps the load-bearing parts, and keeps them not as concessions but as the evidence's clearest findings: the price system, as the irreplaceable mechanism for coordinating knowledge no planner can assemble; private enterprise and the hard budget constraint, whose absence Kornai traced to the taproot of shortage; and the contestability — free entry, and the freedom to fail — that frontier innovation appears to require. What it adds from the socialist side is bounded and specific: decommodify the essentials, broaden the ownership of capital where that can be done without coercion, and hold the whole arrangement inside a democracy. The result is not a split-the-difference compromise between two equal errors. It is the configuration the evidence points to, which happens to sit nearer one pole than the other, and which we decline to relabel for the sake of a symmetry the findings do not support.

Three boundary conditions discipline the synthesis, each conceded to a critic the evidence makes formidable. The first is institutional complementarity: varieties-of-capitalism theory holds that coordinated-economy institutions reinforce one another (Hall & Soskice, 2001), so the mechanisms above cannot be freely bolted onto a liberal-market economy and expected to cohere — the recommendation describes a complementary package, not a menu from which items may be ordered singly. The second is political precondition: the decommodifying welfare state was historically the product of labour mobilisation and encompassing bargaining, not of designers selecting mechanisms, and it does not transplant without the union density, the corporatist institutions, and the social trust that sustain it. The third is symmetric institutional realism: just as command socialism foundered on government failure, the synthesis must be costed against the public-choice pathologies — rent-seeking, regulatory capture, the fiscal common-pool — rather than assuming the benevolent, competent state its critics rightly refuse to grant. And because a recommendation that cannot be broken is not a finding, we commit three falsifiers in advance. If a large economy were to achieve sustained frontier-level growth under comprehensive central planning — if the computational-planning programme delivered at scale what Kornai's analysis says it cannot — the case for retaining market coordination would fail. If liberal-market economies were, over a sustained period, to match the coordinated ones on distribution and development without converging on their decommodification policy, the causal weight placed here on the welfare state would be misattributed. And if a coordinated economy reached absolute ecological sustainability without sacrificing its performance on the other criteria, the framing of sustainability as an unsolved frontier gap would be wrong. None of the three is logically impossible; each is, on current evidence, unrealised; the synthesis stands only until one of them occurs.

A last honesty is owed about how a configuration like this actually arrives, because nothing in the evidence guarantees that the better arrangement wins. The coordinated settlements were not designed in seminars and adopted on their merits; they were forced by organised labour, ratified in moments of emergency, and sustained by institutions — encompassing unions, corporatist bargaining, dense civil society — that the liberal-market economies have spent four decades dismantling. The historical record suggests that institutional paradigms shift not through gradual persuasion but through crisis windows — the Depression and the stagflation of the 1970s are the canonical instances — in which arrangements previously unthinkable become, briefly, available. If that is right, the practical value of an exercise like this one is not that it will move a vote, but that it specifies, ahead of the next window, what the evidence will support building when the window opens. Having the design and the evidence ready is what converts a crisis into a change rather than a spasm.

And here the present paper reaches its own edge, and hands off. The synthesis it recommends is the best configuration the existing evidence supports — but notice what even the best configuration takes as given. It inherits its reward architecture. It asks markets to allocate and a welfare state to redistribute, and it treats the thing the economy actually rewards — extraction or contribution, rent-seeking or value-creation — as a fact of nature to be redistributed after, rather than a variable to be designed. That is the load-bearing assumption this evaluation cannot examine from inside its own frame, because the frame compares the configurations history actually ran, and history did not run an economy designed around motivation. The companion paper in this archive's Motivation thread, Beyond the Binary (Cahill, 2026a), is the attempt to examine it: to argue that capitalism and socialism share a deeper error than either's tabulated failures — a misreading of why people act — and to specify what an economy designed around motivation rather than against it would have to build. The conditional frontier is the evidence that the building should start from the coordinated-market base. What that companion paper then proposes is a regenerative market with a motivation-aligned reward architecture: an economy that keeps the price system and the private enterprise this paper vindicates, retains the decommodification and the broadened ownership this paper credits, and then does the one thing this paper's frame cannot — treats the structure of reward itself, what the economy actually pays people to do, as a variable to be engineered rather than a constant to be redistributed around after the fact. Whether that programme succeeds is a question for that paper and its own evidence, not for this one. But the handoff is exact: the conditional frontier establishes that the coordinated-market base out-performs its alternatives across the widest range of values, and Beyond the Binary asks what the next increment of design, built on that base, would have to specify. This paper ends where that one begins.

§ XWhat Was and Was Not Settled

Some limitations bound these conclusions, and naming them is the last application of the standard. The configurations are not equally well-instrumented: market economies generate dense, comparable statistics while command economies produced opaque and politically curated data, so the distributional and developmental scores for the socialist cases carry wider error bands than the prose can fully convey, and the North Korean figures in particular are external estimates under deep uncertainty. The asymmetry is not only a matter of precision but of direction, which is why we decline to sign it: suppressed prices and hidden shortage flatter the command economies' measured equality and measured output at once, while their unrecorded informal provision and household production penalise them, so the error does not point reliably one way. The indices we lean on carry priors of their own — the two economic-freedom measures, as §6.5 showed, disagree by construction over whether a welfare state subtracts from freedom — and the wellbeing and democracy scores issue from institutions with their own situated vantage. We have tried to use each source only for the comparison it can bear, and to mark the contested ones as contested; a reader who distrusts a given index should substitute their own and re-read the affected row, which the frontier is built to let them do. We observe the configurations history actually ran, not a random assignment, and command socialism was disproportionately adopted in poor, agrarian, war-damaged societies, so some of its underperformance may reflect starting conditions rather than system — the divided-nation design is our main defence against this, but it rests on a handful of cases. Our sharpest negative findings concern command socialism specifically; the market-socialist programme of markets-plus-social-ownership has never been implemented at national scale, and we have tried not to let the failures of central planning stand in for a refutation of a configuration that has barely been tried. And the seven criteria do not exhaust the space of reasonable concerns: we deliberately did not operationalise several values central to the Marxian and republican traditions — exploitation and the wage relation, alienation and meaningful work, class power and non-domination — because they resist the cross-national metrics available here, and a reader who weights them heavily may reach different conclusions about the liberal-market pole in particular. That omission is a choice, not a neutrality.

None of this is offered in the spirit of having ended the argument. The quarrel between these traditions is older than any dataset and will outlive this paper, as it should, because it is partly an argument about what a good life and a good society are — and that is not the kind of question evidence closes. What the evidence can do, and all it can do, is narrow the ground on which the argument is worth having, by clearing away the claims the realised record simply does not support, on either side. It does not support the claim that markets are dispensable. It does not support the claim that a thin safety net is the price of prosperity. With those two set down, the disagreement that remains is smaller, sharper, and more honest than the one we began with — and conducting it on that narrowed ground, rather than in the slogan-against-slogan register the photograph at the start seemed to invite, is the most a paper like this can hope to accomplish.

What remains, when the limitations are granted, is narrower than either camp's slogans and more useful than their truce. Posed as "which is better, capitalism or socialism?", the question cannot be answered. Posed as "which institutional configurations deliver which outcomes, under which conditions, and for which values?", it largely can. Market coordination decisively out-delivers comprehensive central planning on prosperity, dynamism, and resilience, by margins the divided-nation experiments make hard to attribute to anything but system. Liberal democracy decisively out-delivers every authoritarian configuration on the freedoms that liberals and capability theorists alike treat as ends. And the coordinated, social-democratic configuration out-delivers the liberal-market one on distribution, human development, and subjective wellbeing, while matching it on liberty. The verdict is conditional, but its conditionality is bounded: as one's values move from liberty and growth toward equality and capability, the favoured configuration moves from the liberal-market to the social-democratic pole of capitalism — and stops there. It does not travel on to the plan, because the values that motivate socialism turn out, on the realised record, to be better served by a market economy with strong collective provision than by the abolition of the market. What the twentieth century settled is not that one side won. It is that the institutions worth keeping from each can be named — and that they fit together into a configuration whose one unsolved failure, the ecological one, neither tradition has yet answered, and whose one unexamined assumption, the motivational one, is where the next paper begins.

---

Appendix · The Sensitivity Analysis, Executed

The claim in §8 — that the verdict moves with one's values — cannot simply be read off an ordinal table. It requires turning the frontier's marks into numbers and combining them, and since that turning is a value-laden act of exactly the kind this paper warns against, the construction is set out here in full, so that a reader may argue with it rather than take the section on faith.

The scoring. Each cell of the table in §7 is scored: dominant = 2, intermediate = 1, dominated = 0. Three scoring choices are made in the open, since each could be contested. Market socialism is excluded entirely, because its cells record an absence of evidence rather than a level of performance. The subjective-wellbeing cells for the two authoritarian configurations, which §6.4 treats as data gaps rather than findings, are scored as intermediate rather than zero, so that "no data" is not silently equated with "worst." And command socialism's two asterisked cells — its nominal equality on distribution and human development — are scored generously, at 1, even though §6.2 argues they do not survive correction; the closing note shows that scoring them honestly at 0 changes nothing material.

The weightings. Each value-frame is expressed as weights over the seven criteria, with the freedom criterion split into its two axes. Two readings are run, because the result depends entirely on which is used. On the inclusive reading a frame weights its priority criteria three times as heavily as the others, which still count for something. On the exclusive reading a frame counts only its priority criteria and assigns everything else a weight of zero. The priority criteria are: efficiency/growth → prosperity and economic freedom; egalitarian → distribution and human development; capabilities → human development and wellbeing; sustainability → ecological sustainability; liberty → both freedom axes; equal-weight → all seven, evenly.

The result. Under the equal-weight baseline, to show the mechanics, the scores come out CSC 14, liberal-market 9, state-capitalist 5, command-socialist 3. But the inclusive-versus-exclusive split of §8 is itself a simplification: the weight one puts on a frame's priority criteria is a continuous dial, not a two-position switch. Reported as winners, the result is quickly told. Four of the six frames — egalitarian, capabilities, sustainability, and equal-weight — go to the coordinated configuration at every weighting; on those, the liberal-market economy never overtakes, however heavily its rivals' priorities are discounted. The remaining two — efficiency/growth and liberty — also go to the coordinated configuration on the inclusive setting (priority criteria weighted three times the rest), and flip to the liberal-market economy only past a threshold examined just below, the exclusive limit (priority criteria only) being merely the far end of that same dial. Command socialism wins none; and on the sustainability frame the word "winner" denotes only the least-bad among configurations that all fail the absolute test.

Frame	Inclusive 3×	Exclusive	LMC wins at
Efficiency/growth	CSC	LMC	~5–6×
Egalitarian	CSC	CSC	never
Capabilities	CSC	CSC	never
Sustainability	CSC	CSC	never
Liberty	CSC	LMC	~5–6×
Equal weight	CSC	CSC	never

The crossover deserves stating plainly, because it is the one calculation a skeptic will run on these numbers. On the two frames the liberal-market economy can win — efficiency and liberty — its priority criteria outscore the coordinated configuration's only once they are weighted more than about six times the rest (with the freedom-and-growth scores from §7, the two lines cross at a priority weight of six; under the wider 3/1/0 scoring of the robustness check they cross at about five, so the threshold is best read as roughly five-to-six, not a precise six). Below that, the coordinated configuration's breadth prevails even on those frames; on the other four it prevails at every weight. So the honest claim is not that the liberal-market case requires zeroing equality and health — it requires weighting growth-or-liberty more than sixfold above them, which is heavy but finite, and a position a committed growth-first economist could hold without being caricatured.

Two robustness checks. Widening the score gap (dominant = 3, intermediate = 1, dominated = 0) leaves every ranking's order unchanged and command socialism last. And resolving the within-dominant distinction the ordinal table flattens — granting the liberal-market economy its real lead over the coordinated one on raw growth, by docking the latter's prosperity cell from dominant to intermediate — does not flip the inclusive result: the coordinated configuration still leads, 13 to 9 on the equal weight. The win survives; but, as the next paragraph concedes, that the dominant-dominant prosperity tie was scored at all is generous to it.

A third dial: evidential confidence. The aggregation weights criteria, and it weights cardinal scores, but it does not weight identification: a dominant mark earned from the Korean natural experiment and one earned from a Nordic-versus-Anglo matched comparison enter the sum as the same flat two. That matters more than it might, because of where the coordinated configuration's lead actually sits. Decompose its five-point equal-weight margin over the liberal-market economy, cell by cell: +2 on distribution, +1 each on human development, wellbeing, resilience, and sustainability, and −1 on economic freedom, with prosperity and political liberty tied. Four of those five points sit in the distribution–development–wellbeing cluster — precisely the cluster §6.2 concedes is the most softly identified, resting on matched cross-country comparison rather than on a natural experiment. The well-identified cells, prosperity and political liberty, are ties and add nothing to the margin; the better-identified advantages — resilience and sustainability, net of the economic-freedom deficit — supply a single point. So a confidence-honest aggregation, a third dial turned alongside the criterion-weight and cardinalisation dials already run, narrows the result: halving the weight on the three soft cells drops the coordinated lead from five points to three, and zeroing them entirely leaves it ahead by one, carried then only by resilience and sustainability. The verdict survives both, which is the honest headline — but "robust" must be read with the qualification the decomposition forces. The coordinated margin is real and it survives a discount for soft identification; it is nonetheless concentrated in the cells the paper itself trusts least, which is exactly why the §6.2 triangulation — the within-country Thatcher-era and Swedish shifts, the 2008 and 2020 employment evidence — is load-bearing rather than ornamental.

What the construction shows, and what it does not. It shows three things, each with a caveat the honesty of the exercise requires. First, command socialism finishes last or tied-last in every cell — but this result lives in the §6 coding, which the evidence marks dominated on most criteria, rather than emerging from the aggregation; the appendix does not discover it, only declines to disturb it. (The one apparent exception proves the rule: on the exclusive egalitarian frame, crediting command socialism's nominal equality at face value, it edges the liberal-market economy; score that shortage-contaminated equality honestly at zero and it returns to last even there.) Second, on the inclusive reading the coordinated configuration wins all six frames and the liberal-market economy wins none — and here the caveat is sharper, the place this paper's own thesis catches up with it. Adding weighted scores is a compensatory rule: a strength on one criterion offsets a weakness on another, so the rule structurally rewards breadth, and a configuration coded dominant on six of eight rows is built to win it. A real part of "the coordinated configuration wins every inclusive vector" is therefore not an independent discovery but a consequence of near-uniform dominance fed through a compensatory rule — and that rule embeds a value the paper nowhere defends: a low-variance, no-weak-spots stance, close to a soft maximin across criteria. A non-compensatory rule — a minimum threshold on each criterion, or a lexicographic ordering — could in principle crown a configuration excellent on a few axes and poor on the rest, so we ran the non-compensatory alternatives rather than merely name them; asserting a result where the check is cheap to run is exactly the failing this appendix exists to avoid. They do not, in the event, rescue the liberal-market economy. The coordinated configuration also wins under maximin — it is the only configuration with no dominated cell — under a count of dominant cells (six to three), and under a fewest-weaknesses tally (zero dominated cells to two): it leads on peaks as much as on breadth. The single non-compensatory route to a liberal-market win is a lexicographic priority that fixes one criterion — economic freedom, or growth — above all others, and that is just the exclusive weighting of §8 in another hat, already mapped. So the breadth-preference embedded in the additive rule is real — a value-choice smuggled into the arithmetic, the very move §8 warns against, surfacing one storey down — but it is not what carries the result: drop compensation entirely and the coordinated configuration still wins, for the same reason it won the sum, because the evidence codes it strong nearly everywhere; only a refusal to look past one's single favourite criterion returns the other pole. Third, the liberal-market economy reclaims efficiency and liberty not at some exotic extreme but at any priority weight above roughly five- or sixfold, as shown above. What the construction does not deliver, then, is either a precise ranking or a neutral one: the scores are illustrative, the cardinalisation and the aggregation rule are both choices, and the appendix exists to make the dependence of §8's verdict on those choices visible — not to manufacture a decimal, or a value-neutrality, the evidence cannot bear.

---

References

Acemoglu, D., Johnson, S., & Robinson, J. A. (2001). The colonial origins of comparative development: An empirical investigation. American Economic Review, 91(5), 1369–1401. https://doi.org/10.1257/aer.91.5.1369

Acemoglu, D., & Robinson, J. A. (2012). Why nations fail: The origins of power, prosperity, and poverty. Crown.

Bank of Korea. (2023). 2022 estimate of North Korea's economic growth rate. Seoul: Bank of Korea.

Berdine, G., Geloso, V., & Powell, B. (2018). Cuban infant mortality and longevity: Health care or repression? Health Policy and Planning, 33(6), 755–757. https://doi.org/10.1093/heapol/czy033

Cahill, M. (2026a). Beyond the binary: Motivation-aligned institutional design for post-growth market economies. Armchair Scholar.

Cahill, M. (2026b). The equality-indexed monetary-trade system: Reforming global money and trade for equality. Armchair Scholar.

Choi, M., Sempungu, J. K., Lee, E. H., & Lee, Y. H. (2023). Changes in contributions of age- and cause-specific mortality to the widening life-expectancy gap between North and South Korea, 1990–2019. SSM–Population Health, 23, 101445. https://doi.org/10.1016/j.ssmph.2023.101445

Cockshott, W. P., & Cottrell, A. (1993). Towards a new socialism. Spokesman.

Cohen, G. A. (2009). Why not socialism? Princeton University Press.

Craig, B., & Pencavel, J. (1995). Participation and productivity: A comparison of worker cooperatives and conventional firms in the plywood industry. Brookings Papers on Economic Activity: Microeconomics, 1995, 121–174.

Crippa, M., et al. (2024). GHG emissions of all world countries — 2024 report. European Commission, Joint Research Centre (EDGAR).

Dapprich, J. P., & Cockshott, W. P. (2023). Input–output planning and information. Journal of Economic Behavior & Organization, 205, 412–422. https://doi.org/10.1016/j.jebo.2022.10.043

Doucouliagos, C. (1995). Worker participation and productivity in labor-managed and participatory capitalist firms: A meta-analysis. ILR Review, 49(1), 58–77. https://doi.org/10.1177/001979399504900104

Esping-Andersen, G. (1990). The three worlds of welfare capitalism. Princeton University Press.

Fischer, S., Sahay, R., & Végh, C. A. (1992). Stabilization and economic reform in Russia. Brookings Papers on Economic Activity, 1992(1), 77–174.

Fraser Institute. (2025). Economic freedom of the world: 2025 annual report. Vancouver: Fraser Institute.

Freedom House. (2025). Freedom in the world 2025. Washington, DC: Freedom House.

Friedman, M. (1962). Capitalism and freedom. University of Chicago Press.

Goodkind, D., & West, L. (2001). The North Korean famine and its demographic impact. Population and Development Review, 27(2), 219–238. https://doi.org/10.1111/j.1728-4457.2001.00219.x

Greco, S., Ehrgott, M., & Figueira, J. R. (Eds.). (2016). Multiple criteria decision analysis: State of the art surveys (2nd ed.). Springer. https://doi.org/10.1007/978-1-4939-3094-4

Hall, P. A., & Soskice, D. (Eds.). (2001). Varieties of capitalism: The institutional foundations of comparative advantage. Oxford University Press. https://doi.org/10.1093/0199247757.001.0001

Hayek, F. A. (1945). The use of knowledge in society. American Economic Review, 35(4), 519–530.

Helliwell, J. F., Layard, R., Sachs, J. D., De Neve, J.-E., Aknin, L. B., & Wang, S. (Eds.). (2025). World happiness report 2025. Wellbeing Research Centre, University of Oxford.

Heritage Foundation. (2025). 2025 index of economic freedom. Washington, DC: The Heritage Foundation.

International Monetary Fund. (2014). 25 years of transition: Post-communist Europe and the IMF. Washington, DC: IMF.

Kornai, J. (1980). Economics of shortage. North-Holland.

Kornai, J. (1986). The soft budget constraint. Kyklos, 39(1), 3–30. https://doi.org/10.1111/j.1467-6435.1986.tb01252.x

Kornai, J. (1992). The socialist system: The political economy of communism. Princeton University Press; Oxford University Press.

Lambert, K. J., & Fegley, T. (2023). Economic calculation in light of advances in big data and artificial intelligence. Journal of Economic Behavior & Organization, 206, 243–250. https://doi.org/10.1016/j.jebo.2022.12.009

Lange, O. (1936). On the economic theory of socialism: Part one. Review of Economic Studies, 4(1), 53–71. https://doi.org/10.2307/2967

Lange, O. (1937). On the economic theory of socialism: Part two. Review of Economic Studies, 4(2), 123–142. https://doi.org/10.2307/2967609

Lavoie, D. (1985). Rivalry and central planning: The socialist calculation debate reconsidered. Cambridge University Press.

Lerner, A. P. (1944). The economics of control: Principles of welfare economics. Macmillan.

Medina, E. (2011). Cybernetic revolutionaries: Technology and politics in Allende's Chile. MIT Press.

Milanovic, B. (2019). Capitalism, alone: The future of the system that rules the world. Belknap Press of Harvard University Press.

Mises, L. von. (1935). Economic calculation in the socialist commonwealth. In F. A. Hayek (Ed.), Collectivist economic planning (pp. 87–130). Routledge. (Original work published 1920)

Naughton, B. (2018). The Chinese economy: Adaptation and growth (2nd ed.). MIT Press.

Nölke, A., ten Brink, T., May, C., & Claar, S. (2019). State-permeated capitalism in large emerging economies. Routledge.

North, D. C. (1990). Institutions, institutional change and economic performance. Cambridge University Press. https://doi.org/10.1017/CBO9780511808678

Nozick, R. (1974). Anarchy, state, and utopia. Basic Books.

Nussbaum, M. C. (2011). Creating capabilities: The human development approach. Belknap Press of Harvard University Press.

OECD. (2025). Income distribution database (IDD), key indicators (December 2025 update). Paris: OECD.

Piketty, T. (2014). Capital in the twenty-first century (A. Goldhammer, Trans.). Belknap Press of Harvard University Press.

Rawls, J. (1971). A theory of justice. Belknap Press of Harvard University Press.

Roemer, J. E. (1994). A future for socialism. Harvard University Press.

Sen, A. (1985). Commodities and capabilities. North-Holland.

Sen, A. (1999). Development as freedom. Alfred A. Knopf.

Sinn, H.-W. (2000). Germany's economic unification: An assessment after ten years (NBER Working Paper No. 7586). National Bureau of Economic Research.

Stuckler, D., King, L., & McKee, M. (2009). Mass privatisation and the post-communist mortality crisis: A cross-national analysis. The Lancet, 373(9661), 399–407. https://doi.org/10.1016/S0140-6736(09)60005-2

United Nations Development Programme. (2025). Human development report 2025. New York: UNDP.

V-Dem Institute. (2026). Democracy report 2026 and V-Dem dataset v16 (Liberal Democracy Index, 2025 data). Gothenburg: V-Dem Institute, University of Gothenburg.

Wood, G. T., & Allen, M. M. C. (2020). Comparing capitalisms: Debates, controversies and future directions. Sociology, 54(3), 482–500. https://doi.org/10.1177/0038038519895937

World Bank. (2022). Four decades of poverty reduction in China: Drivers, insights for the world, and the way ahead. Washington, DC: World Bank.

World Bank. (2024). China overview. Washington, DC: World Bank.

World Inequality Lab. (2025). World inequality report 2026: Country sheets. Paris: World Inequality Lab.

Wright, E. O. (2010). Envisioning real utopias. Verso.

---

Editor's note · Motivation thread

This is the first paper in the Motivation thread, and it is positioned above the second. Beyond the Binary (No. 002) picks up from where this paper ends: it argues that capitalism and socialism share a buried error — both treat human motivation as a fixed input to be unleashed or constrained — and designs an economy around motivation instead. The Conditional Frontier is the evidence underneath that move: a same-standards evaluation of the configurations history actually ran, ending at the finding that the live disagreement is intra-capitalist, that the coordinated-market base out-performs both the liberal-market variety and the plan across the widest range of values, and that even this winning configuration inherits its reward architecture rather than designing it. Read the frontier first for the evidence, or the design first for the destination; they were written to close a loop. The throughline of the thread holds: name the load-bearing assumption the dominant framing will not — here, that an economy's incentives are given rather than chosen — and then say something specific enough to be wrong about.

The appendix executes the sensitivity analysis described in §8; the source ledger (appendix-B-source-ledger.md) records confidence flags for every cited figure. Comments welcome at contactme​@​marshallcahill.com.