AXIOCRACY: Rule by Value Created

You can see the future first in the ledger.

Open any company’s books, any ministry’s budget, any research lab’s grant report, and you will find the same quiet fiction. Money goes in. Value comes out. And somewhere in between, a decision gets made about who created it: who gets the bonus, the promotion, the equity, the credit, the next grant. That decision is almost never written down. It is made after the work is done, by whoever holds the pen, according to rules nobody agreed to in advance. We call it management. We call it the market. We call it merit. Mostly, it is power wearing the costume of judgement.

For most of history this was tolerable, because there was no alternative. You could not break a year of collaborative work into its contributions. You could not trace which decision, which line of code, which phone call, which quiet act of maintenance moved the goal. The calculation was impossible, so we outsourced it to hierarchy and to price, and we told ourselves that whatever came out was roughly fair.

That excuse is expiring. Work now leaves a trace. Every commit, every document, every decision, every handoff between a person and a machine is logged somewhere. Within a few years, a large share of the work in any serious organisation will be done by software agents, and the question “who (or what) created this value?” will stop being philosophical and become the central operating problem of the economy. Every firm deploying agents will have to answer it, every day, at scale, or it will not be able to steer them at all.

When that happens, there are two roads. On the first, credit assignment is done by opaque systems optimising metrics nobody chose, owned by a handful of platforms: the Uber rating, generalised to all of human life. On the second, the goals and the measures are agreed openly, before the work starts, by the people who will be measured, and value is credited against them, transparently, contestably, and in public.

The second road needs a name. I propose one: Axiocracy, from the Greek axia, worth, and kratia, rule. Rule by value created.

This is not a utopia. Every piece of it already exists somewhere: in commons that have survived for centuries, in health systems that pay for outcomes, in aid contracts that pay per child immunised, in a sovereign state that ties its ministers’ bonuses to the income of its poorest fifth. It has also failed, repeatedly and instructively, and those failures are the most valuable part of the record. What has never existed is the whole: a coherent principle, a name, a charter, and machines capable of running it.

Axiocracy is not a claim about what people are worth. It is a claim about how we should decide, together and in advance, what counts. It makes one bet: that the most legitimate way to distribute the fruits of shared work is to agree what we are trying to achieve before we start, and then to honour that agreement when we finish.

Let me show you why I think that bet is now winnable.

I. The Broken Ledger: Who Creates Value, and Who Gets Paid

The question is older than economics, and it is the question economics was founded to answer: who creates value, and are they paid for it?

Adam Smith and David Ricardo thought the answer lay in labour. The marginalists thought it lay in the last unit of whatever was added. John Bates Clark, at the end of the nineteenth century, wrote the most beautiful sentence in the whole debate. The natural law of distribution, he argued, gives “to each what he creates.” It was a normative ideal dressed as a description. Clark believed that competitive markets would deliver it automatically. They did not.

The wedge

Here is the simplest measurement of the gap between Clark’s ideal and the world we actually live in. Between 1979 and 2017, net productivity in the United States kept climbing, and median compensation did not keep up. The wedge between the two lines opened to roughly 43 percentage points. Had the typical worker’s hourly compensation tracked the value their economy was producing, it would have been about $33 in 2017 instead of about $23. Nearly ten dollars an hour, every hour, for the median worker, went somewhere else.

You can argue about the measurement. Economists do: how to deflate, what counts as compensation, whether productivity is being mismeasured. The more careful studies find that pay and productivity are still partly linked for the typical worker. But no serious reading of the data says that the typical person is credited with what they create. The link has been stretched until it is barely a link at all.

And this is the aggregate story, the one that averages everything out. Go down to the level of a single team, and it gets worse.

Credit after the fact

Think about the last serious project you were part of. A product launch, a policy, a paper, a building. Now answer honestly: when the rewards were handed out (the bonuses, the promotions, the author order, the speaking slot, the next budget), were they decided according to rules everyone knew before the work began?

Almost certainly not. They were decided afterwards. By a manager, a committee, a senior partner, a principal investigator. According to criteria that were partly stated, partly implied, and partly invented on the spot. The person who fixed the thing at 2 a.m. and the person who presented it to the board were weighed on scales that nobody could see. Those who were close to the money, close to the decision, close to the story, did well. Those who did the maintenance, the mentoring, the invisible glue work, did not.

This is not a moral failing of managers. It is a structural feature of how we assign credit. Value is measured after the fact, by the powerful, against goals that were never made explicit. Under those conditions the outcome is predictable. Credit flows toward visibility and proximity, not toward contribution.

Mariana Mazzucato has spent a decade showing the same pattern at the scale of whole economies. Whoever defines what counts as “productive” controls the distribution of the rewards. Public research that made the smartphone possible goes uncredited, while those who capture value late in the chain present themselves as its creators. The line between value creation and value extraction is drawn by whoever holds the ledger.

Why meritocracy did not fix it

We already have a word for the idea that rewards should follow contribution. It is meritocracy, and it has failed in a specific, instructive way.

Michael Young coined “meritocracy” in 1958 as satire. His book imagined a Britain where intelligence plus effort had become the only measure of a person, and it ended with the meritocrats overthrown by the people they had ranked. Scholars largely ignored the book. The word entered the language overnight, and within two decades it had flipped: the warning became an ideal, then a justification for the winners.

Meritocracy failed for three reasons, and each of them is a lesson.

First, it measured the wrong thing. Merit is a latent property of a person: talent, credentials, IQ, the right schools. It is an estimate of what someone might contribute, frozen into a status. Frank Knight saw the problem in 1923. Productive capacity, he wrote, comes from “a complex mixture of inheritance, luck, and effort, probably in that order of relative importance.” Only the effort has any ethical claim.

Second, it measured on one scale. Merit became a single ladder, and as Raymond Williams put it, a ladder “is a device that can only be used individually; you go up the ladder alone.” One scale means one hierarchy, and one hierarchy means humiliation for everyone near the bottom. Michael Sandel and Daniel Markovits have documented where that leads: winners who believe they deserve everything, and losers told that they deserve nothing.

Third, and most important, nobody ever agreed what merit was. It was defined by those who already had it.

The calculation problem

Why have we put up with this for so long? Because the alternative was impossible.

To credit people by the value they actually created, you would need to decompose a collaborative outcome into its contributions. You would need to know, for each piece of work, how much it moved the goal. For most of history that information did not exist. Work happened in heads, in rooms, on paper, and left almost no trace. The only feasible way to allocate credit was to hand the job to a hierarchy (a boss decides) or to a market (a price decides), and to accept whatever unfairness came with it.

The planners of the twentieth century hit the same wall from the other side. They wanted to set goals centrally and measure progress continuously, and they drowned in data they could not collect or process. In 1971, Stafford Beer’s Cybersyn project in Chile tried to wire a national economy’s factories to a control room using telex lines. It was a glimpse of the future, and it was decades early.

That wall is coming down. Work now leaves a trace. The collaborative outcome of a modern organisation is recorded in version control, in ticket systems, in documents with edit histories, in message threads, in the logs of the software agents that increasingly do the work alongside us. Decomposing value is no longer computationally impossible. It is merely unaddressed.

And that is the danger. When something becomes technically possible, it gets built, whether or not anybody agrees on how it should work. If we do not decide how value should be credited, the answer will be decided for us, inside systems we cannot see.

The broken ledger is about to be rewritten. The only question is who holds the pen.

II. Rule by Value Created: What Axiocracy Is

New orders need new words. Reinhart Koselleck, the great historian of concepts, showed that words like capitalism or citizen were not merely labels attached to things that already existed. They created fields of action. They gave people something to organise toward, argue about, and build. “Capitalism” was coined as an insult, around 1850, and became a neutral description of the world. “Meritocracy” was coined as satire and became an ideal. The coiner of a word does not control what it becomes. That is a reason to define it with care.

So let me be precise.

The definition

Axiocracy (axia, worth or value, and kratia, rule) is rule by value created. An organisation, network or economy is axiocratic when it has three features:

1. Decomposition. The work is broken into identifiable contributions, by people, teams and machines, so that each can be seen.
2. Ex-ante agreement. The goals, the measures of progress toward them, and the rule that turns measured value into credit are agreed before the work starts, by the people who will be measured.
3. Proportional credit. Each contribution is credited according to its measured value against those agreed goals, and the record is open to inspection and appeal.

That is the whole of it. Everything else in this series is an elaboration, a defence, or a set of guardrails.

The second feature is the one that matters most. It is the move that distinguishes Axiocracy from every previous attempt to “reward contribution.” In almost every existing system, value is judged after the fact. Axiocracy insists that the judgement be constituted before it. The measurement becomes a contract rather than a verdict.

Why “before” changes everything

Philosophers have argued for centuries about whether values can be compared at all. Ruth Chang’s answer is, I think, the strongest anchor for this whole project. Comparison is always comparison with respect to something: a “covering value.” A sonnet and a bridge cannot be ranked in the abstract. They can be ranked with respect to a purpose. Fix the purpose, and comparison becomes possible.

That is what an ex-ante agreement does. It fixes the covering value. Once a team has agreed that this is what we are trying to achieve, and this is how we will know, contributions to that goal can be compared, and the comparison can be justified to the people being compared.

John Rawls, who is usually read as the great enemy of desert, supplies the other half. Rawls rejected the idea that people deserve their natural talents. But he accepted legitimate expectations: those who do what a just system has announced it will reward are entitled to that reward. The announcement is what creates the entitlement. Axiocracy is a system built entirely around the announcement. Declare the goals, declare the measures, declare the credit rule, and then honour them.

This is also why Axiocracy makes a much smaller moral claim than meritocracy, and a much more defensible one. Meritocracy says: you are worth more. Axiocracy says: we agreed what we were trying to do, and this is how much your work moved it. Value credit is not moral worth. A person with little credit in one project is not a lesser person. The number describes a contribution to a goal, not a soul.

What Axiocracy is not

The “-ocracies” are a crowded family. Axiocracy has to be distinguished from its siblings, or it will be stretched into all of them.

It is not meritocracy. Meritocracy rewards latent merit, the estimated capacity of a person, measured by credentials and tests and frozen into status. Axiocracy rewards realised value, what a contribution actually achieved, measured against a goal the contributor helped set. Merit is a property of people. Value created is a property of work.

It is not technocracy. Technocracy, and its cousin epistocracy, gives authority to those who know. Axiocracy gives authority to prior consent. Experts may design the measurement instruments, but the goals and the weights are agreed politically, by the people affected. The political layer (what counts?) and the technical layer (how much did this move it?) are deliberately kept apart. The engineers of Technocracy Inc. in the 1930s wanted to measure everything in a single unit, energy, and to abolish politics. Axiocracy wants the opposite: plural measures, and more politics at the moment the rules are set.

It is not algocracy. “Algocracy,” the rule of algorithms, is already here. Authority is exercised every day through systems that rank, score and route people, with objectives chosen by whoever owns the system. Axiocracy uses the same machinery for the opposite purpose. The objectives are public, agreed and versioned. The code that computes credit is treated like law: readable, amendable, and appealable.

It is not holacracy. Holacracy distributes authority through roles and governance meetings. It says a great deal about who decides and almost nothing about how value is credited. Axiocracy is about the ledger.

Where the idea comes from

Axiocracy is a new word for a very old intuition, and it has more ancestors than inventors.

Aristotle’s distributive justice held that goods should be divided “according to worth,” kat’ axian. The word axia is right there. Elinor Ostrom, studying commons that had survived for centuries, found that the durable ones share a design principle: the benefits people draw are proportional to what they put in, and the rules are made and changed by those they bind. A 2010 review of the evidence found that this proportionality principle is significantly associated with commons that last.

The closest modern statement comes from a small open-hardware network in Montreal, Sensorica, which tried to run itself on contribution accounting. Its members wrote that the method for sharing benefits “must be established at the beginning of the economic process, in a transparent way. It constitutes a contract among participants.” That is the axiocratic principle almost word for word.

Axiocracy takes that intuition, which was proven in the commons, stated in the cooperatives and attempted in the networks, and proposes it as a general principle of organisation, at a moment when machines finally make it practical.

A test for any institution

The definition gives us a simple test. For any organisation you belong to, ask three questions:

• Can you see who contributed what?
• Were the goals and the measures agreed before the work started, and did you have a say?
• When the rewards were distributed, could you check how they were computed, and could you appeal?

Most institutions fail all three. Some pass one. Almost none pass all three. The distance between where we are and where we could be is the subject of the rest of this series.

III. The Grammar Already Exists

The most common reaction to Axiocracy is that it sounds utopian. Credit everyone by the value they created, against goals agreed in advance? Nobody does that.

In fact, quite a lot of people do. They just don’t call it anything, and they have never been put side by side. When you do, a shared grammar appears, one that has been reinvented independently in health care, in foreign aid, in public finance, in social investment, in firms and in government:

Agree the outcome. Publish the price. Verify independently. Pay for value delivered.

Let me walk up the scale.

The commons

Start with the oldest. Elinor Ostrom won the Nobel Prize in economics for showing that communities can govern shared resources (irrigation systems, fisheries, forests, pastures) for centuries without either a state or a market. Her design principles for durable commons have since been tested against dozens of cases. One of the best supported says that the benefits people draw from the commons are proportional to what they contribute, in labour, material or money. Another says that those affected by the rules help make and change them.

Put those two together and you have the axiocratic core, running in villages that have never heard of either Ostrom or Axiocracy. Contribution and benefit are linked, and the link is set by the contributors.

Health care

In 2010 Michael Porter argued in the New England Journal of Medicine that the overarching goal of health care should be value, defined as the health outcomes achieved per dollar spent. Not the number of procedures. Not the number of beds. Outcomes, measured over the full cycle of care for a medical condition. He noted that of 78 standard quality measures then in use in the US, all but five measured process, and none measured true outcomes.

Value-based health care has since become a global movement. It is imperfect and contested, but it has changed the question hospitals ask about themselves. That shift, from what did we do? to what did it achieve?, is the axiocratic shift.

Aid between nations

The Center for Global Development proposed Cash on Delivery Aid: a contract in which a donor pays a fixed amount for each unit of confirmed progress toward an agreed goal. The donor does not dictate how. Progress is verified by a third party, and the contract, the progress and the payments are all public. In one worked example, a donor pays $20 for every child completing primary school up to the baseline, and $200 for every child above it, rewarding only the increment. The vaccine alliance GAVI had already done something similar, paying $20 per additional child immunised.

Agree the what, not the how. Pay for the increment. Verify independently. Publish everything.

A nation’s clinics

Rwanda became one of the pioneers of performance-based financing, scaling it nationally in the late 2000s. Clinics were paid for results (quantity and quality of care) rather than inputs. Inside the clinics, the money had to be divided among the staff. In one scheme, facilities distributed payments “among personnel according to previously agreed criteria that captured the relative contributions of staff.”

That is Axiocracy, in a health centre in East Africa, a decade and a half before this essay.

Social investment

Impact bonds let investors pre-finance a social programme, and outcome funders repay them only if agreed outcomes are independently verified. Outcome funds publish rate cards, public price lists for units of social value, and let providers compete to deliver them.

The best-documented case is the Educate Girls bond in Rajasthan, 2015 to 2018. Payments were weighted 20% on enrolment and 80% on learning. In year one the programme hit only 23% of its learning target. By year three, with the freedom to change course that outcome contracts allowed, it reached 160% of its three-year learning target, verified by an independent evaluator.

A firm of 60,000 people

Haier, the Chinese appliance maker, reorganised itself into thousands of autonomous micro-enterprises under a model called rendanheyi, “the integration of people and goals.” Each unit’s performance is measured against a combination of financial targets and user value added, and its members’ pay is tied to that performance. Even internal functions such as legal sell their services to other units for a fee. Every contribution has a price.

A sovereign cabinet

The most striking precedent sits at the very top of a state. Since 2012, a large part of Singapore’s ministerial pay has been a National Bonus tied to four indicators agreed in advance and weighted equally: real median income growth, real income growth of the lowest 20%, unemployment, and real GDP growth. The rulers’ bonus depends on the income of the bottom fifth. The Prime Minister receives no individual performance bonus at all, “as there is no one to assess his individual performance.”

Whatever one thinks of Singapore’s politics, this is a working answer to a question most democracies never ask: what would it mean to pay those who govern by the value they create for the governed?

Governments keeping value books

Underneath all of this sits the ordinary machinery of public appraisal. The UK Treasury’s Green Book requires specific objectives to be set before a project is appraised, and provides tables for putting a value on things once thought unpriceable. New Zealand’s 2019 Wellbeing Budget required bids to be justified against agreed dimensions of wellbeing, not just fiscal cost. The OECD tracks well-being on a dashboard of eleven dimensions rather than a single number.

Governments already run large machines that value outcomes against objectives set in advance. They use them to choose projects. Axiocracy proposes using the same logic to credit contributions.

And the honest caveat

The grammar exists. So does the counter-evidence, and it should be read with equal care.

The World Bank’s review of performance-based financing, covering nearly forty countries over fifteen years, found that health facilities improved, but that the performance-pay link itself was probably not the driving force. Giving facilities autonomy and flexible money did most of the work. Worse, rewarding measured dimensions of care led to the neglect of unmeasured ones. Verification could consume up to a third of administrative costs. The evidence base for impact bonds remains thin. Haier’s model depended on labour laws that would not survive in Europe. And Singapore’s meritocracy has become self-reproducing, with more than a billion dollars a year spent on private tuition.

These are not reasons to abandon the grammar. They are its specification. Reward every valued dimension or none. Verify cheaply, with algorithms and risk-based sampling. Pair credit with genuine autonomy. Protect labour rights by design. Weight shared goals toward the median and the bottom.

The precedents prove the grammar can be spoken. The failures teach it how to speak well. Axiocracy’s contribution is to say, clearly and in one place, that this is a single principle, and that it can be generalised.

IV. The Machines Make It Possible, and Necessary

For a century, the strongest objection to any scheme of crediting value was practical: you cannot compute it. The information was not there, and even if it were, nobody could process it. That objection is dissolving in front of us, faster than almost anyone outside the technology industry realises.

The trace

Start with the simplest fact. Knowledge work now leaves a record.

Code lives in version control, where every change has an author, a timestamp, a description and a review. Decisions live in documents with complete edit histories. Tasks live in ticket systems that record who picked them up, who unblocked them, and when they closed. Conversations live in searchable threads. Customer outcomes live in analytics. None of this was designed for crediting value. All of it is the raw material for doing so.

Ten years ago, turning this exhaust into a credible account of who moved which goal would have required an army of analysts. Today a capable model can read an entire project’s history in minutes and produce a structured, cited account of contributions: what was done, by whom, and how it connects to the stated objectives. It will be imperfect. So is every performance review ever written. The difference is that the model’s account can be inspected, challenged, re-run and audited, which cannot be said of the manager’s recollection in a December calibration meeting.

Evaluation is easier than production

AI alignment researchers have converged on an insight that turns out to be the formal backbone of Axiocracy. Jan Leike and his colleagues built their approach to supervising powerful systems on a simple assumption: evaluating an outcome is easier than producing it. You may not be able to write a great paper, but you can tell a great paper from a poor one. You may not be able to design a bridge, but you can check whether it stands.

That asymmetry is what makes decomposition tractable. You do not need to know in advance how every piece of work should be done. You need to agree what success looks like, and then evaluate each contribution against it, recursively, through a hierarchy of evaluators, human and machine. The same idea that allows humans to supervise systems smarter than they are allows an organisation to credit contributions it could never have planned.

Constitutions for machines

The second convergence is even more direct. The leading approach to shaping AI behaviour, Constitutional AI, governs an automated evaluator with a short, written list of principles set in advance. The model’s outputs are judged against the constitution, not against the ad-hoc preferences of whoever happens to be reviewing them.

Then, in 2023, Anthropic and the Collective Intelligence Project went a step further. They asked a representative sample of about 1,000 Americans to write the constitution themselves, through an online deliberation platform. Participants cast more than 38,000 votes on over a thousand proposed principles. The resulting public constitution was used to train a model, which performed as well as the one trained on the company’s own constitution while showing less bias on several measures.

Read that again with Axiocracy in mind. A thousand people agreed, in advance and in public, on the principles an automated evaluator would apply. The evaluator applied them at scale. That is the axiocratic loop, already working, in the most advanced technology on earth.

The agentic firm

Now add the development that makes all of this urgent rather than merely possible.

Within a few years, a large share of the work in serious organisations will be done by software agents. Not chatbots answering questions, but agents executing multi-step tasks: writing and shipping code, drafting and filing documents, researching, negotiating, reconciling. They will work alongside people, hand work back and forth with them, and increasingly with each other.

An organisation like that has a problem that no organisation has faced before. It must continuously answer: which of these thousands of contributions, by people and by machines, actually moved our goals? Without an answer it cannot allocate budget between agents. It cannot tell which human judgement is worth keeping in the loop. It cannot decide which workflows to scale and which to shut down. It cannot pay anyone fairly.

Credit assignment stops being a question of fairness and becomes the operating system of the firm. Every company deploying agents at scale will build some version of it. The only question is what kind.

The two roads, again

This is why Axiocracy is not an academic exercise. The machinery of measuring value will be built regardless. It is already being built, inside the metrics dashboards of every platform and every agentic workflow product.

The default version will look like the systems we already know: objectives chosen by the owner, metrics nobody else can read, scores that follow you around, ratings by strangers with no appeal. Trebor Scholz, writing about the platform economy, put it bluntly: every Uber has its Unter. A badly designed credit-assignment engine is the Uber rating generalised to all of working life.

The axiocratic version uses the same machinery with different rules. Goals are agreed before the work. Measures are public and versioned, like law. Every credit decision has an audit trail, an explanation and an appeal. The constitution that governs the evaluators is written by those it governs.

There is a warning here from the alignment research itself. Pan and colleagues showed that more capable optimisers exploit badly specified objectives harder, and that the exploitation can arrive suddenly, in “phase transitions” that give little warning. Agents will game measures faster and more thoroughly than people ever did. That is not a reason to avoid measurement. It is a reason to design the measures in the open, with the people who will be measured, and to revise them constantly. Opaque metrics, gamed by capable agents and defended by nobody, are the worst of all worlds.

The machines have made rule by value created possible. They are about to make some form of rule by measured value inevitable. The only open question is whether it will be agreed or imposed.

V. The Challenges

Every “-ocracy” carries the seed of its own corruption. Democracy can become the tyranny of the majority. Meritocracy became an alibi for inherited privilege. Technocracy became rule by people who could not be voted out.

Axiocracy is no exception, and it would be dishonest to present it without its failure modes. I take the four most serious in turn. Each has a long evidence base, and each points to a specific design answer.

Va. The Gamed Measure

The oldest objection to any system that rewards measured performance has a name, or rather several. Goodhart’s law: when a measure becomes a target, it ceases to be a good measure. Campbell’s law: the more a quantitative indicator is used for social decision-making, the more it distorts the processes it was meant to monitor.

The evidence is overwhelming, and it should be taken seriously.

How measures eat what they measure

Theodore Porter, the historian of quantification, tells the story of the US Forest Service, which raised its projected timber growth rates to “draw the teeth” from a law requiring sustainable harvests. When managers are judged by the accounts, Porter observes, they learn to optimise the accounts. Wendy Espeland and Michael Sauder showed how law school rankings remade the schools they ranked. Institutions reorganised themselves around the formula, not around legal education. Sally Engle Merry documented how global indicators crystallise over decades, locking in the choices of whoever designed them first.

The health evidence from the previous chapter adds a sharper finding. When the World Bank studied performance-based financing across dozens of countries, it found that paying for measured services increased idle capacity on the unmeasured ones. Providers did what they were paid for and dropped what they were not.

And the AI evidence adds a frightening one. Pan and colleagues found that as optimisers become more capable, they exploit misspecified objectives more aggressively, and that behaviour can shift abruptly at capability thresholds, with little warning. An organisation full of capable agents, each optimising a proxy, is Goodhart’s law with a turbocharger.

Why this does not defeat Axiocracy

Here is the uncomfortable truth that critics of measurement rarely confront: every organisation already runs on measures. Revenue, headcount, hours billed, papers published, tickets closed, visibility to the boss. The choice is not between measurement and no measurement. It is between measures that are chosen openly and revised deliberately, and measures that are chosen by default and never examined at all.

The worst gaming happens under three conditions: a single measure, chosen by someone else, frozen in place. Axiocracy’s design attacks each of them.

Plural measures, never one number. Value is a dashboard, not a scalar. A goal is expressed through several measures, weighted by agreement, with floors on the dimensions that must not be traded away (safety, ethics, quality). You cannot game a dashboard as cheaply as you can game a number.

Reward every valued dimension, or none. The idle-capacity finding is precise: neglect follows the boundary of what is rewarded. So the boundary must be drawn deliberately. If a dimension of the work matters, it goes into the measures. If it cannot be measured, it is protected outside the ledger rather than silently starved inside it.

Measure the increment, net of luck. Following the Cash on Delivery model, credit the value added above a baseline, not the raw output that would have happened anyway. Social return on investment practice supplies the deductions: deadweight (what would have happened regardless), attribution (what others contributed), displacement and drop-off.

Revise on a schedule. Measures get sunset clauses. Each cycle, participants review whether the proxies still track the goal, and change them if not. Divergence between the proxy and the real outcome is monitored like any other operational risk, because with capable agents it can arrive suddenly.

Keep islands of judgement. Numbers do not rule alone. Qualitative, deliberative review sits alongside the metrics, and the way qualitative judgement is translated into credit is itself documented.

Goodhart’s law is not an argument against Axiocracy. It is an argument against the unaccountable, single-number metric regimes we already live under. Axiocracy is the only proposal that makes the choice of measure a public, revisable, collective act, and that is the only known cure.

Vb. The Aristocracy of Scores

The deepest criticism of Axiocracy is not that it will fail. It is that it will succeed, and produce a new aristocracy.

The meritocracy trap

We have watched this happen once already. “Meritocracy” began as Michael Young’s satire of a society that ranked people by intelligence plus effort. Within a generation it had become a sincere ideal, and then an ideology. Jo Littler traces how it turned into a tool for legitimising inequality: if the system is fair, then the winners deserve to win, and the losers deserve to lose. Michael Sandel calls the result meritocratic hubris. Daniel Markovits shows how the elite now passes its advantages down not through land but through education, which is a more efficient form of inheritance than titles ever were.

Marion Fourcade and Kieran Healy add a sharper diagnosis for a data-rich age. Scored societies, they argue, moralise outcomes: when everything is measured, “everyone seems to get what they deserve.” The scores hide luck and structure. They create new “classification situations,” hierarchies built not on property or occupation but on data. Your credit score, your ratings, your rankings accumulate into a kind of übercapital that follows you everywhere.

An axiocracy built carelessly would be the purest version of this. A public ledger of the value each person created, rendered as a single number, accumulated over a lifetime, used to allocate everything. It would be the most efficient machine for producing smug winners and humiliated losers ever designed.

Luck is not contribution

Frank Knight saw the root problem a century ago. Productive capacity, he wrote, comes from inheritance, luck and effort, “probably in that order.” Crediting the value someone creates, without adjustment, credits their inheritance and their luck. Critics of contribution-based pay, from Michael Albert and Robin Hahnel on the left to Rawls in the centre, make the same point. The output of your work reflects talents and circumstances you did not choose.

There is also the problem of interdependence. Each person’s capacity to contribute depends on the contributions of others, as Samuel Scheffler and Elizabeth Anderson have argued. Isolating “your” contribution from a web of joint production can be an arbitrary cut, like crediting the general for a victory his cook made possible.

The answers

Axiocracy cannot make these objections disappear. It can build its answers into the definition itself, rather than leaving them as optional extras that will be dropped under pressure.

Value credit is not moral worth. This is written into the charter, not into a footnote. An axiocratic credit describes how much a contribution moved an agreed goal. It says nothing about the worth of a person. Rawls himself allows entitlements to agreed rewards while rejecting the idea that people deserve their talents. Axiocracy claims entitlement, never desert.

Status stays plural. There is no single leaderboard, no lifetime score, no portable number that follows a person from context to context. Credit is scoped to a project and a goal, and it expires. Young’s rebels rejected meritocracy “in the name of multidimensional merit.” Axiocracy starts there.

Adjust for circumstance. Where it can be done, measured value is adjusted for circumstances outside a contributor’s control, following John Roemer’s approach of comparing people within their “type.” Rwanda’s health financing added an isolation bonus for remote clinics. It is a small example of the right instinct.

Credit relative to others. Contribution is always measured relative to the joint goal and to the others working on it, using marginal or Shapley-style attribution, not as a solitary quantity. That answers the interdependence objection directly: the method is built on the recognition that nobody creates value alone.

A floor outside the ledger. Axiocracy governs the distribution of surplus, not access to a decent life. Basic security, dignity and the means to participate sit outside the axiocratic ledger entirely. The Peruvian irrigation commons Ostrom studied paired proportional benefit with a subsistence floor. So must we.

The difference test. Finally, a system-level test borrowed from Rawls: an axiocratic scheme is legitimate only if it improves the position of those it credits least, compared to the alternative they would otherwise live under. Singapore wrote a version of this into its ministers’ pay, where a quarter of the National Bonus depends on the income growth of the poorest fifth. Every axiocratic institution should be able to show the same.

An aristocracy of scores is the natural failure of any measurement society. Axiocracy’s claim is not that it is immune, but that it is the first design to name the failure in its own definition and to build the defences in from the start.

Vc. The Watched Worker

If Axiocracy decomposes work into contributions and records each one, then it records a great deal about people. That should make anyone uneasy, and it makes me uneasy.

The surveillance objection

Shoshana Zuboff has described how measurement becomes behavioural control: systems that observe us in order to predict us, and predict us in order to steer us. Trebor Scholz, studying platform labour, argues that constant rating by strangers undermines the dignity of work. A driver’s livelihood hangs on a star rating from people who will never see them again, with no explanation and no appeal. Michel Bauwens and Vasilis Niaros warn that contribution ledgers in peer networks create permanent public profiles of people’s economic lives, open to capture by anyone who can read them.

Elizabeth Anderson makes a subtler point. Fine-grained measurement of individual responsibility is not only intrusive but demeaning. It treats people as objects of constant appraisal rather than as equals.

An axiocracy that logged every keystroke and scored every conversation would deserve every one of these criticisms.

The gift objection

There is a second, quieter danger. Some of the most valuable work humans do is done as a gift: mentoring a junior colleague, answering a stranger’s question in an open-source forum, caring for someone who is struggling. The anthropological tradition from Marcel Mauss onward shows that gift relationships create a kind of social bond that transactions cannot. In peer production, community recognition often matters more than money.

What happens when you put a price on the gift? Sometimes it disappears. The best evidence comes from inside the contribution-accounting movement itself. When the Backfeed protocol was piloted at OuiShare, a European collaborative-economy network, members were asked to evaluate each other’s contributions. The pilot ran into trouble. Defining the scope of contributions was hard, and the model “failed to take into account the feelings that emerged when people had to evaluate the contributions of others.” Members feared it “would actually reduce many social relations … into mere transactions.”

That is a warning from friends, not enemies, and it should be heeded.

The answers

The answer is not to abandon the ledger. The answer is to give it walls.

Only agreed goals enter the ledger. Axiocracy does not measure people. It measures contributions to goals that were agreed in advance. Work that is not tied to an agreed goal is not recorded, not scored, and not anyone’s business. The scope is set by the contract, and it is narrow by design.

Protect a sphere that is never accounted. Care, friendship, mentoring as gift, play and conversation are deliberately kept outside the ledger. Where recognition matters, it flows through non-monetary channels: thanks, acknowledgement, honour. These stay separate from credit, so that pricing does not crowd them out.

Prefer the contract to the crowd. The OuiShare lesson points to the clearest design choice in the whole literature. Emergent, after-the-fact peer evaluation of everything felt like commodification and did not scale. Sensorica’s approach, a value equation agreed at the beginning by the participants, is a contract rather than a popularity contest. Axiocracy takes the contract side, explicitly.

Data sovereignty. Contribution records belong to the contributors. They are scoped, time-bound and portable, and are shared only with consent. The hypercert model developed for impact funding is one design for this: fractional claims defined by contributor, scope of work and time, created only with the contributor’s agreement. No record should follow a person into contexts they did not agree to.

Due process, always. Every credit decision has an audit trail, a notice, an explanation in plain language, and a human appeal. Danielle Citron’s work on “technological due process” gives a ready specification: when automated systems make decisions about people, those people are owed the same procedural protections they would get from a human official.

The right to log off. Participation in any axiocratic scheme is voluntary at the level of the goal. A person can decline to have their work in a project accounted, and accept the consequences for credit, without losing access to the floor described in the previous chapter.

The watched worker is the most likely dystopia of the agentic economy, whether or not anyone ever says the word Axiocracy. The measurement will be built anyway. The only protection is a principle that insists the ledger has a boundary, and that the people inside it hold the keys.

Vd. Who Sets the Measures?

Every argument so far rests on a single phrase: agreed in advance. Which raises the most important question of all. Agreed by whom?

Measurement is power

Joseph Stiglitz, Amartya Sen and Jean-Paul Fitoussi opened their famous report on measuring economic performance with a simple observation: what we measure affects what we do. Measurement is performative. Whoever sets the metric governs.

Mariana Mazzucato makes the same point about the boundary of production. Whoever decides what counts as productive activity decides who gets rewarded. And the long history of accounting shows how fragile the neutral appearance of numbers can be. In the 1970s, British companies briefly published a Value Added Statement showing how the wealth a firm created was split between employees, shareholders, lenders and the state. Making the split visible intensified conflict over it. Managers stopped publishing it, and it disappeared. Researchers have since shown that methodological choices in such statements can swing the apparent split by a factor of four.

If goals and measures are set by the powerful and merely “agreed” by the rest, Axiocracy becomes a machine for laundering existing power through arithmetic. The worst kind of domination is the kind that looks like a fair calculation.

Arrow’s shadow

There is also a deep theoretical limit. Kenneth Arrow proved that no method of aggregating individual preferences into a collective ranking satisfies every reasonable fairness condition at once. Any procedure for agreeing goals and weights will involve a choice that someone can object to. There is no perfectly neutral constitution.

This is true, and it is also true of every democratic institution ever built. The response is not to seek a perfect procedure. It is to choose a good enough one openly, name it, and make it revisable.

The answers

Separate the political layer from the technical layer. Setting goals, choosing measures and fixing weights is political. It is done by the people affected, through deliberation and voting. Measuring contributions against those goals is technical, and is done by instruments and evaluators. Experts can advise on the first and build the second. They never own the first.

Bargaining parity at the table. Those being measured have at least equal standing with those commissioning the work when goals and measures are set. This is the direct answer to the power asymmetries behind the productivity-pay wedge. If the people who create value have no say in how it is defined, the old ledger is simply rebuilt in new code.

Bridging, not majority. Collective Constitutional AI showed that a thousand people can write a set of principles together, using methods that surface bridging statements, ones supported across groups that usually disagree, rather than letting the majority win. For weighting goals, quadratic voting lets people express how much they care, not just which side they are on. The aggregation rule is chosen explicitly and published.

Fix the conventions before the work. Every accounting choice that could swing the result later (gross or net value, the time horizon, which forms of capital count, the discount rate) is fixed in the ex-ante agreement. Axiocracy applies its own principle to its own accounting, and so defuses the disputes that killed the Value Added Statement.

Symmetric disclosure. Value splits are disclosed for everyone, including leaders. The Value Added Statement died because disclosure was optional and controlled by directors. In an axiocratic institution, the people at the top are measured by the same rules, in the same ledger, in public.

Code as law. Following Lawrence Lessig’s insight that code regulates behaviour as surely as law does, the software that computes credit is treated as law. It is versioned, public and readable, has a test suite, and is amended only through the same process that agreed the goals.

Locally sovereign rules. Ostrom’s warning applies: one size does not fit all, and rigid universal formulas destroy the commons they are meant to protect. There is no global axiocratic formula. Each community, firm or project writes its own value constitution within a shared charter of principles.

Measure the legitimacy. Finally, the system is itself measured. Periodically and anonymously, the people governed by an axiocratic scheme say whether they accept it as fair. A scheme that loses the consent of those it measures has lost its only claim to authority, and must be rewritten.

The question “who sets the measures?” is not a flaw in Axiocracy. It is Axiocracy. The whole principle is an answer to it: the people who will be measured, in advance, in public, with the right to change their minds.

VI. The Axiocratic Charter

Principles are only useful if someone can adopt them. What follows is written to be adopted: by a team starting a project next week, a cooperative rewriting its rules, a company deploying its first fleet of agents, or a government ministry deciding how to fund outcomes. Each principle is short. Each is backed by the evidence in the preceding chapters.

The definition

An institution is axiocratic when it decomposes its work into visible contributions, agrees its goals and measures before the work begins with the people who will be measured, and credits each contribution by its measured value against those goals, openly and contestably.

The twelve principles

1. Agree before you work. Goals, measures, weights and the credit rule are fixed before the work starts, and written down. Nothing is credited against a rule invented afterwards.

2. The measured hold the pen. The people whose contributions will be measured take part in setting the goals and measures, with at least equal standing to those commissioning the work.

3. Value credit is not moral worth. A credit describes how much a contribution moved an agreed goal. It never describes the worth of a person, and it is never presented as desert.

4. Measure plurally. Value is a dashboard, not a single number. Goals are expressed through several measures, with floors on the dimensions that must never be traded away.

5. Reward every valued dimension, or none. What matters goes into the measures. What cannot be measured is protected outside the ledger, not silently starved inside it.

6. Credit the increment, net of luck. Credit value added above a baseline, after deducting what would have happened anyway and what others contributed. Where possible, adjust for circumstances outside the contributor’s control.

7. Nobody creates alone. Contribution is always measured relative to the joint goal and to the others working on it, never as a solitary quantity.

8. Keep a floor outside the ledger. Axiocracy distributes surplus, not the right to a decent life. Basic security and dignity are never subject to measured value.

9. Protect the unaccounted. Care, friendship, gifts and play stay outside the ledger. Only work tied to agreed goals is recorded. Everyone has the right to log off.

10. Treat the code as law. The rules and software that compute credit are public, versioned and readable, and are amended only through the process that agreed the goals. Every credit decision has an audit trail, an explanation and an appeal to a human.

11. Revise on a schedule. Every measure carries a sunset date. Each cycle, participants check whether the proxies still track the goal, and change them if they do not.

12. Pass the difference test. The scheme must improve the position of those it credits least, compared to the alternative. Leaders are measured in the same ledger, by the same rules, in public. And the consent of the governed is itself measured. A scheme that loses it must be rewritten.

Starting tomorrow

None of this requires new laws, new technology or anyone’s permission. It can start with one team and one project:

• Before the kickoff, write one page. What are we trying to achieve? How will we know? How will credit, bonuses, author order or equity be shared when we succeed? Everyone signs it.
• During the work, keep the trace honest. Let the record of who did what accumulate where everyone can see it.
• At the end, compute the credit by the rule you agreed. Publish the calculation. Hear the appeals. Then ask everyone what they would change next time.

That is an axiocracy of five people. It will be imperfect. It will also be more legitimate than almost any system of credit its members have ever worked under, because for the first time they will have decided in advance, together, what counts.

Scale follows legitimacy. It always has.

VII. Parting Thoughts

Every era has a question it cannot avoid.

For the eighteenth century it was who has the right to rule? The answer, painfully and incompletely, was the people. For the nineteenth it was who owns what the machines produce? We have been fighting about that answer ever since. For the twentieth it was can a free society also be a fair one? We built welfare states and social insurance, and that answer is still being written.

The question of the coming decade is what is it worth? It is worth asking who created the value, and whether they were credited for it. We will be asking it about every piece of work done in an economy where people and machines collaborate so closely that nobody can tell where one contribution ends and the next begins.

There will be an answer. There always is. The only choice is whether that answer is imposed after the fact, by whoever owns the systems doing the measuring, or agreed in advance, openly, by the people being measured.

What if we are right?

Suppose the argument of this series holds. Suppose that the calculation barrier really has fallen. Suppose that the grammar of agree, publish, verify, pay can be generalised from clinics and aid contracts and commons to firms, networks and states. Suppose that the failure modes can be designed against rather than merely lamented.

Then some things that seem permanent are not.

The productivity-pay wedge is not a law of nature. It is the output of a ledger that was never agreed. The humiliation of those at the bottom of a meritocracy is not the price of efficiency. It is the result of measuring people instead of contributions, on one scale instead of many. The invisibility of maintenance, care and glue work is not inevitable. It follows from counting only what the powerful chose to count. The dread that AI will concentrate all value in the hands of whoever owns the models is not destiny. It is a question of who writes the constitution for the evaluators.

None of these things will change on their own. Measurement is coming whether we like it or not. The only question is its constitution.

The word

I proposed a word at the start of this series, and I will end with a note of caution about it. The coiners of “-ocracies” do not control what their words become. Michael Young watched “meritocracy” turn from a warning into a boast within his lifetime. Axiocracy could suffer the same fate. It could become a respectable name for surveillance, or a new alibi for the winners.

That is why the definition carries its own defences, why the charter says in its third line that value credit is not moral worth, and why the last principle requires the system to earn, continuously, the consent of those it measures. A word cannot protect itself. The people who use it can.

So here is the invitation. Take the charter. Try it on one project. Tell us where it breaks. Rewrite the parts that are wrong. That, too, is the axiocratic method: agree what we are trying to do, try it, measure it honestly, and revise.

It’s time to decide, together and in advance, what counts.

About

Axiocracy: Rule by Value Created is an essay series by Jakub Bareš, published with ENSI in September 2026.

The argument draws on a research library of 129 primary documents across fifteen angles: how political concepts are coined and how they drift; philosophical and economic theories of value; desert theory; valuation and measurement; contribution accounting; public value and measures beyond GDP; technocracy and algorithmic governance; the sociology of quantification; value alignment in AI; commons and peer-value economies; alternative economic systems; the strongest critiques of rule by measured value; and the institutions that have already tried something like it.

Every factual claim in the series traces back to a document in that library. The figures on productivity and pay come from the Economic Policy Institute. The commons evidence comes from Ostrom and from Cox, Arnold and Villamayor-Tomás. The health-financing evidence comes from Porter, the Center for Global Development and the World Bank. The impact-bond results come from Brookings and the K4D helpdesk. The Singapore figures come from the government’s 2012 White Paper on ministerial salaries, and the alignment research from Anthropic, the Collective Intelligence Project, and Pan and colleagues.

Axiocracy is a proposal, not a doctrine. Criticism, counter-evidence and field reports from anyone who tries the charter are the most useful things you can send.